Network path calculation method and network equipment
Patent Information
- Application Number
- CN202280095101.4
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2022-04-29
- Publication Date
- 2024-11-26
AI Technical Summary
The classic Interior Gateway Protocol (IGP) distributed algorithm cannot calculate the optimal network path in a multi-objective factor scenario, resulting in the inability to meet the strict requirements of the industrial Internet for multiple network performance factors, such as latency and bandwidth.
A network path calculation method is proposed. By recording additional advantageous network paths, partial order comparison and iterative calculation are performed in multi-objective factor scenarios to ensure that the multi-objective factor requirements of the application are met. Non-optimal path records are introduced to accurately calculate end-to-end end optimal path.
It achieves accurate calculation of the optimal network path that meets application requirements in a multi-objective factor scenario, improves the optimization capabilities of resource allocation and resource control, breaks through the limitations of existing solutions, reduces computing overhead, and improves the efficiency of network services. real-time.
Smart Images

Figure CN119032591A_ABST
Abstract
Description
A network path calculation method and network device Technical Field
[0001] The present application relates to the field of data transmission, and in particular to a method for calculating a network path and a network device. Background Art
[0002] With the gradual transition from the consumer entertainment internet to the industrial internet, networks are carrying an increasing number of industrial applications (such as smart manufacturing and remote control). Unlike internet services such as social entertainment and consumer shopping, industrial applications typically have more stringent service-level agreement (SLA) requirements, such as latency less than 100 milliseconds (ms) and packet loss less than 10%. Furthermore, industrial internet SLAs often require multiple network performance factors. For example, smart manufacturing applications require network latency, reliability, and packet loss simultaneously. Therefore, the underlying network needs to change its traditional best-effort routing approach based on a single objective factor. It can now calculate the optimal path based on multiple objective factors according to the SLA requirements of different applications, thereby providing personalized quality of service (QoS) guarantees for different applications.
[0003] Classic interior gateway protocols (IGPs), such as Open Shortest Path First (OSPF), use a path calculation method (represented by the Dijkstra algorithm) based on the principle of "local optimal path + local optimal path + ... = global optimal path." Each network node independently calculates and records the optimal path to the destination node. When a network node receives a packet forwarded by a neighboring node, it directly forwards the packet along the locally calculated optimal path to the destination node. Each hop in the end-to-end path follows a local optimal path, resulting in the optimal end-to-end optimal path. As shown in the network topology in subdiagram (a) of Figure 1, there is one path, Path0, from node X to node Z (with a delay of 10 ms, ignoring the dimension here). There are three paths from node Z to destination node Y: Path1 (with a delay of 5 ms), Path2 (with a delay of 10 ms), and Path3 (with a delay of 15 ms). Path1 is the local optimal path from node Z to node Y (because it has the lowest delay). Therefore, in the classic distributed IGP algorithm, node Z only records the locally optimal path Path1. When neighbor node X transmits a service data packet to node Z, node Z will use the locally optimal path Path1 to transmit the data packet to the destination node Y. Therefore, when both hops X→Z and Z→Y follow the locally optimal path, the end-to-end optimal path X→Z→Y is obtained. This principle of "local optimal path + local optimal path + ... = global optimal path" holds true in scenarios with a single objective. However, the principle of "local optimal path + local optimal path + ... = global optimal path" that the classic distributed IGP algorithm relies on does not hold true in scenarios with multiple objective factors. As shown in subdiagram (b) of Figure 1, the same network topology X→Z→Y is used. Assume that each path now has two performance metrics: latency and bandwidth (i.e., two objective factors). There is one path from node X to node Z, Path0 (bandwidth 10, latency 10; the dimensions are ignored here). There are three paths from node Z to node Y: Path1 (bandwidth 20, latency 10), Path2 (bandwidth 10, latency 5), and Path3 (bandwidth 15, latency 15). Assuming the network carries video services, the network performance requirement is to first select the path with the highest bandwidth. If multiple paths have equal bandwidth, the path with the lowest latency is selected. This network performance requirement is a dual-objective optimization of "bandwidth + latency." Based on the application's network performance requirements, the optimal path from Z to Y is Path1, as it has the highest bandwidth.According to the classic IGP algorithm, when a data packet arrives from node X to node Z, node Z will use Path 1 to transmit the data to the destination node Y. The performance of the end-to-end path X→Z(Path 1)→Y is: end-to-end bandwidth 10 (the minimum of Path 0 and Path 1 bandwidths) and end-to-end delay 20 (the sum of Path 0 and Path 1 delays). However, it can be seen that after the data packet is transmitted from node X to node Z, node Z does not use the locally optimal path Path 1, but instead uses the non-optimal path Path 2. The resulting end-to-end path X→Z(Path 2)→Y has an end-to-end bandwidth of 10 (the minimum of Path 0 and Path 2 bandwidths) and an end-to-end delay of 15 (the sum of Path 0 and Path 2 delays), which is actually better than if node Z took the locally optimal path.
[0004] This shows that in multi-objective scenarios, the optimal end-to-end path isn't simply a concatenation of locally optimal paths at each hop. Taking suboptimal paths between certain nodes can often yield an end-to-end path with better performance than taking the optimal path at each hop. Therefore, in multi-objective scenarios, the principle of "local optimal path + local optimal path + ... + local optimal path ≠ global optimal path" makes it impossible to calculate optimal network paths based on classic IGP distributed algorithms.
[0005] Summary of the Invention
[0006] The first aspect of the present application provides a method for calculating a network path, which is used to address the problem that the classic IGP distributed algorithm cannot solve the optimal network path for multi-objective factors. A method for calculating the distributed multi-objective factors of an advantageous network path is designed that can calculate any business needs, thereby improving the ability of resource allocation optimization and resource control optimization.
[0007] Based on this, the embodiments of the present application provide the following technical solutions:
[0008] In the first aspect, an embodiment of the present application provides a method for calculating a network path, which can be distributedly deployed in network nodes or traversed and implemented in a centralized controller; it can be applied to an Overlay network or an Underlay network, and can calculate the optimal network path according to any business requirements. The method may include: first, determining a reachable network path with a source node as the starting point and a directly connected node of the source node as the end point. The directly connected node refers to a node that is directly connected to the source node and does not pass through other nodes in the middle. The reachable network path can be called a first reachable network path (which can be one or more). It should be noted that the reachable network path not only includes the network nodes involved in the reachable network path, but also includes the overall network performance of the entire reachable network path (such as the delay, bandwidth, packet loss rate, etc. of the reachable network path). Among them, the source node is a network node in the network (which can be called the target network) to which the target device (i.e., the device with data to be sent) is connected. The network node can be a router or a server in the target network, etc., which is specifically determined by the type of the target network involved, and this application does not limit this. Afterwards, a node (which may be referred to as the first node) is selected from the set of nodes to be added to obtain a reachable network path with the source node as the starting point and the first node as the relay. The reachable network path may be referred to as the second reachable network path. The set of nodes to be added is a collection of other network nodes in the target network except the source node. It should be noted that, in some embodiments of the present application, the intermediate nodes of the second reachable network path may not include other network nodes in the set to be added except the first node. Moreover, the intermediate nodes of the second reachable network path do not include the first network node (i.e., the source node) and the last destination node (i.e., the destination node) of the second reachable network path, but refer to any network node in the "middle". For example, assuming that the second reachable network path is A→B→C→…→X, where A, B, C,…, and X are all network nodes, then the intermediate nodes of the second reachable network path refer to any network node in B, C,…, but do not include the source node A and the destination node X. All reachable network paths included in the first reachable network path and the second reachable network path can be called target paths (one or more). Afterwards, it is determined whether there are m (m≥2) network paths reaching the same destination node in the target path. If so, these m network paths are compared based on n (n≥2) preset target factors to obtain a dominant network path (one or more). The dominant network path can be called a first dominant network path, and the target path is updated based on the first dominant network path. The updated target path is used as the new first reachable network path. Therefore, updating the target path can also be called updating the first reachable network path to obtain the updated first reachable network path.
[0009] In the above-mentioned embodiment of the present application, on the basis of the classic IGP distributed algorithm, the recording of non-optimal paths is introduced. That is, the embodiment of the present application improves the traditional distributed IGP algorithm. By recording additional advantageous network paths (that is, the first advantageous network path obtained after comparing m network paths based on n objective factors), it is possible to accurately calculate the end-to-end optimal path that meets the network performance requirements of the application multi-objective factors in the multi-objective factor scenario of "local optimal path + local optimal path + ... + local optimal path ≠ global optimal path", breaking through the limitations of all existing solutions and improving the ability of resource allocation optimization and resource control optimization.
[0010] In a possible implementation of the first aspect, after obtaining the updated first reachable network path, the method may further include: deleting the first node from the set of nodes to be added, and repeating the steps of selecting the first node from the set of nodes to be added and subsequent steps until the set of nodes to be added is an empty set.
[0011] In the above-mentioned embodiment of the present application, through an iterative process, it can be calculated that in the target network, all network paths from a source node to any target node as a terminal are advantageous network paths, and the calculation is accurate.
[0012] In a possible implementation of the first aspect, m network paths are compared based on n target factors to obtain a first advantageous network path by performing a partial order comparison on the m network paths based on n target factors to obtain a second advantageous network path, wherein each second advantageous network path serves as the first advantageous network path. This implementation is generally applicable to networks with relatively small scale (such as a real-time network (RTN) with a maximum scale of 100 network nodes), so the number of second advantageous network paths screened out by the partial order comparison method is relatively small. Therefore, even if subsequent iterative calculations are performed based on the full amount of second advantageous network paths, the computational overhead of the overall multi-target factor scenario is not large. For example, in a real-time network with 100 network nodes, the calculation of the optimal path for all end-to-end multi-target factor scenarios can be completed in hundreds of milliseconds.
[0013] In the above-mentioned embodiment of the present application, each second advantageous network path obtained is used as the first advantageous network path to participate in subsequent iterative calculations, so that the first reachable network path obtained by the last update must contain the end-to-end optimal path, with completeness and calculation accuracy.
[0014] In a possible implementation of the first aspect, the method may further include: in the case where the set of nodes to be added is an empty set after the first node, that is, there is no network node in the set of nodes to be added, then the first reachable network path obtained in the last update is the first reachable network path finally obtained in the current round. In the case where each second advantageous network path is used as the first advantageous network path, since the end-to-end optimal path must be included in the first reachable network path updated in the last update of the current round, the first target network path with the source node as the starting point and the destination node corresponding to the data to be transmitted as the end point can be directly determined from the first reachable network path obtained in the last update of the current round according to the network performance requirements of the data to be transmitted on the target device. The first target network path is used to transmit the data to be transmitted.
[0015] In the above-mentioned embodiment of the present application, when multiple iterative calculations are performed until the set of nodes to be added is empty, if each second advantageous network path is used as the first advantageous network path, then the first reachable network path obtained by the last update must contain the end-to-end optimal path of each destination node, and can be transmitted along the end-to-end optimal path according to the business network performance requirements, thereby realizing the calculation of the optimal path in a multi-objective factor scenario of any business requirement.
[0016] In a possible implementation of the first aspect, m network paths are compared based on n target factors to obtain a first advantageous network path. First, a partial order comparison is performed on the m network paths based on the n target factors to obtain a second advantageous network path. When there are at least p second advantageous network paths, q are selected from the p second advantageous network paths as first advantageous network paths, where m≥p≥2 and p≥q≥1. The selected second advantageous network path is the first advantageous network path, wherein the selection method may be random selection or selection based on certain selection rules (e.g., selecting a path with fewer hops or a larger bandwidth, etc.), which is not limited in this application.
[0017] In the above-mentioned embodiment of the present application, when the number of second advantageous network paths reaching the same destination node reaches a certain number p (e.g., p=4), in order to ensure the efficiency of the calculation, the subsequent iterative calculation is no longer based on the full amount of second advantageous network paths, but q (e.g., q=2) paths are selected from the currently obtained second advantageous network paths for iterative calculation, thereby optimizing the calculation overhead.
[0018] In a possible implementation of the first aspect, the method may further include: if the set of nodes to be added is an empty set after deleting the first node, that is, there is no network node in the set of nodes to be added, then according to the network performance requirements of the data to be transmitted on the target device, determine a second target network path with the source node as the starting point and the destination node corresponding to the data to be transmitted as the end point from the first reachable network path obtained by the last update, and the second target network path is used to transmit the data to be transmitted.
[0019] In the above-mentioned embodiment of the present application, this implementation method can be used as an improvement solution for large-scale networks. For services with special needs (such as voice, video and other services with high real-time requirements), the network performance requirements of the data to be transmitted on the target device can be used to select an optimal path from the first reachable network path obtained from the last update for transmission first, which can ensure the real-time performance of the service and improve the user experience.
[0020] In a possible implementation of the first aspect, the first reachable network path obtained in the last update of the current round is obtained based on the selected second dominant network path. In this way, the first reachable network path calculated in the current round based on partial information may not include the end-to-end optimal path. Based on this, the embodiment of the present application adopts a multi-round calculation method to make the end-to-end path calculated based on the partially selected second dominant network path gradually converge to the global optimal path. Specifically, it can be first determined that when there are at least p second dominant network paths, the current set of nodes to be added is the target node set; then, the target node set is used as the new set of nodes to be added, and the steps of selecting the first node from the set of nodes to be added, obtaining the second reachable network path with the source node as the starting point and the first node as the relay, and the subsequent steps are repeated until the first reachable network path obtained in the last round meets the convergence condition or reaches a preset number of rounds (e.g., 50 times). Among them, performing one round of selecting the first node from the set of nodes to be added, obtaining the second reachable network path with the source node as the starting point and the first node as the relay, and the subsequent steps is one round. It should be noted here that in some embodiments of the present application, if multiple rounds of calculations need to be repeated, the operation of selecting q paths from p second-dominant network paths as the first-dominant network paths in the first round can be randomly selected, but in subsequent rounds of calculations, the optimal path among the first-dominant network paths selected in the previous round can be iterated to the next round as one of the feasible solutions. For example, assuming that there are currently 10 full-scale dominant paths, namely paths 1 to 10, and paths 2, 3, and 7 are randomly selected in the first round, and it is found that the optimal path among these three paths is 3, then path 3 will be recorded. In the next iterative calculation, three more paths 1, 5, and 8 are randomly selected; together with the local optimal path 3 from the previous time, there will be a total of four paths, 1, 3, 5, and 8, participating in the calculation of subsequent paths. The current local optimal path (such as 5) is then recorded, and iteration continues until the first reachable network path obtained in the last round meets the convergence condition or reaches the preset number of rounds. It should be noted that in some embodiments of the present application, the method of multi-round calculation can also be: repeatedly executing the steps of determining the first reachable network path with the source node as the starting point and the directly connected node of the source node as the end point and subsequent steps, until the target path obtained in the last round meets the convergence condition or reaches the preset number of rounds.
[0021] In the above-mentioned implementation of the present application, the real-time performance of network services in multi-objective factor scenarios in large-scale networks is improved. The implementation of the present application does not pursue the accurate calculation of the end-to-end optimal path in one go, as this would incur excessive computational overhead and affect the real-time performance of network services. In contrast, the implementation of the present application improves the method of calculating the optimal path in one go, and designs a progressive calculation method that selects some advantageous paths in combination with multiple rounds, i.e., the first round of calculation quickly calculates an approximate end-to-end optimal path based on a small amount of path information, so that real-time services can initiate service data transmission on the approximate optimal path in a timely manner, and then continuously optimizes the accuracy of the approximate optimal solution through multiple rounds of iterative calculations until it converges to the true optimal path.
[0022] In a possible implementation of the first aspect, it is assumed that according to the network performance requirements of the data to be transmitted on the target device, the target network path that is currently transmitting the data to be transmitted is the second target network path, and after the next round or multiple rounds of calculation, the target network path (such as the optimal target network path) that meets the preset conditions is the third target network path, which means that the network performance of the third target network path is better than the second target network path. In this case, the transmission path of the data to be transmitted can be switched from the second target network path to the third target network path. The switching mode includes but is not limited to: switching mode at any time and switching mode once. Among them, the switching mode at any time is to switch as long as there is a network performance (i.e., the third target network path) in the first reachable network path obtained by the next round of calculation that is better than the second target network path that is currently transmitting data. The principle of this switching is: as long as a better target network path is generated, switch. The one-shot switching mode is that until the first reachable network path obtained in the last round reaches the convergence condition or reaches the preset number of rounds, the data to be transmitted is transmitted via the second target network path obtained in the current round (i.e., the first round). After all rounds are completed, the third target network path that meets the preset conditions (e.g., the optimal target network path in all rounds) is selected from the first reachable network paths obtained in all rounds for switching.
[0023] In the above-mentioned implementation of the present application, if a better path becomes available during the iteration process, the transmission path for the data to be sent can be switched at any time, or it can be switched all at once after finding the optimal path. The advantage of switching at any time is that the transmission path for the data to be transmitted always remains on the currently calculated optimal network path, improving transmission efficiency; the advantage of switching all at once is that the transmission path for the data to be transmitted does not need to be frequently changed, ensuring the stability of the transmission path. In specific application scenarios, the switching mode can be selected based on actual needs, providing flexibility.
[0024] In a possible implementation of the first aspect, the method of switching the transmission path of the data to be transmitted from the second target network path to the third target network path can be: obtaining a switching instruction, which is used to instruct the source node to wait for a preset period of time before initiating the transmission of untransmitted data packets in the data to be transmitted on the third target network path.
[0025] In the above-mentioned embodiment of the present application, since the transmission of the data to be transmitted has already begun after the first round of calculation of the first reachable network path is completed, the optimal path to the destination node corresponding to the data to be transmitted may be updated after each subsequent round of calculation is completed. Therefore, it is possible to frequently adjust the data transmission path. When the transmission path switches during the data transmission process, it may cause the data transmission to be out of order. Based on this, the embodiment of the present application provides a solution to the data out of order problem and reduces the data packet loss rate.
[0026] In a possible implementation of the first aspect, a specific implementation method for performing a partial order comparison on the m network paths based on n target factors to obtain the second advantageous network path may be: first, based on the n target factors, construct m performance arrays of the m network paths, each performance array including n elements arranged in order, one network path corresponds to one performance array, and one element corresponds to the value of one target factor on the corresponding network path. It should be noted that the value of each element can be a specific value (such as bandwidth 50, latency 15, etc.), or a specific value obtained for a value interval (such as a bandwidth interval [20,50] has multiple values, and the specific value obtained by calculating the average, median, weighted sum, normalization, etc.) is not limited in this application. For example, assuming that the n preset target factors are metric-1, metric-2, ..., metric-n (a metric can be any indicator, such as a network performance indicator such as bandwidth, latency, or packet loss rate, or any conditional constraint target, such as the total number of hops in a path not exceeding 5), and there are two network paths to the same destination node (i.e., m = 2), namely path A and path B, then two performance arrays are constructed: a performance array corresponding to path A (metric-a1, metric-a2, ..., metric-an) and a performance array corresponding to path B (metric-b1, metric-b2, ..., metric-bn), where metric-a1, metric-a2, ..., metric-an are the specific values of the corresponding target factors on path A, and metric-b1, metric-b2, ..., metric-bn are the specific values of the corresponding target factors on path B. Afterwards, the values of the elements at each corresponding position in the m performance arrays are compared. If the value of each element in the first performance array is greater than or equal to the element at the corresponding position in the second performance array, then the network path corresponding to the first performance data is determined to be the second dominant network path, and the network path corresponding to the second performance data is determined to be a non-dominant network path, where the first performance array and the second performance array are any two of the m performance arrays. For example, if each metric in the performance array of path A is better than or equal to the metric at the corresponding position in the performance array of path B, that is, metric_a1≥metric_b1, metric_a2≥metric_b2, ..., metric_an≥metric_bn, then path A is considered to be the second dominant network path, and correspondingly, path B is considered to be a non-dominant path, and path B will be removed from the first reachable network path (i.e., not recorded).
[0027] In the above-mentioned embodiments of the present application, the specific implementation method of the partial order comparison is specifically described, which is feasible.
[0028] In a possible implementation of the first aspect, the comparison result of the values of the elements at each corresponding position in the m performance arrays may also have another situation, that is, the value of the first element in the first performance array is greater than or equal to the value of the element at the corresponding position in the second performance array, and the value of the second element is less than or equal to the value of the element at the corresponding position in the second performance array. In this case, it is determined that the network path corresponding to the first performance array and the network path corresponding to the second performance array are both second-advantage network paths, wherein the first element and the second element are any two elements in the first performance array. For example, since the performance array corresponding to path A is (metric-a1, metric-a2, ..., metric-an), and the performance array corresponding to path B is (metric-b1, metric-b2, ..., metric-bn), if each metric in the performance array of path A is better than or equal to the metric at the corresponding position in the performance array of path B, that is, metric_a1 ≥ metric_b1, metric_a2 ≥ metric_b2, ..., metric_an ≥ metric_bn, then path A is considered to be the second-dominant network path (i.e., the first case). Otherwise (for example, metric_a1 ≥ metric_b1, metric_a2 < metric_b2), it is considered that path A and path B cannot be compared in terms of superiority (ignoring the network performance requirements of the specific application), and both paths need to be retained as the second-dominant paths.
[0029] In the above-mentioned embodiment of the present application, a processing method after another comparison result appears after comparison using the partial order comparison method is specifically described, so that the first reachable network path finally obtained has completeness.
[0030] In a possible implementation of the first aspect, a method of selecting a network node from the set of nodes to be added as the first node may be random selection, or it may be based on the network performance requirements of the data to be transmitted on the target device, and a network node that meets the preset performance requirements is selected from the set of nodes to be added as the first node.
[0031] In the above-mentioned embodiments of the present application, several methods of selecting the first node from the set of nodes to be added are specifically described, which are optional.
[0032] In a possible implementation of the first aspect, the implementation method of updating the first reachable network path based on the obtained first advantageous network path may be: when there are m network paths to the same destination node, only the first advantageous network path is retained in the first reachable network path.
[0033] In the above embodiment of the present application, the network path that becomes non-dominant in the comparison will be eliminated from the first reachable network path and will no longer participate in the subsequent iterative process. The computational overhead is minimized by filtering out non-dominant network paths that do not need to be recorded in real time (i.e., real-time pruning).
[0034] In a possible implementation of the first aspect, the target factor can be a network performance indicator. Since in the industry, whether it is a centralized algorithm or a distributed approximate optimal path algorithm (such as an IGP distributed path algorithm based on weighted normalization of multiple objective factors, an approximate optimal path algorithm based on a single objective factor KSP, an optimal path algorithm based on multiple single objective factors of network sharding, etc.), the calculation logic is presented in the single dimension of "seeking the optimal", such as seeking the optimal path for latency, the optimal path for packet loss, etc. In actual application scenarios, not all indicators are required to be optimal. Sometimes, some indicators are often optimized to within a certain threshold, such as the delay <300ms, bandwidth >500Mbps, and the number of hops not exceeding 4 hops. In the embodiments of the present application, such target factors that do not require the optimal solution are called conditional constraints. Therefore, in some embodiments of the present application, the target factor can also be a conditional constraint target, such as delay <300ms, bandwidth >500Mbps, and the total number of hops of the network path not exceeding a preset number of hops. Specifically, the present application does not limit the specific type of the target factor, and it can be set based on the specific application scenario and requirements. When the network performance requirements of an application include both network performance indicators and conditional constraints, the conditional constraint is converted into a specific indicator value and participates in the calculation of the first dominant path before the conditional constraint reaches the threshold. When the conditional constraint reaches the threshold, it no longer participates in the calculation of the first dominant path (i.e., the corresponding network path is directly eliminated).
[0035] In the above-mentioned embodiment of the present application, in response to the limitation that the classic IGP distributed algorithm cannot calculate the optimal path under conditional constraints, a processing method for converting conditional constraint targets into optimization indicators is designed. The parameters corresponding to the conditional constraint targets are uniformly processed according to the optimization indicators before reaching the threshold, and participate in the partial order comparison of the performance array together with other network performance targets, thereby ensuring that the optimal solution is not missed during the dominant path screening process. The embodiment of the present application can ensure the accuracy of network path calculation in multi-objective factor scenarios under any business requirements, that is, in any complex multi-objective factor scenario, the end-to-end optimal path can be correctly calculated.
[0036] In a possible implementation manner of the first aspect, the network performance indicator includes at least any one of the following: bandwidth, delay, packet loss rate, jitter, and cost.
[0037] In the above-mentioned embodiments of the present application, several commonly used forms of network performance indicators are specifically described, which have wide applicability and flexibility.
[0038] In one possible implementation of the first aspect, the target network can be an overlay network. A typical overlay network scenario is a real-time network (RTN) carrying audio and video services. The RTN will comprehensively calculate the optimal end-to-end overlay network path that meets the audio and video transmission requirements based on the different network performance factors such as latency, packet loss, and jitter of the audio or video services. Alternatively, the target network can be an underlay network. A typical underlay network scenario is a scenario where service data is transmitted between campus networks or cloud networks. In campus network scenarios, the optimal data forwarding path for the router needs to be planned based on the different network performance parameters required by different office applications (e.g., video conferencing services require low jitter and low packet loss, while desktop cloud services require high bandwidth and low latency). In cloud network scenarios, cloud service providers can provide users with different performance paths based on their VIP levels, such as a gold path with latency <10ms and packet loss <5%, a silver path with latency <15ms and packet loss <5%, and a bronze path with latency <20ms and packet loss <10%. Specifically, this application does not limit the specific type of the target network.
[0039] In the above-mentioned embodiments of the present application, the networks to which the network path calculation method provided in the embodiments of the present application is applicable are specifically described, and the method has universal applicability.
[0040] A second aspect of the present application provides a network device that implements the method of the first aspect or any possible implementation of the first aspect. The function can be implemented by hardware or by hardware executing corresponding software. The hardware or software includes one or more modules corresponding to the above-mentioned functions.
[0041] A third aspect of an embodiment of the present application provides a computer device, which may include a memory, a processor, and a bus system, wherein the memory is used to store programs, and the processor is used to call the programs stored in the memory to execute the method of the first aspect of the embodiment of the present application or any possible implementation method of the first aspect.
[0042] A fourth aspect of an embodiment of the present application provides a computer-readable storage medium, which stores instructions. When the computer-readable storage medium is run on a computer, the computer can execute the method of the above-mentioned first aspect or any possible implementation of the first aspect.
[0043] A fifth aspect of the embodiments of the present application provides a computer program, which, when executed on a computer, enables the computer to execute the method of the first aspect or any possible implementation of the first aspect.
[0044] In a sixth aspect of an embodiment of the present application, a chip is provided, which chip (e.g., a CPU) includes at least one processor and at least one interface circuit, the interface circuit is coupled to the processor, the at least one interface circuit is used to perform a transceiver function, and send instructions to the at least one processor, the at least one processor is used to run a computer program or instruction, which has the function of implementing the method of the first aspect or any possible implementation of the first aspect, the function can be implemented by hardware, or by software, or by a combination of hardware and software, the hardware or software including one or more modules corresponding to the above functions. In addition, the interface circuit is used to communicate with other modules outside the chip. BRIEF DESCRIPTION OF THE DRAWINGS
[0045] FIG1 is a schematic diagram of an example of calculating an optimal network path using the classic IGP routing calculation principle provided in an embodiment of the present application;
[0046] FIG2 is a diagram of an application scenario of the Internet Overlay technology provided in an embodiment of the present application;
[0047] FIG3 is a schematic diagram of an architecture of a centralized algorithm provided in an embodiment of the present application;
[0048] FIG4 is a schematic diagram showing a principle of calculating a network path using a classic IGP distributed algorithm according to an embodiment of the present application;
[0049] FIG5 is a schematic diagram showing a principle of a method for calculating a network path according to an embodiment of the present application;
[0050] FIG6 is a schematic diagram showing the effect of introducing non-optimal path records on computational complexity according to an embodiment of the present application;
[0051] FIG7 is a flow chart of a method for calculating a network path according to an embodiment of the present application;
[0052] FIG8 is a schematic diagram of an example of determining a first reachable network path according to an embodiment of the present application;
[0053] FIG9 is a schematic diagram of an example of determining a second reachable network path according to an embodiment of the present application;
[0054] FIG10 is a schematic diagram of an example of a method for performing partial order comparison provided in an embodiment of the present application;
[0055] FIG11 is a schematic diagram showing that the number of second advantageous network paths in a large-scale network gradually increases with the iteration process according to an embodiment of the present application;
[0056] FIG12 is a schematic diagram of an example of performing iterative calculations according to an embodiment of the present application;
[0057] FIG13 is a schematic diagram of a design of a multi-objective factor routing table provided in an embodiment of the present application;
[0058] FIG14 is a schematic diagram of a calculation process for performing multiple rounds of random selection to calculate a target path according to an embodiment of the present application;
[0059] FIG15 is a schematic diagram of a method for resolving data transmission disorder caused by path switching according to an embodiment of the present application;
[0060] FIG16 is a schematic diagram of an application scenario of audio and video service routing in an RTN network provided by an embodiment of the present application;
[0061] FIG17 is a schematic diagram comparing the calculation accuracy and performance of the present solution and the existing multi-path solution of the consumer cloud RTN under different RTN network scales provided by an embodiment of the present application;
[0062] FIG18 is a schematic diagram comparing the calculation results of the present application solution and the existing solution provided in an embodiment of the present application;
[0063] FIG19 is a schematic diagram of a scenario in which the calculation method of Example 1, provided in an embodiment of the present application, exhibits inaccurate calculations when there are conditional constraints on the target;
[0064] FIG20 is a schematic diagram of a scenario of a calculation method of Example 2 when there are conditional restrictions on the target provided by an embodiment of the present application;
[0065] FIG21 is a schematic diagram of the structure of a network device provided in an embodiment of the present application;
[0066] Figure 22 is another structural diagram of the network device provided in an embodiment of the present application. DETAILED DESCRIPTION
[0067] The first aspect of the present application provides a method for calculating a network path, which is used to address the problem that the classic IGP distributed algorithm cannot solve the optimal network path for multi-objective factors. A method for calculating the distributed multi-objective factors of an advantageous network path is designed that can calculate any business needs, thereby improving the ability of resource allocation optimization and resource control optimization.
[0068] The terms "first", "second", etc. in the specification and claims of the present application and the above-mentioned drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequential order. It should be understood that the terms used in this way can be interchangeable under appropriate circumstances, and this is merely a way of distinguishing the objects of the same attributes when describing them in the embodiments of the present application. In addition, the terms "including" and "having" and any of their variations are intended to cover non-exclusive inclusions, so that the process, method, system, product or equipment comprising a series of units need not be limited to those units, but may include other units that are not clearly listed or inherent to these processes, methods, products or equipment.
[0069] The embodiments of this application involve a lot of network-related knowledge. In order to better understand the solutions of the embodiments of this application, the following first introduces the relevant terms and concepts that may be involved in the embodiments of this application. It should be understood that the interpretation of related concepts may be limited by the specific circumstances of the embodiments of this application, but it does not mean that this application is limited to only such specific circumstances. The specific circumstances of different embodiments may also vary, which is not specifically limited here.
[0070] (1) Underlay Network
[0071] The underlay network refers to the underlying forwarding architecture of current data center networks. As long as any two points on the data center network are reachable, it serves as the physical foundation. Improvements to the underlay network can be made through technological improvements to physical network equipment, expansion of the number of devices, and bandwidth scale. It encompasses all existing traditional network technologies.
[0072] (2) Overlay Network
[0073] Because the Internet is comprised of multiple carriers, Internet service providers (ISPs), and autonomous domains, and the interconnections between these domains are characterized by complex business relationships, Internet transmission paths are often not the shortest paths in the network. For example, transmission between two Asian nodes may be routed through Europe, increasing end-to-end latency. Furthermore, because Internet routing is unaware of path performance, failures or congestion along the path are often difficult to avoid, and can require lengthy convergence times.
[0074] To address the issue of suboptimal internet paths, the industry has proposed internet overlay technology. In network technology, overlay refers to a virtualization technology model superimposed on network architecture. Its general framework enables application hosting on the network without major modifications to the underlying network, while maintaining separation from other network services. It primarily relies on IP-based network technology. Overlay technology builds a virtual network on top of the existing physical network, with upper-layer applications interacting exclusively with the virtual network.
[0075] Specifically, by placing forwarding units (also known as network nodes) in data centers in different regions of the Internet, interconnecting and scheduling each other, a new virtual overlay network (i.e., Overlay network) is constructed on the basis of the existing public Internet (i.e., Underlay network). The network nodes deployed in the middle can be called forwarding relays (which can be denoted as PoPs). For ease of understanding, please refer to Figure 2, which is an application scenario diagram of Internet Overlay technology. In Figure 2, the default link between the source device and the destination device is the network path shown by the dotted line. Assuming that the cross-domain link between domain A and domain B is currently facing severe congestion, then routing according to this default link will result in end-to-end performance degradation. Through Overlay technology, PoPs are deployed in domains A, B, and C respectively. Among them, the PoP deployed in domain A is called PoP1, the PoP deployed in domain B is called PoP2, and the PoP deployed in domain C is called PoP3. The overlay backbone network controls the route from PoP1 to PoP3 as PoP1→PoP2→PoP3, effectively bypassing the congested link between domains A and B and improving overall end-to-end performance.
[0076] (3) Real-time network (RTN)
[0077] A real-time network is a specific type of overlay network and is a completely new architecture for real-time transmission of audio and video. Similar to a live content delivery network (CDN), a real-time network is designed for scenarios where real-time audio and video are strongly required. In principle, global end-to-end latency can be controlled to less than 300ms through a real-time network. In the embodiments of this application, the network nodes included in the real-time network may also be referred to as RTN nodes.
[0078] Because classic IGP distributed algorithms cannot calculate optimal paths for multiple factors, the industry generally adopts centralized algorithms when multi-factor network path calculation is required. Specifically, a centralized architecture aggregates all path information and the network performance requirements of different network applications into a centralized network controller (referred to as a centralized controller). The centralized controller then calculates the optimal path based on this full path information, meeting the network performance requirements of each application. The typical centralized multi-factor network path calculation architecture is shown in Figure 3. After the centralized controller calculates the optimal end-to-end path, it distributes the path to each network node along the path to complete data forwarding.
[0079] Because centralized algorithms record all path information, unlike classic IGP distributed algorithms that only record locally optimal paths, they traverse a larger space and can accurately search the entire path set to find the optimal path that meets the application's network performance requirements. This eliminates the inaccuracy issues often seen with classic IGP distributed algorithms in scenarios with multiple factors. However, centralized algorithms suffer from poor real-time and scalability. Centralized algorithms typically have significant computational overhead, and as network size increases, the time complexity of accurately determining the optimal path that meets application SLAs becomes unacceptable. Field-measured data shows that in a network with 2K nodes and 10K links, a centralized algorithm can complete the calculation of the optimal path based on multiple factors to meet specific application SLA requirements in approximately one day, while the real-time requirements of the service are sub-second. Therefore, due to these real-time and scalability limitations, centralized algorithms are not suitable for large-scale network environments or real-time network services.
[0080] In addition to centralized algorithms, the industry also employs distributed approximate optimal path algorithms. These include: IGP distributed path algorithms based on weighted normalization of multiple objective factors, approximate optimal path algorithms based on a single objective factor, KSP, and optimal path algorithms based on multiple single objective factors using network sharding. For example, the IGP distributed path algorithm based on weighted normalization of multiple objective factors uses a weighted approach to normalize multiple objective factors into a single composite parameter, and then calculates the IGP single-factor optimal path based on this composite parameter. With this weighted normalization approach, the weights assigned to different objective factors ultimately affect the calculation results. Therefore, these methods focus on designing the weight distribution of different objective factors so that the calculation results meet the constraints. While this weighted normalization approach can comprehensively consider the impact of each objective factor on application performance—that is, it can use weighting to amplify the dominance of key objective factors in path calculation—it cannot precisely match parameters according to application logic (for example, it cannot accurately calculate paths with latency <300ms, bandwidth >500Mbps, and jitter <50ms; it can only maximize the weights of latency, bandwidth, or jitter to approximately meet the application's network performance requirements). It can only perform relatively coarse-grained weighting based on the importance of each objective factor corresponding to the application's network performance requirements, failing to meet the precise SLA requirements of network applications. To address this issue, a new class of distributed methods proposes first calculating the top K optimal paths based on a single factor to form a candidate solution set. Then, from this candidate solution set, the optimal path is selected based on the application's network performance requirements, and this path is approximately equivalent to the global optimal path (i.e., an approximate optimal path algorithm based on the single objective factor, KSP). While the top K solution space for a single objective factor calculated by the KSP algorithm can accurately calculate the optimal path based on application network performance requirements, paths with optimal performance based on multiple objective factors may not appear in the top K solution space for a single objective factor. (For example, using latency as the objective factor, the KSP algorithm calculates the top 10 latency paths, but these top 10 latency paths may have poor performance in other objective factors such as packet loss, bandwidth, or cost. In this case, the solution space for the top 10 latency paths will miss the globally optimal path.) Optimal path algorithms based on multiple single objective factors in network slicing combine logical network slicing with physical links to calculate single objective factor routes based on different objective factors in different network slices. A representative solution of this type is the Flex-Algo solution jointly developed by Cisco and Juniper. The Flex-Algo solution still uses the standard IGP algorithm, but it allows switching between different objective factors within the same network topology. For example, using latency as the objective factor and bandwidth as the objective factor may result in different end-to-end optimal routes due to different link performance within the same network topology.By combining logical sharding with physical networks, different optimal end-to-end routes can be calculated for different applications within different network shards, based on bandwidth or latency factors, thereby providing differentiated quality of service. However, as previously analyzed, each application typically has performance requirements for multiple target factors (for example, latency, bandwidth, jitter, and other metrics). Therefore, each network shard often requires path calculation based on multiple target factors. The Flex-Algo solution cannot achieve this function.
[0081] In summary, the main problem with existing distributed approximate optimal path algorithms is their inability to accurately calculate multi-objective optimal paths that meet application business requirements. This problem arises because existing distributed approximate optimal path algorithms do not fundamentally alter the IGP algorithm. Therefore, by introducing additional objective factors without changing the algorithm, they typically only calculate approximate optimal paths, often with inaccurate results. Consequently, they are unable to find the global optimal path in multi-objective scenarios where "local optimal path + local optimal path + ... + local optimal path ≠ global optimal path."
[0082] Therefore, to solve the above problems, this application improves the traditional distributed IGP algorithm. By recording additional locally advantageous network paths, it can accurately calculate the end-to-end optimal path that meets the network performance requirements of the application's multiple objective factors with minimal computational overhead in the multi-objective factor scenario of "local optimal path + local optimal path + ... + local optimal path ≠ global optimal path", breaking through the limitations of all existing solutions and improving the ability to optimize resource allocation and resource control.
[0083] The embodiments of the present application are described below in conjunction with the accompanying drawings. Those skilled in the art will appreciate that, with the development of technology and the emergence of new scenarios, the technical solutions provided in the embodiments of the present application are also applicable to similar technical problems.
[0084] First, the core principle of the network path calculation method provided by the embodiment of this application is explained. Taking the dual-objective factor scenario of video services as an example, it is assumed that the network performance requirements of video services are: first, select the path with the largest bandwidth. When multiple paths with equal bandwidth are found, select the path with the lowest latency. This is a "bandwidth + latency" dual-objective optimization scenario. This application can use a performance array containing two elements to identify the bandwidth and latency of each hop link. For example, (20,10) indicates that the link bandwidth is 20 and the latency is 10 (ignoring the dimension here).
[0085] As shown in Figure 4, in a traditional IGP distributed algorithm, the source node (here, assumed to be node u in Figure 4) adds a reachable node in Figure 4 to the optimal path tree during each iteration and updates the current optimal path and path performance from the source node to all other destination nodes (w, v, x) within the current optimal path tree. For example, during the third iteration, nodes w and v are added to the shortest path tree. The routing table entry now records the optimal path from node u to node w as u→w, with a performance of (20, 1), meaning a bandwidth of 20 and a latency of 1. The routing table entry also records the optimal path from node u to node v as u→w→v, with a performance of (20, 5). It should be noted here that there is another path u→v from node u to node v, with a performance of (10,2). However, according to the network performance requirements of the application, the path with the largest bandwidth is preferred. Therefore, the classic IGP distributed algorithm believes that the performance of the path u→w→v (20,5) is better than the performance of the path u→v (10,2) because the primary performance indicator of the path u→w→v is larger, "bandwidth". However, the classic IGP distributed algorithm only records the optimal path to each destination node. Therefore, the routing table entry for node u to node v only records the path u→w→v with a performance of (20,5), and does not record the path u→v with a performance of (10,2). When further calculating the optimal path from node u to node x, node u finds that it can reach node x by transiting through node w or node v. The classic IGP distributed algorithm will compare the performance of the end-to-end path from node u to node x in the two cases in turn and record the path with the best performance. The specific process is as follows:
[0086] 1) If the performance of the optimal path u→w from node u to node w is (20, 1), and the performance of the direct path from node w to node x is (5, 1), the resulting end-to-end path u→w→x is (20, 1) + (5, 1) = (5, 2). In other words, the bandwidth is the minimum of the two hops, min(20, 5) = 5, and the latency is the sum of the two hops, add(1, 1) = 2.
[0087] 2) If the optimal path u→w→v from node u to node v is transferred through node v, the performance is (20, 5). Adding the performance of the direct path from node v to node x is (10, 3), the resulting end-to-end path u→w→v→x has a performance of (20, 5) + (10, 3) = (10, 8).
[0088] In this case, source node u compares the performance of the two end-to-end paths and finds that the path through node v has a performance of (10, 8) better than the path through node w (5, 2) because the primary optimization objective, "bandwidth," is greater. Therefore, using the classic IGP distributed algorithm, the optimal path from source node u to destination node x is calculated as: u→w→v→x, with a performance of (10, 8).
[0089] However, it can be found that when transiting through node v, if node u to node v does not take the locally optimal path u→w→v, but instead takes the locally non-optimal path u→v, and then takes the direct path from node v to node x, the performance of the resulting end-to-end path u→v→x is (10,2)+(10,3)=(10,5). This is better than the performance of the IGP optimal path u→w→v→x (10,8), which is obtained by taking the optimal path at each hop in the classic IGP distributed algorithm. This is because the bandwidth of the path u→v→x is the same as the end-to-end path u→w→v→x calculated by the classic IGP distributed algorithm, but the latency is smaller. However, during the iterative calculation process of the classic IGP distributed algorithm, when calculating the optimal path to the destination node v, the performance of the path u→w→v (20, 5) is better than the performance of the path u→v (10, 2) (because the bandwidth of the path u→w→v is larger). Therefore, the local non-optimal path u→v is not recorded in the classic IGP routing table, and therefore the end-to-end optimal path u→v→x cannot be calculated based on the local non-optimal path u→v.
[0090] To address this issue, the embodiment of the present application introduces the recording of local non-optimal paths on the basis of the classic IGP distributed algorithm. As shown in Figure 5, after the node w is added to the optimal path tree in the second iteration, the embodiment of the present application will calculate that there are two paths to the node v, namely the path u→w→v and the path u→v, where the performance of the path u→w→v is (20,5) and the performance of the path u→v is (10,2). However, unlike the classic IGP distributed algorithm, in the embodiment of the present application, although the performance of the path u→v is not as good as the performance of the path u→w→v according to the network performance requirements of the business, and it is not the optimal path, the path will also be recorded in the embodiment of the present application. In this way, when calculating the optimal path to node x in the next iteration, the embodiment of the present application can calculate the true end-to-end optimal path u→v→x based on the local non-optimal path u→v.
[0091] While the inclusion of non-optimal paths addresses the accuracy of optimal paths in multi-objective scenarios, it may also introduce a new problem: reduced path calculation efficiency. Introducing non-optimal path records requires iterative calculations to traverse and compare more paths. As shown in Figure 6, consider a three-hop network topology with N1 paths from the source node to intermediate node 1, N2 paths from intermediate node 1 to intermediate node 2, and N3 paths from intermediate node 2 to the destination node. Using a classic IGP distributed algorithm, only the local optimal path is calculated and recorded for each hop, and the hop-by-hop optimal paths are concatenated into the end-to-end optimal path. This computational complexity is O(N1+N2+N3), minimizing computational overhead but potentially leading to errors in the optimal path. As more non-optimal paths are introduced, the path calculation becomes increasingly accurate. In the most extreme case, when all non-optimal paths (i.e., the full set of paths) are recorded, the complexity of calculating the global optimal path increases to O(N1*N2*N3). While this algorithm can still accurately traverse the optimal path in end-to-end multi-objective scenarios, the computational overhead can be significant. Based on this, another goal of the embodiment of the present application is to find the minimum computational overhead that can ensure accurate path calculation in a multi-objective factor scenario, that is, to record the minimum number of local non-optimal paths to ensure the accuracy of the multi-objective factor scenario.
[0092] To address the problem of optimizing computational complexity, the embodiment of the present application designs a partial order comparison method based on a performance array, which can filter out local non-optimal paths that do not need to be recorded in real time during the iterative calculation process, and minimize computational overhead through real-time pruning.
[0093] After introducing the principle of the method for calculating the network path of the embodiment of the present application, the specific process of the method for calculating the network path provided by the embodiment of the present application is described below. It should be noted that, in an actual network system (i.e., the target network), each network node can independently execute the method for calculating the network path described in the embodiment of the present application to calculate the advantageous network path of multiple target factors with the network node as the source node and reaching the final destination node, and form a local routing table based on the obtained advantageous network path. In this case, the subject that executes the following steps 701 to 707 is the source node.
[0094] However, it should be noted that in other embodiments of the present application, if there is a unified performance test network node (also referred to as a control node, management node, etc.) in the network system, the network path calculation method described in the embodiments of the present application can also be executed by the control node to calculate the optimal network path of multiple target factors with a certain network node as the source node and reaching the final destination node. The calculated optimal network path is then distributed to other network nodes in the network system to form a local routing table. In this case, the main body executing the following steps 701 to the subsequent steps 707 is the control node.
[0095] Therefore, in an embodiment of the present application, the execution subject of the network path calculation method provided in the embodiment of the present application can be the current source node or the control node with unified management in the target network. According to the situations of different execution subjects, the specific operations of different execution subjects in each step will be introduced respectively.
[0096] Specifically, if the execution subject is a source node, the source node and other network nodes in the target network can obtain network topology and link performance information (such as latency, bandwidth, packet loss rate, etc.) through the notification of IGP information. When an application on a target device connected to the source node needs to transmit data (i.e., the data to be sent as described above), the application on the target device formalizes its own application requirements (such as network performance requirements) in a data packet, and then passes the data packet to the source node. After receiving and parsing the data packet, the source node calculates the network path based on the application requirements and the network topology and link performance information. If the execution subject is a control node, the network topology and link performance information is notified to other network nodes in the target network through the control node. When an application on a target device connected to the source node needs to transmit data, the application on the target device formalizes its own application requirements in a data packet, and then passes the data packet to the control node. After receiving and parsing the data packet, the control node calculates the network path based on the application requirements and the network topology and link performance information, and then notifies the calculated target path to other network nodes in the target network.
[0097] Please refer to FIG. 7 for details. FIG. 7 is a flowchart of a method for calculating a network path provided in an embodiment of the present application, which may include the following steps:
[0098] 701. Determine a first reachable network path starting from a source node and ending at a node directly connected to the source node.
[0099] First, determine a reachable network path with the source node as the starting point and the directly connected node of the source node as the end point. The reachable network path can be called the first reachable network path (which can be one or more). Among them, the directly connected node refers to a node that is directly connected to the source node and is not relayed by other nodes in the middle. It should be noted that the reachable network path not only includes the network nodes involved in the reachable network path, but also includes the overall network performance of the entire reachable network path (such as the delay, bandwidth, packet loss rate, etc. of the reachable network path). Among them, the source node is a network node in the network (which can be called the target network) to which the target device (i.e., the device with data to be sent) is connected. The network node can be a router or a server in the target network, etc., which is specifically determined by the type of the target network involved, and this application does not limit this. For example, assuming that the target network is an Overlay network, a device 1 (such as a mobile phone, a host, etc.) is connected to the network node A in the Overlay network nearby. For the device 1, the network node A is the source node. Similarly, if another device 2 is connected to a nearby network node B in the overlay network, then for that device 2, network node B is the source node. In other words, each network node in the target network can become a source node for a specific target device, where the source node is for a specific device to send data. For ease of explanation, in the embodiments of this application, only a specific target device and the network node it is connected to are used as the source node.
[0100] It should be noted that in some embodiments of the present application, the target network may be an overlay network. A typical overlay network scenario is a real-time network (RTN) that carries audio and video services. The RTN will comprehensively calculate the optimal end-to-end overlay network path that meets the audio and video transmission requirements based on the different requirements of the audio or video services for network performance factors such as latency, packet loss, and jitter. In addition, the target network may also be an underlay network. A typical underlay network scenario is a scenario where service data is transmitted between campus networks or cloud networks. In campus network scenarios, it is necessary to plan the optimal data forwarding path of the router based on the different network performance parameters required by different office applications (e.g., video conferencing services require low jitter and low packet loss, desktop cloud services require large bandwidth and low latency, etc.). In cloud network scenarios, cloud service providers can provide users with different performance quality paths based on the VIP level of different cloud users, such as a gold path with a latency of <10ms and a packet loss of <5%, a silver path with a latency of <15ms and a packet loss of <5%, and a bronze path with a latency of <20ms and a packet loss of <10%. Specifically, this application does not limit the specific type of the target network.
[0101] To facilitate understanding of step 701, the following example is provided: Please refer to Figure 8 for details. Assume that the network topology determined in the target network includes a total of 10 network nodes, namely nodes 1-9, and assume that the source node is node 0 and the other network nodes are nodes 1-9. Assume that based on the network topology, it can be known that there are 4 adjacent nodes (i.e., directly connected nodes) that can be reached by node 0, namely node 1, node 2, node 3, and node 4. Therefore, there are 4 first reachable network paths with the source node as the starting point and the directly connected node of the source node as the end point, namely: node 0→node 1, node 0→node 2, node 0→node 3, node 0→node 4. For the sake of convenience, these 4 first reachable network paths can be recorded as {0→1, 0→2, 0→3, 0→4}.
[0102] Since the execution subject of step 701 can be either the source node or the control node, the specific execution process of step 701 will be slightly different depending on the execution subject, which is described below:
[0103] (1) The execution subject is the source node
[0104] When the execution subject is a source node, a specific implementation method for determining the first reachable network path starting from the source node and ending at a directly connected node of the source node may be: the source node and other network nodes in the target network may mutually obtain network topology and link performance information (e.g., latency, bandwidth, packet loss rate, jitter, etc.) through IGP information advertisements. The source node then determines the first reachable network path starting from the source node and ending at a directly connected node of the source node based on the network topology and link performance information.
[0105] (2) The execution subject is the control node
[0106] When the execution subject is a control node, a specific implementation method for determining the first reachable network path starting from the source node and ending at a directly connected node of the source node may be: the control node obtains network topology and link performance information. For example, the control node may collect performance measurement data of the target network at regular intervals (e.g., 1 minute) and use the collected performance measurement data to determine the connectivity between each pair of network nodes in the target network, thereby integrating the overall network topology. In addition, the control node may also obtain link performance information based on the collected performance measurement data. Thereafter, based on the network topology and link performance information, the control node determines the first reachable network path starting from the source node and ending at a directly connected node of the source node.
[0107] It should be noted that there are multiple ways to implement the determination of the target network's network topology. As an example, when the network performance between two network nodes, such as latency, bandwidth, jitter, and packet loss, reaches or fails to reach a preset threshold, the two network nodes are considered to be disconnected. For example, when the packet loss rate between two network nodes reaches a preset threshold (e.g., 80%), the two network nodes are considered to be disconnected. As another example, it is also possible to customize whether two network nodes in the target network are connected. For example, when maintenance personnel need to disconnect the connection between network node A and network node B based on other restrictions, the maintenance personnel can customize the setting of network node A and network node B to be disconnected.
[0108] 702 . Determine whether the node set to be added is an empty set. If not, execute step 703 ; if so, execute step 707 .
[0109] Next, a determination is made as to whether the node set to be added is an empty set, where the node set to be added is the set of network nodes in the target network excluding the source node. If it is determined that the node set to be added is not an empty set (i.e., there are network nodes in the node set to be added), step 703 is executed; if it is determined that the node set to be added is an empty set (i.e., there are no network nodes in the node set to be added), step 707 is executed.
[0110] To facilitate understanding of step 702, the following is still illustrated using FIG8 as an example: Assuming that the network topology determined in the target network includes a total of 10 network nodes, namely nodes 0-9, and assuming that the source node is node 0, the node set to be added includes nodes 1-9. For ease of explanation, the node set to be added can be recorded as {1,2,3,4,5,6,7,8,9}. When executing step 702, it can be determined that the node set to be added is not an empty set (because there are still 9 network nodes 1-9). In this case, the process will jump to step 703 to execute subsequent steps 703-706. It should be noted that step 706 is a repetitive process that is repeated until the node set to be added is empty. If the node set to be added is empty, step 707 is executed.
[0111] It should be noted that no matter the execution subject is a source node or a control node, the process of executing step 702 by the execution subject is similar, so they will not be described separately here.
[0112] 703. Select a first node from the set of nodes to be added, and obtain a second reachable network path with the source node as the starting point and the first node as the relay.
[0113] When it is determined that the set of nodes to be added is not empty, a node (which may be referred to as the first node) is selected from the set of nodes to be added to obtain a reachable network path with the source node as the starting point and the first node as the relay. The reachable network path may be referred to as the second reachable network path.
[0114] It should be noted that, in some embodiments of the present application, the intermediate nodes of the second reachable network path do not include other network nodes in the set to be added except the first node. In addition, the intermediate nodes of the second reachable network path do not include the first network node (i.e., source node) and the last destination node (i.e., destination node) of the second reachable network path, but refer to any network node in the "middle". For example, assuming that the second reachable network path is A→B→C→…→X, where A, B, C,…, X are all network nodes, then the intermediate nodes of the second reachable network path refer to any network node in B, C,…, but do not include the source node A and the destination node X.
[0115] It should also be noted that, in some other embodiments of the present application, the method of selecting a network node from the set of nodes to be added as the first node can be a random selection, or it can be based on the network performance requirements of the data to be transmitted on the target device, and a network node that meets the preset performance requirements is selected from the set of nodes to be added as the first node.
[0116] To facilitate understanding of the above process of selecting the first node, Figure 8 is still used as an example for explanation below: Assume that the network performance requirement for the data to be transmitted is: give priority to the path with the lowest latency. When multiple paths have the lowest latency at the same time, give priority to the path with the lowest packet loss rate (that is, the dual objective factors of latency + packet loss rate). In some embodiments of the present application, a network node can be randomly selected from the set of nodes to be added {1, 2, 3, 4, 5, 6, 7, 8, 9} as the first node. The advantage of this selection method is that no additional factors need to be considered. In other embodiments of the present application, based on the network performance requirements of the data to be transmitted (i.e., the path with the lowest latency is prioritized, and when multiple paths have the lowest latency at the same time, the path with the lowest packet loss rate is prioritized), a network node with the lowest latency in the expected local path with node 0 (assuming that all network nodes in the target network are connected to the source node, so it is called the "expected local path" rather than the actual local path) can be selected from the set of nodes to be added {1, 2, 3, 4, 5, 6, 7, 8, 9} as the first node. Assuming that the network node with the lowest latency in the expected local path is node 3, the first node selected is node 3. The advantage of this selection method is that a relatively more advantageous network node can be pre-selected to reduce the number of subsequently obtained advantageous network paths and the number of comparisons, thereby improving iteration efficiency.
[0117] To facilitate understanding of step 703, FIG9 is used as an example below: Assuming that the network topology determined in the target network includes a total of 10 network nodes, namely nodes 1-9, and assuming that the source node is node 0, the set of nodes to be added includes nodes 1-9. For ease of explanation, the set of nodes to be added can be recorded as {1,2,3,4,5,6,7,8,9}. Assuming that the first node selected from the set of nodes to be added is node 3, in the existing network topology, a second reachable network path with node 0 as the starting point and node 3 as the relay is found, and the intermediate nodes of this second reachable network path cannot include nodes {1,2,4,5,6,7,8,9}. As shown in FIG9, there are 5 second reachable network paths with node 0 as the starting point and node 3 as the relay, namely {0→3→1, 0→3→2, 0→3→7, 0→3→8, 0→3→9}.
[0118] It should be noted that no matter whether the execution subject is a source node or a control node, the process of the execution subject executing step 703 is similar, so it will not be explained separately here.
[0119] 704. If there are m network paths reaching the same destination node in the target path, then compare the m network paths based on n target factors to obtain a first dominant network path, and update the target path based on the first dominant network path to obtain an updated target path, where the target path includes a first reachable network path and a second reachable network path, where m ≥ 2 and n ≥ 2.
[0120] In an embodiment of the present application, all reachable network paths included in the first reachable network path and the second reachable network path can be referred to as target paths (one or more). For example, as shown in Figure 9, there are 4 first reachable network paths, namely {0→1, 0→2, 0→3, 0→4}, and there are 5 second reachable network paths, namely {0→3→1, 0→3→2, 0→3→7, 0→3→8, 0→3→9}. The resulting target paths are 4+5=9, namely {0→1, 0→3→1, 0→2, 0→3→2, 0→3, 0→4, 0→3→7, 0→3→8, 0→3→9}.
[0121] Afterwards, determine whether there are m (m≥2) network paths reaching the same destination node in the target path. If so, compare these m network paths based on n (n≥2) preset target factors to obtain a dominant network path (one or more), which can be called a first dominant network path. The target path is updated based on the first dominant network path to obtain an updated target path.
[0122] It should be noted that in some embodiments of the present application, the target factor can be a network performance indicator, such as bandwidth, delay, packet loss rate, jitter, cost, etc. It should be noted here that, whether it is a centralized algorithm or a distributed approximate optimal path algorithm (such as the above-mentioned IGP distributed path algorithm based on weighted normalization of multiple objective factors, the approximate optimal path algorithm based on the single objective factor KSP, the optimal path algorithm based on multiple single objective factors of network sharding, etc.), the calculation logic is presented in the single dimension of "seeking the best", such as seeking the optimal path for delay, the optimal path for packet loss, etc. In actual application scenarios, not all indicators are required to be optimal. Sometimes, some indicators can be optimized to within a certain threshold, such as the delay <300ms, bandwidth >500Mbps, and the number of hops not exceeding 4. In the embodiments of the present application, such target factors that do not require the optimal solution are called conditional constraints. Therefore, in some embodiments of the present application, the target factor may also be a conditional target, such as latency < 300ms, bandwidth > 500Mbps, the total number of hops in the network path not exceeding a preset number of hops, etc. Specifically, the present application does not limit the specific type of the target factor, which can be set based on the specific application scenario and requirements. In an embodiment of the present application, when the network performance requirements of the application include both network performance indicators and conditional targets, then before the conditional target reaches the threshold, the conditional target is converted into a specific indicator value and participates in the calculation of the first advantageous path; when the conditional target reaches the threshold, it no longer participates in the calculation of the first advantageous path (i.e., the corresponding network path is directly eliminated).
[0123] To facilitate understanding of step 704, Figure 9 is used as an example. As can be seen from the above description, nine target paths are obtained: {0→1, 0→3→1, 0→2, 0→3→2, 0→3, 0→4, 0→3→7, 0→3→8, 0→3→9}. Of these nine target paths, there are two groups of at least two network paths with node 0 as the source node and reaching the same destination node. One group consists of two network paths reaching the same destination node 1: {0→1, 0→3→1}. The other group consists of two network paths reaching the same destination node 2: {0→2, 0→3→2}. Taking the first group of network paths {0→1, 0→3→1} as an example, these two network paths {0→1, 0→3→1} are compared based on n (n≥2) preset target factors to obtain one or more first dominant network paths corresponding to the first group of network paths, where the number of first dominant network paths is less than or equal to the number of network paths in the first group. Similarly, for the second group of network paths {0→2, 0→3→2}, these two network paths {0→2, 0→3→2} are also compared based on the n preset target factors to obtain one or more first dominant network paths corresponding to the second group of network paths, where the number of first dominant network paths is less than or equal to the number of network paths in the second group. Finally, the target path is updated based on the obtained first dominant network paths corresponding to each group, obtaining an updated target path.
[0124] It should be noted that, in some embodiments of the present application, the implementation method of updating the target path based on the obtained first dominant network path can be: for the case where there are m network paths to the same destination node, only the first dominant network path is retained in the target path, that is, the network path that becomes non-dominant in the comparison will be eliminated from the target path and will no longer participate in the subsequent iterative process, thereby minimizing the computational overhead by filtering out non-dominant network paths that do not need to be recorded in real time (i.e., real-time pruning).
[0125] To facilitate understanding of the above real-time pruning process, let's continue with the above example: assuming that the first dominant network path obtained after comparing the first set of network paths is {0→1}, and the first dominant network path obtained after comparing the second set of network paths is {0→3→2}, then the paths {0→3→1, 0→2} are deleted from these nine target paths. The updated target paths are then seven, namely {0→1, 0→3→2, 0→3, 0→4, 0→3→7, 0→3→8, 0→3→9}.
[0126] It should be noted that in some embodiments of the present application, if there are m network paths reaching the same destination node in the target path, then the m network paths are compared based on n target factors (here ignoring the specific network performance requirements of the application), and the specific implementation method for obtaining the first advantageous network path can be: based on the n target factors, the m network paths are partially compared, and the network paths that are more advantageous after the partial order comparison or the network paths that cannot be compared are both regarded as advantageous network paths. The advantageous network path obtained at this time can be called the second advantageous network path. The following is an introduction to a specific implementation method of how to perform partial order comparison:
[0127] First, based on the n target factors, m performance arrays of the m network paths are constructed. Each performance array includes n elements arranged in order. One network path corresponds to one performance array, and one element corresponds to the specific value of a target factor on the corresponding network path. For example, assume that there are n preset target factors, namely metric-1, metric-2, ..., metric-n (a metric can be any network performance indicator, such as bandwidth, latency, packet loss rate, or any conditional constraint, such as a total path hop count of no more than 5). Furthermore, there are two network paths (i.e., m = 2) to the same destination node, namely path A and path B. Two performance arrays are constructed: the performance array corresponding to path A (metric-a1, metric-a2, ..., metric-an) and the performance array corresponding to path B (metric-b1, metric-b2, ..., metric-bn). Metric-a1, metric-a2, ..., metric-an are the specific values of the corresponding target factors on path A, and metric-b1, metric-b2, ..., metric-bn are the specific values of the corresponding target factors on path B. Subsequently, the values of the elements at each corresponding position in these m performance arrays are compared, and the comparison results may be as follows:
[0128] (1) The value of each element in the first performance array is greater than or equal to the element at the corresponding position in the second performance array.
[0129] In this case, the network path corresponding to the first performance data is determined to be the second dominant network path, and the network path corresponding to the second performance data is determined to be the non-dominant network path, wherein the first performance array and the second performance array are any two of the m performance arrays.
[0130] For ease of understanding, let's continue with the above example: since the performance array corresponding to path A is (metric-a1, metric-a2, ..., metric-an), and the performance array corresponding to path B is (metric-b1, metric-b2, ..., metric-bn), if each metric in path A's performance array is better than or equal to the metric at the corresponding position in path B's performance array, that is, metric_a1 ≥ metric_b1, metric_a2 ≥ metric_b2, ..., metric_an ≥ metric_bn, then path A is considered the second dominant network path. Correspondingly, path B is considered a non-dominant path and is removed from the target path (i.e., not recorded).
[0131] (2) There exists a first element in the first performance array whose value is greater than or equal to the value of the element at the corresponding position in the second performance array, and a second element whose value is less than or equal to the value of the element at the corresponding position in the second performance array.
[0132] In this case, it is determined that the network path corresponding to the first performance array and the network path corresponding to the second performance array are both second advantageous network paths, wherein the first element and the second element are any two elements in the first performance array.
[0133] For ease of understanding, let's use the above example again for illustration: since the performance array corresponding to path A is (metric-a1, metric-a2, ..., metric-an), and the performance array corresponding to path B is (metric-b1, metric-b2, ..., metric-bn), if each metric in path A's performance array is better than or equal to the metric at the corresponding position in path B's performance array, that is, metric_a1 ≥ metric_b1, metric_a2 ≥ metric_b2, ..., metric_an ≥ metric_bn, then path A is considered the second-dominant network path (i.e., the first case). Otherwise (e.g., metric_a1 ≥ metric_b1, metric_a2 < metric_b2), it is considered that paths A and B cannot be compared (ignoring the network performance requirements of the specific application), and both paths need to be retained as the second-dominant paths.
[0134] To further deepen our understanding of the partial order comparison method, the following is an example to illustrate the above comparison process: Please refer to Figure 10 for details. Taking the two target factors "bandwidth + delay" as an example, it is assumed that there are 5 network paths (i.e., m = 5) starting from the same network node x and reaching the same destination node y (the network path may cross one or more other network nodes in the middle), namely P0, P1, P2, P3, and P4. The network nodes involved in the specific network paths are shown in Figure 10, and their performance arrays are P0: (bandwidth 20, delay 5), P1: (bandwidth 10, delay 2), P2: (bandwidth 20, delay 10), P3: (bandwidth 10, delay 6), and P4: (bandwidth 18, delay 10). According to the partial order comparison screening method described in the embodiment of the present application, paths P2 and P4 do not need to be recorded because both their attributes are worse than or equal to P0; path P3 does not need to be recorded because both its attributes are worse than or equal to P1; and paths P0 and P1 cannot be compared to determine their superiority (because the bandwidth of path P0 is better than that of path P1, and the latency of path P1 is better than that of path P0), so both paths are recorded as second-best paths.
[0135] It can be theoretically proved that the partial order comparison method based on the performance array described in the embodiment of the present application achieves maximum path information reduction, that is, it records the minimum number of non-optimal paths while ensuring the accuracy of multi-objective factor requirements. The proof process is as follows: Assuming that all the second-dominant network paths obtained by the partial order reduction method based on the performance array in the embodiment of the present application cannot guarantee the accuracy of the multi-objective factor requirements, it means that the local path of one or several hops in the optimal end-to-end path is not included in the second-dominant network path reduced by the method of the embodiment of the present application; here it is assumed that a certain hop PathM in the optimal end-to-end path is not included in the second-dominant network path calculated by the method of the embodiment of the present application, and "PathM+other hop paths" constitute the true optimal path of the multi-objective factors (that is, the path calculation of this scheme is wrong), then according to the calculation rules of the second-dominant network path, another path PathN must be found in the second-dominant network path calculated by the method of the embodiment of the present application, so that all performance indicators of PathN are better than or equal to PathM (otherwise, PathN and PathM cannot be compared, and then PathM will also appear in the second-dominant network path, which contradicts the assumption). Further, it can be concluded that the performance of "PathN+other hop path" is better than or equal to "PathM+other hop path", that is, the present application has found an end-to-end path with better performance than the end-to-end optimal path, which is obviously illogical. Therefore, the assumption is wrong, that is, the second advantageous network path selected based on the method of the embodiment of the present application must include an end-to-end optimal path with multiple objective factors.
[0136] In summary, after performing a partial order comparison on m network paths based on n target factors, the network path that is more advantageous after the partial order comparison or the network path that cannot be compared to determine the superiority or inferiority is taken as the second advantageous network path. Then, based on the specific needs, the first advantageous network path is obtained based on the obtained second advantageous network path. It should be noted that in the embodiments of the present application, there are different situations in which the first advantageous network path is obtained based on the second advantageous network path. The different situations are described below, including but not limited to:
[0137] Scenario 1: Each second advantageous network path serves as a first advantageous network path.
[0138] This scenario generally applies to relatively small networks (e.g., RTN networks with a maximum of 100 nodes). Therefore, the number of second-best network paths selected using the partial-order comparison method of the performance array is relatively small. Therefore, even if subsequent iterative calculations are performed based on the full set of second-best network paths, the computational overhead for the overall multi-objective factor scenario is low. For example, in an RTN network with 100 nodes, the optimal path calculation for all end-to-end multi-objective factor scenarios can be completed in hundreds of milliseconds.
[0139] Case 2: When there are at least p second advantageous network paths, q are selected from the p second advantageous network paths as first advantageous network paths, where m≥p≥2 and p≥q≥1.
[0140] Assume that when the network path calculation method described in the embodiment of the present application is applied to an ultra-large-scale network (e.g., the number of network nodes reaches tens of thousands), the number of records of the second advantage network path may gradually increase as the iterative calculation proceeds. At this time, the iterative calculation based on the full amount of the second advantage network path will also have a relatively large computational overhead. As shown in Figure 11, in an ultra-large-scale network, the second advantage network path recorded in each iteration will become larger and larger: at the beginning, there are only two second advantage network paths to a certain destination node: Path0 and Path1. As the iterative calculation proceeds, the second advantage network path to a subsequent destination node has been expanded to 8 (Path0~Path7), which also leads to lower and lower efficiency of network path calculation. At this time, if the business carried by the network is a real-time business, the degradation of the computing efficiency of the network path will affect the real-time nature of the business.
[0141] To address this issue, the embodiments of the present application provide an improved solution to further optimize the computational overhead of multi-objective factor scenarios in large-scale networks, thereby meeting the demand for fast computing for real-time services.
[0142] The specific improvement scheme is the scheme described in the second scenario, that is, when there are at least p second advantageous network paths, q are selected from the p second advantageous network paths as the first advantageous network paths, m≥p≥2, p≥q≥1. In other words, when the number of second advantageous network paths reaching the same destination node reaches a certain number p (e.g., p=4), in order to ensure the efficiency of the calculation, the subsequent iterative calculation is no longer based on the full amount of second advantageous network paths, but q (e.g., q=2) are selected from the currently obtained second advantageous network paths for iterative calculation, thereby optimizing the calculation overhead. The selected second advantageous network path is the first advantageous network path, wherein the selection method can be random selection or selection based on certain selection rules (e.g., selecting fewer hops, larger bandwidth, etc.), and this application does not limit this. It should be noted here that in some embodiments of the present application, if multiple rounds of calculations need to be repeated, the operation of selecting q paths from p second-dominant network paths as first-dominant network paths in the first round can be randomly selected, but in subsequent rounds of calculations, the optimal path among the first-dominant network paths selected in the previous round can be iterated to the next round as one of the feasible solutions. For example, assuming that there are currently 10 full-scale advantageous paths, namely paths 1 to 10, and paths 2, 3, and 7 are randomly selected in the first round, and it is found that the optimal path among these three paths is 3, then path 3 will be recorded. In the next iterative calculation, three more paths 1, 5, and 8 are randomly selected; together with the local optimal path 3 from the previous time, there will be a total of four paths, 1, 3, 5, and 8, participating in the calculation of subsequent paths. The current local optimal path (such as 5) is then recorded, and iteration continues until the target path obtained in the last round meets the convergence condition or reaches the preset number of rounds.
[0143] It should be noted that no matter the execution subject is a source node or a control node, the process of executing step 704 by the execution subject is similar, so they will not be described separately here.
[0144] 705 : Delete the first node from the set of nodes to be added, and use the updated target path as a new first reachable network path.
[0145] After obtaining the updated target path, the first node may be deleted from the set of nodes to be added, and the updated target path may be used as a new first reachable network path.
[0146] For example, using Figure 9 as an example, there are four first-reachable network paths, {0→1, 0→2, 0→3, 0→4}, and five second-reachable network paths (the first node selected from the set of nodes to be added is node 3), {0→3→1, 0→3→2, 0→3→7, 0→3→8, 0→3→9}. The resulting target paths are 4+5=9, {0→1, 0→3→1, 0→2, 0→3→2, 0→3, 0→4, 0→3→7, 0→3→8, 0→3→9}. These nine target paths are the target paths before the update. Assume that among these nine target paths, there are two groups of at least two network paths with node 0 as the source node and reaching the same destination node. One group is two network paths reaching the same destination node 1, {0→1, 0→3→1}, and the other group is two network paths reaching the same destination node 2, {0→2, 0→3→2}. Assume that the first dominant network path obtained after comparing the first set of network paths is {0→1}, and the first dominant network path obtained after comparing the second set of network paths is {0→3→2}. Then, the paths {0→3→1, 0→2} are deleted from the nine target paths, and the updated target paths are seven, namely {0→1, 0→3→2, 0→3, 0→4, 0→3→7, 0→3→8, 0→3→9. The updated target paths {0→1, 0→3→2, 0→3, 0→4, 0→3→7, 0→3→8, 0→3→9} serve as the new first reachable network paths and are subsequently repeated in steps 702 to 705. In this iteration, after deleting node 3, the network nodes included in the node set to be added are {1, 2, 4, 5, 6, 7, 8, 9}.
[0147] It should be noted that no matter the execution subject is a source node or a control node, the process of executing step 705 by the execution subject is similar, so they will not be described separately here.
[0148] 706. Repeat steps 702 to 705.
[0149] After the updated target path is used as the new first reachable network path, steps 702 to 705 are repeatedly performed until the set of nodes to be added is empty.
[0150] To facilitate understanding of this repeated execution process, Figure 12 is used as an example below: Assume that the network topology determined in the target network includes a total of 10 network nodes, namely nodes 1-9, and assume that the source node is node 0 and the other network nodes are nodes 1-9. Assume that based on the network topology, there are 4 adjacent nodes (i.e., directly connected nodes) that can be reached by node 0, namely node 1, node 2, node 3, and node 4. Therefore, there are four first reachable network paths starting from the source node and ending at its directly connected node, recorded as {0→1, 0→2, 0→3, 0→4} (the number of iterations can be recorded as 0 in this case). Now, we determine whether the set of nodes to be added is empty. As shown in Figure 12, the set of nodes to be added is {1, 2, 3, 4, 5, 6, 7, 8, 9}. Assuming that the first node selected from the set of nodes to be added is node 3, in the existing network topology, we can find the second reachable network paths starting from node 0 and relayed by node 3 as {0→3→1, 0→3→2, 0→3→7, 0→3→8, 0→3→9}. The target paths are then {0→1, 0→3→1, 0→2, 0→3→2, 0→3, 0→4, 0→3→7, 0→3→8, 0→3→9}, for a total of 9 paths. After performing partial order comparison on the network paths reaching the same destination node among the 9 target paths based on the above method, the path {0→3→1, 0→2} is deleted from the 9 target paths. The updated target paths are 7, namely {0→1, 0→3→2, 0→3, 0→4, 0→3→7, 0→3→8, 0→3→9}. The set of nodes to be added is {1, 2, 4, 5, 6, 7, 8, 9} (the number of iterations can be recorded as 1 at this time). Afterwards, in the second iteration, (i.e., the number of iterations is 2), and we determine whether the set of nodes to be added is empty. As shown in Figure 12, the set of nodes to be added is {1, 2, 4, 5, 6, 7, 8, 9}. Assuming that the first node selected from the set of nodes to be added is node 6, in the existing network topology, we can find the second reachable network paths with node 0 as the starting point and node 6 as the relay as {0→3→6→1, 0→3→6→2, 0→6→3, 0→6→4, 0→6→5, 0→6}, for a total of 6 paths. Since the new first reachable network paths obtained in the first iteration are {0→1, 0→3→2, 0→3, 0→4, 0→3→7, 0→3→8, 0→3→9}, for a total of 7 paths.Then the target paths obtained in the second iteration are 6+7=13, which are {0→1, 0→3→6→1, 0→3→2, 0→3→6→2, 0→3, 0→6→3, 0→4, 0→6→4, 0→6→5, 0→6, 0→3→7, 0→3→8, 0→3→9}. Similarly, after performing partial order comparison on these 9 target paths based on the above method, a partial order comparison is performed on the network paths reaching the same destination node among these 9 target paths, that is, a partial order comparison is performed on the four groups of network paths {0→1, 0→3→6→1}, {0→3→2, 0→3→6→2}, {0→3, 0→6→3}, and {0→4, 0→6→4}. The result of the comparison assumes that {0→1, 0→3→6→1} are all second-dominant network paths (i.e., there is no Method to compare the advantages and disadvantages), and {0→3→6→2}, {0→3}, and {0→4} are the second-dominant network paths, then delete the paths {0→3→2, 0→6→3, 0→6→4} from these 13 network paths, and the updated target paths are 10, namely {0→1, 0→3→6→1, 0→3→6→2, 0→3, 0→4, 0→6→5, 0→6, 0→3→7, 0→3→8, 0→3→9}. These 10 target paths are used as the first reachable network paths in the third iteration. At this time, the set of nodes to be added is {1, 2, 4, 5, 7, 8, 9} (the number of iterations at this time can be recorded as 2); after that, continue to determine whether the set of nodes to be added is empty and the subsequent operations according to the above similar steps until the set of nodes to be added is empty.
[0151] It should be noted that no matter whether the execution subject is a source node or a control node, the process of the execution subject executing step 706 is similar, so it will not be explained separately here.
[0152] 707. Based on the network performance requirements of the data to be transmitted on the target device, determine a target network path with the source node as the starting point and the destination node corresponding to the data to be transmitted as the end point from the first reachable network path obtained from the last update. The target network path is used to transmit the data to be transmitted.
[0153] If it is determined that the set of nodes to be added is empty, that is, there is no network node in the set of nodes to be added, then the first reachable network path obtained by the last update is the first reachable network path finally obtained in the current round.
[0154] It should be noted here that, in the embodiment of the present application, the round is different from the number of iterations. The round refers to the number of times steps 701 to 706 are executed, and executing steps 701 to 706 once is one round; while the number of iterations refers to the number of times steps 702 to 706 are repeatedly executed until the set of nodes to be added is empty, and executing steps 702 to 705 once is one iteration. A round may contain one or more iterations.
[0155] It should also be noted that, as can be seen from step 704, there are different situations in which the first advantageous network path is obtained based on the second advantageous network path, such as the above-mentioned situation 1 and situation 2. In some embodiments of the present application, the situations are slightly different for different situations, which are described below:
[0156] (1) For scenario 1: each second dominant network path serves as a first dominant network path.
[0157] Since this scenario generally applies to relatively small networks (e.g., RTN networks with a maximum of 100 nodes), the number of second-best network paths selected using the partial order comparison method of the performance array is relatively small. Therefore, even if subsequent iterative calculations are performed based on the full set of second-best network paths, the overall computational overhead of the multi-objective factor scenario is not high.
[0158] Therefore, in this case, the end-to-end optimal path must be included in the first reachable network path updated in the last round of the current round. Based on this, the first target network path with the source node as the starting point and the destination node corresponding to the data to be transmitted as the end point can be directly determined from the first reachable network path obtained in the last update of the current round according to the network performance requirements of the data to be transmitted on the target device, and the data to be transmitted is transmitted based on the first target network path. For example, a network node (such as a source node, a control node, etc.) can record the first reachable network path obtained in the last update to reach all other network nodes in a multi-target factor routing table, and then calculate the optimal path in the dominant network path to each destination node according to the network performance requirements of the data to be transmitted, and mark the optimal path. After the optimal path to all destination nodes is calculated, the network node reads the destination node IP address in the application data packet header, and forwards each data packet of the data to be sent to the next hop network node according to the optimal path to the IP address.
[0159] To facilitate understanding of this process, Figure 12 is used as an example. Assume that the first reachable network path obtained from the last update of the current round contains 12 network paths, namely {0→1, 0→3→6→1, 0→7→2, 0→8→3, 0→4, 0→6→5, 0→6, 0→3→9→7, 0→2→8, 0→3→8, 0→1→8, 0→7→9}. If data to be transmitted on the target device (assuming host 1, connected to node 0) needs to be sent to the destination device (assuming host 2), which is connected to node 8, and the network performance requirement for the data to be transmitted is: prioritize the path with the lowest latency. If multiple paths have the lowest latency at the same time, prioritize the path with the lowest packet loss rate (i.e., considering both latency and packet loss rate). From the above, it can be seen that the destination node corresponding to the data to be transmitted is node 8, and from the first reachable network path obtained by the last update of the current round, it can be seen that there are 3 network paths with node 0 as the starting point and the destination node 8 corresponding to the data to be transmitted as the end point, namely {0→2→8, 0→3→8, 0→1→8}. Then, according to the network performance requirements of the data to be transmitted described above, the optimal path obtained is {0→3→8}, then the path {0→3→8} is the first target network path. Finally, the data to be transmitted is transmitted based on the first target network path {0→3→8}.
[0160] It should be noted here that, in traditional routing, once the IP address is determined, the routing table entries are also fixed (e.g., always taking the most cost-effective route), and each network node runs address-by-address routing. Therefore, as long as the source / destination IP address is determined, the routing table will calculate a unique and definite route for the pair of addresses. Unlike the IGP single-target factor scenario, the embodiment of the present application is aimed at a multi-target factor scenario. For the same network performance requirement, any two network nodes in the network may need to record multiple advantageous paths rather than a single optimal path. In addition, the same source / destination node in the multi-target factor scenario itself may have multiple different optimal paths due to differences in application network performance requirements, and network nodes also need to distinguish the paths of different business data. To this end, the present application also designs a distributed multi-target factor routing table by application (which can be by specific application or by application type (e.g., social, game, etc.)), as shown in Figure 13. Since the embodiment of the present application is a multi-target factor scenario, it can perform multi-target factor path calculation based on the SLA requirements of different applications. The same source / destination node may take different paths due to different services carried, so it is routed by service. Therefore, the situation shown in Figure 13 will occur. For the same source / destination address pair, due to the different network performance requirements of application A and application B, different routing tables are required to record the optimal paths for different applications. For each application, the routing table needs to record different performance indicators (Metric1, Metric2, Metric3, ...) of each path to facilitate iterative calculation; in addition, using the network path calculation method described in the embodiment of the present application, even in the same application routing table, multiple different first reachable network paths (i.e., updated target paths) and corresponding path performance will be recorded for the same destination address, so it is necessary to indicate the optimal path for a specific application in the table entry. After the user enters the command to display the routing table, in addition to displaying the routing table for each application, it is also necessary to specify the specific network performance requirements for each type of business, as well as the optimal path among the multiple first reachable network paths calculated using the solution of this application under the network performance requirements. As shown in Figure 13, * is used to mark the optimal path, and the other recorded items in the routing table are all first reachable network paths. The embodiment of the present application realizes differentiated routing based on different applications between the same source / destination address pairs by designing a multi-objective factor routing table that matches the network path calculation method described in the embodiment of the present application, breaking through the limitation of traditional routing tables that can only perform address-by-address routing.
[0161] It should be noted that in the embodiments of the present application, the application A and application B can be specific applications (e.g., application A is WeChat, application B is QQ, etc.), or they can represent application types (e.g., application A refers to a social application, application B refers to a shopping application). The application types can be distinguished in advance, and each specific application can be divided into a specific application type. Then, the network node can maintain a routing table under each specific application or application type, and one application or application type corresponds to a network performance requirement.
[0162] It should also be noted that Figure 13 is a specific implementation of the multi-objective factor routing table illustrated in the embodiment of the present application. In other implementations of the present application, there may also be routing table designs in other forms, and the present application does not limit this. As long as the routing table adopts application-by-application (also known as business-by-business) routing, one routing table for each type of application (different from other solutions that route by IP, all businesses share one table entry), and the routing table records multiple performance parameters of each path, it can be called the multi-objective factor routing table described in the embodiment of the present application. It should be noted that the multi-objective factor routing table needs to specify the specific network performance requirements of each type of application, as well as the optimal path among the multiple advantageous paths calculated by the embodiment of the present application under the network performance requirements, otherwise the data cannot be forwarded correctly.
[0163] In summary, the network path calculation method for scenario 1, by introducing records of local non-optimal paths, allows for iterative end-to-end path calculation based on non-optimal local paths. This ensures the accuracy of the optimal path in the multi-objective factor scenario where "local optimal path + local optimal path + ... + local optimal path ≠ global optimal path." The dominant path screening method, based on partial order comparison of performance vectors, ensures accurate path calculation while minimizing the number of non-optimal path records, thereby maximizing computational overhead and ensuring efficient optimal path calculation. The dominant path screening method, based on partial order comparison of performance arrays, can filter out non-dominant paths that do not need to be recorded during the iterative calculation process to the greatest extent possible, thereby maximizing computational overhead and ensuring efficient calculation.
[0164] (2) For scenario 2: when there are at least p second dominant network paths, q are selected from the p second dominant network paths as first dominant network paths, m≥p≥2, p≥q≥1.
[0165] Scenario 2 is an improved solution for large-scale networks. The specific implementation can be simply summarized as follows: as the iterative calculation proceeds, when the number of second advantageous network paths recorded to reach the same destination node reaches a certain number p (e.g., p=4), in order to ensure the efficiency of the calculation, the subsequent iterative calculation is no longer based on the full amount of second advantageous network paths, but q (e.g., q=2) are selected from the currently obtained second advantageous network paths for iterative calculation, thereby optimizing the calculation overhead. For services with special requirements (such as voice, video and other services with high real-time requirements), the second target network path with the source node as the starting point and the destination node corresponding to the data to be transmitted as the end point can be determined from the first reachable network path obtained by the last update of the current round according to the network performance requirements of the data to be transmitted on the target device, and the data to be transmitted is transmitted based on the second target network path. This process is similar to the processing process corresponding to the above scenario 1 and will not be repeated here.
[0166] However, the first reachable network path obtained in the last update of the current round is obtained based on the selected second dominant network path. In this way, the first reachable network path calculated based on partial information in the current round may not include the end-to-end optimal path. Based on this, in other embodiments of the present application, a multi-round calculation method is adopted to make the end-to-end path calculated each time based on the partially selected second dominant network path gradually converge to the global optimal path. Specifically, the above steps 701 to 706 can be repeatedly executed until the first reachable network path obtained in the last round meets the convergence condition or reaches a preset number of rounds (e.g., 50 times).
[0167] Taking the random selection method as an example, the specific method of multiple rounds of calculation is as follows: when calculating the first reachable network path in the next round, when the number of second dominant network paths reaching the same destination node reaches a certain number p (for example, p = 4), the method of randomly selecting q second dominant network paths from them for subsequent iterative calculations is also adopted. However, unlike the previous round of calculations, this round of calculations can also add the local optimal path recorded in the previous round to the second dominant network path to be randomly selected in this round, and participate in subsequent iterative calculations together. In this way, after multiple rounds of repeated calculations, the end-to-end path calculated based on the random partial information will eventually converge to the global optimal solution.
[0168] It should be noted that in some embodiments of the present application, assuming that based on the network performance requirements of the data to be transmitted on the target device, the target network path that is currently transmitting the data to be transmitted is the second target network path, and after the next round or multiple rounds of calculations, the target network path that meets the preset conditions (e.g., the optimal target network path) is the third target network path, which means that the network performance of the third target network path is better than the second target network path. In this case, the transmission path of the data to be transmitted can be switched from the second target network path to the third target network path. It should be noted that the switching method may include but is not limited to:
[0169] a. Switch modes at any time
[0170] This switching mode applies to a single round of calculations. Specifically, if the network performance of the first reachable network path (i.e., the third target network path) obtained in the next round of calculations is better than the second target network path currently being transmitted, the switch will be initiated. The principle of this switching mode is to switch whenever a more optimal target network path is available. The advantage of this on-the-fly switching mode is that the transmission path of the data to be transmitted always remains on the currently calculated optimal network path, improving transmission efficiency.
[0171] In addition, it should be noted that in some embodiments of the present application, the transmission path may be switched after each round of calculation of the first reachable network path of the current round. This is because in each next round of calculation, if the record of the optimal path of the previous round is introduced, each switch can be optimized towards a better path. Specifically, the present application does not limit the specific implementation method of the random switching mode.
[0172] b. Switch mode once
[0173] This switching mode is applicable to multiple rounds of calculation. Specifically, until the first reachable network path obtained in the last round reaches convergence conditions or reaches a preset number of rounds, the data to be transmitted continues to be transmitted via the second target network path obtained in the current round (i.e., the first round). After all rounds are completed, a third target network path that meets the preset conditions (e.g., the optimal target network path among all rounds) is selected from the first reachable network paths obtained in all rounds and switched to. The advantage of this single-shot switching is that it eliminates the need to frequently change the transmission path for the data to be transmitted, ensuring the stability of the transmission path.
[0174] For ease of understanding, the calculation process shown in Figure 14 is used as an example for explanation below: the horizontal part of Figure 14 represents the iterative calculation process of the first reachable network path of the advantage path tree in each round, and each box records the number of advantage paths currently reaching the current destination node; while the vertical part represents the calculation of the first reachable network path of the advantage path tree in multiple rounds, and the information interaction relationship between each round and the next round of calculation.
[0175] As shown in Figure 14, during the first round of first-reachable network path calculation (i.e., the first round of target path calculation), when the third destination node (assuming node 3) is calculated, the number of second-dominant network paths reaches four (Path0, Path1, Path2, Path3). Assume here that p = 4. To optimize computational overhead, two of the four second-dominant network paths are randomly selected (i.e., q = 2) to participate in the iterative calculation of the destination path for the next network node (assuming node 4). In Figure 14, during the first round of calculation, node 3 randomly selects Path1 and Path2, and Path2 has the better performance of the two paths. Assuming that four paths are also calculated when calculating the second-dominant network path for node 4, two paths (Path0 and Path2) are also randomly selected to participate in the iterative calculation of the destination path for the next network node (assuming node 5). Path0, the path with the best performance of the two, is recorded. This process continues until the destination path calculation for all destination nodes is completed. Since the amount of information required for iterative calculations is reduced starting at node 3, each end-to-end path calculated based on partial information is not globally optimal, necessitating subsequent recalculations of the first reachable network path. Furthermore, after the current round of first reachable network path calculations is complete, latency-sensitive services can initially transmit data along the currently calculated approximate optimal path. If subsequent calculations optimize the approximate optimal path, the service data transmission path can be switched at any time, further improving service quality.
[0176] During the second round of first-reachable network path calculation, four second-dominant network paths are calculated when the third destination node is reached. Again, two paths are randomly selected to participate in the destination path calculation for node 4. Assuming that Path0 and Path1 are randomly selected in the second round, plus the optimal random path Path2 of node 3 in the previous round, a total of three paths (Path0, Path1, and Path2) participate in the destination path calculation for node 4. Path0, the path with the best performance among the three paths, is also recorded. Similarly, after node 4 calculates four second-dominant network paths, it also randomly selects two paths to calculate the destination path for node 5. Assuming that Path2 and Path3 are randomly selected in the second round, plus the optimal random path Path0 in the previous round, a total of three paths (Path0, Path2, and Path3) participate in the destination path calculation for node 5. This process continues in this way until the destination paths corresponding to all destination nodes are calculated. Since the randomly selected path in the second round of calculation may be different from that in the previous round, and the optimal path in the previous round is also included in the path calculation in this round, the optimal network path to each destination node may have changed after this round of calculation. At this time, the data to be transmitted that has been transmitted along the results of the previous round of calculation can be switched to the new path (i.e., the third target network path) for continued transmission (here taking the switching mode at any time as an example).
[0177] Similarly, when calculating the first reachable network path in the third round, each destination node still randomly selects two second-dominant network paths, plus the optimal path among the paths selected in the second round, a total of three paths participate in the iterative calculation of the destination path of the next node, and then further update the optimal path from the source node to each destination node. In this way, after multiple rounds of repeated calculations, the end-to-end path calculated based on the randomly selected local path will eventually converge to the global optimal path, which is the same as the calculation result based on the full-scale dominant path. However, the advantage of the embodiment of the present application is that it can first calculate an approximate optimal path for real-time services to initiate data transmission, and gradually optimize and adjust the current transmission path to the optimal path during the service transmission process, which is particularly suitable for large-scale networks.
[0178] It should also be noted that in the embodiment of the present application, since the transmission of the data to be transmitted has already begun after the first round of calculation of the first reachable network path is completed, the optimal path to the destination node corresponding to the data to be transmitted may be updated after each subsequent round of calculation is completed. Therefore, it is possible to frequently adjust the data transmission path. When the data transmission path is switched during the transmission process, it may cause data transmission disorder problems.
[0179] Figure 15 illustrates a data transmission out-of-order problem caused by path switching: After the service data transmits packet 17, the network node calculates a new optimal path and switches subsequent data transmission (data after packet 18) to the new path. However, because the latency of the new path is longer than the original path, the order in which the packets arrive at the destination node becomes disordered. For example, packets 18 and 19 arrive at the destination node before packet 15. When the classic TCP protocol detects data out-of-order, it treats the out-of-order data as lost and initiates data retransmission, which degrades transmission performance.
[0180] Based on this, an embodiment of the present application also provides a solution to the problem of data disorder, that is, when a network node (such as a source node) obtains a switching instruction, it waits for a preset period of time before initiating the transmission of untransmitted data packets in the data to be transmitted on a new path (i.e., the third target network path). The specific solution is also shown in Figure 15: after calculating the new transmission path, the network node waits for a period of time Δt before initiating subsequent data transmission on the new path. In some embodiments of the present application, the waiting time Δt can be the delay difference between the new path and the old path, so that the first data packet transmitted by the new path (i.e., data packet No. 18) can be ensured to arrive at the destination node no earlier than the last packet transmitted by the original path (i.e., data packet No. 17), thereby achieving order preservation of data transmission during path switching.
[0181] In summary, the network path calculation method corresponding to Scenario 2 improves the real-time performance of network services in multi-objective factor scenarios in large-scale networks. This solution does not pursue the accurate calculation of the end-to-end optimal path in one go, as this would incur excessive computational overhead and affect the real-time performance of network services. In contrast, this solution improves the traditional routing calculation method of calculating the optimal path in one go, and designs a progressive calculation method that randomly selects some advantageous paths and combines multiple rounds: the first round of calculation quickly calculates an approximate end-to-end optimal path based on a small amount of path information, so that real-time services can initiate service data transmission on the approximate optimal path in a timely manner. Subsequently, the accuracy of the approximate optimal solution is continuously optimized through multiple rounds of iterative calculations until it converges to the true optimal path. In addition, in order to address the problem of data transmission disorder that may be caused by path switching, delayed transmission is used to ensure the order of service data during path switching.
[0182] It should be noted that the network path calculation method described in the embodiment of the present application is a general method, which can be distributedly deployed in network nodes or traversed and implemented in a centralized controller; it can be applied to overlay networks or underlay networks; and the optimal path can be calculated according to any network performance requirements.
[0183] To further understand the embodiments of the present application, two examples are used as typical examples to illustrate the implementation process of the network path calculation method provided in the embodiments of the present application in specific application scenarios.
[0184] Example 1: Audio and video service routing scenario in RTN network.
[0185] FIG16 is a schematic diagram of an application scenario for routing audio and video services in an RTN network provided by an embodiment of the present application, wherein the network nodes in the RTN network can be referred to as RTN nodes or RTN servers. The network path calculation method provided by the embodiment of the present application can be implemented in the RTN network shown in FIG16. Since the RTN network belongs to an overlay network, the routing strategy can be customized through application layer software on a centralized control node without the need to modify the hardware of the physical device, which makes it easier to implement. At the same time, the audio and video services carried by the RTN network usually need to comprehensively consider multiple factors such as network latency, jitter, and packet loss when calculating the transmission path. The application scenario matches the solution of the embodiment of the present application.
[0186] In this example, assume that the network performance requirements for audio and video services are summarized as follows:
[0187] Case 1: When the packet loss rate of multiple end-to-end paths is less than 10%, the path with the lowest latency is prioritized. When multiple paths have the lowest latency simultaneously, the path with the lowest latency jitter is prioritized.
[0188] Case 2: When the packet loss rates of multiple end-to-end paths are between 10% and 20%, a 50ms delay is converted into a 10% packet loss rate, thereby unifying delay and packet loss rate into a comprehensive performance metric. The path with the best comprehensive performance metric is preferentially selected.
[0189] Case 3: When the packet loss rates of multiple end-to-end paths are all above 20%, the path with the lowest packet loss rate is prioritized. When the packet loss rates of multiple paths are simultaneously the lowest, the path with the lowest latency is prioritized.
[0190] When two or more of Case 1, Case 2, and Case 3 exist simultaneously, for example, among all end-to-end paths, some have a packet loss rate less than 10%, some have a packet loss rate greater than 20%, and some have a packet loss rate between 10% and 20%, the path is selected in the order of "Case 1 is better than Case 2, which is better than Case 3": If a path in Case 1 exists, the optimal path in Case 1 is returned first; if no path in Case 1 exists, the optimal path in Case 2 is returned first.
[0191] In addition, the RTN network has a unified management node (i.e., a control node. Generally, RTN networks have such a management node. If there is no management node, each RTN node needs to calculate its own target path. Here, the control node is used as the execution subject for illustration). This node is responsible for collecting performance measurement data between network nodes (including latency, jitter, and packet loss rate measurements) and distributing it to each RTN node. The path calculation process is also completed uniformly at the control node, and the specific calculation results are then distributed to each RTN node. Based on the existing network calculation logic of the RTN network, the specific implementation steps of this example solution in this multi-objective factor scenario can be as follows:
[0192] Step 1: The control node collects RTN network information.
[0193] This operation is performed by the control node of the RTN network and is performed periodically during the operation of the entire RTN network (regardless of whether there is a specific data transmission service). The specific actions are as follows:
[0194] 1.1. The control node collects performance measurement data of the RTN network at regular intervals (e.g., 1 minute). Specific network performance indicators include latency, jitter, and packet loss rate.
[0195] 1.2. The control node determines the connectivity between RTN nodes based on performance measurement results (i.e., performance measurement data) (for example, if the packet loss rate exceeds 90%, the two nodes are considered disconnected). Based on the connectivity between the two RTN nodes, the control node integrates the topology of the RTN network, i.e., the network topology.
[0196] 1.3. When a device (e.g., a mobile phone, personal computer, smartwatch, or other terminal device) connects to an RTN node, the RTN node notifies the control node of the device's IP address and RTN node identifier. The control node then completes the mapping table between the device's IP address and the RTN node. Once the mapping table is complete, the control node responds to the corresponding RTN node.
[0197] 1.4. After the above steps are completed, the control node obtains a periodically updated device IP and RTN node mapping table, as well as a periodically updated RTN network topology and performance measurement data.
[0198] Step 2: The source device announces the network performance requirements of the service.
[0199] This step is performed by the source device and the RTN node together, and is initiated when an application on a specific source device needs to initiate data transmission. The specific action flow is as follows:
[0200] 2.1. A source device (eg, host 1) is connected to an RTN node in the RTN network. The RTN node notifies the control node of the information of host 1 through step 1.3 described above.
[0201] 2.2. After receiving the response from the control node that has executed step 1.3, the RTN node responds to host 1 and allows it to initiate data transmission.
[0202] 2.3. Host 1 notifies the destination node IP and the network performance requirements of the service to the RTN node it is connected to, and the RTN node further feeds back to the control node.
[0203] It should be noted that whenever a new service initiates data transmission on host 1, steps 2.1 to 2.3 need to be executed to notify the RTN node and control node of the communication addresses of different services and the network performance requirements of the services.
[0204] Step 3: The control node calculates the target path in a multi-target scenario and distributes the routing table. This step begins after the control node receives the service network performance of host 1 from the RTN node. Host 1 can only perform data transmission operations after this step is completed. The specific action flow is as follows:
[0205] 3.1. The control node receives the network performance requirements of the service on host 1 forwarded by the RTN node.
[0206] 3.2. The control node selects the network performance indicators of interest to the application (in this example, the performance indicators of interest to the application are latency, packet loss rate, and jitter) from the local measurement information record (i.e., the performance measurement data in 1.1), composes a hop-by-hop performance array (latency, packet loss rate, and jitter), and matches the performance array to each overlay path in the RTN network topology.
[0207] 3.3. The control node finds the IDs of the RTN nodes connected to host 1 and host 2 in the node mapping table based on the IP address of host 1 and the IP address of the destination device (e.g., host 2). These two RTN nodes are referred to as the source RTN node and the destination RTN node, respectively.
[0208] 3.4. The control node uses the source RTN node as the source node, all other RTN nodes as the destination nodes, and the indicators in the selected performance array (such as delay, packet loss, and jitter) as target factors to execute the network path calculation method described in the embodiment of the present application. The specific execution steps are as follows:
[0209] 3.4.1. Include the source RTN node in the target path (also called the dominant path tree), and include all other RTN nodes in the set of nodes to be added.
[0210] 3.4.2. Calculate the reachable network path from the source node to the reachable destination node (i.e., the directly connected node), and update the path performance of the reachable network path in the dominant path tree.
[0211] 3.4.2. Determine whether the set of nodes to be added is empty, and proceed to different branch steps based on the judgment result:
[0212] a. If the set of nodes to be added is not empty, proceed to step 3.4.4 and update a node to be added (ie, the first node mentioned above) to the dominant path tree.
[0213] b. If the set of nodes to be added is empty, proceed to step 3.4.5 to calculate the optimal path to each RTN node (i.e., the target path described above).
[0214] 3.4.4. Select a new RTN node from the set of nodes to be added and add it to the dominant path tree. The specific steps are as follows:
[0215] 3.4.4.1. From the set of nodes to be added, select the RTN node with the lowest latency among all nodes in the dominant path tree and add it to the dominant path tree.
[0216] 3.4.4.2. Using the newly added node as a relay, update the set of reachable paths to all destination nodes that pass through the new node.
[0217] 3.4.4.3. For all reachable paths to each RTN node in the dominant path tree, perform a partial order comparison of the three factors of "delay, packet loss, and jitter" and record the dominant path.
[0218] 3.4.4.4. Record the new dominant path to each node in the dominant path tree and delete the selected node from the set of nodes to be added.
[0219] 3.4.4.5. The dominant path tree is updated, and the process returns to step 3.4.3 to determine whether the dominant path tree needs to be updated.
[0220] 3.4.5. Calculate the optimal path from the source RTN node to any destination RTN node (the destination RTN node is determined when the service data is sent to a specific destination device, so all other RTN nodes may be destination RTN nodes). The specific execution steps are as follows:
[0221] 3.4.5.1. All the advantageous paths to the destination node are divided into three categories according to the packet loss rate range: Case 1 (packet loss rate < 10%), Case 2 (10% ≤ packet loss rate ≤ 20%), and Case 3 (packet loss rate > 20%).
[0222] 3.4.5.2. Depending on whether the number of paths in Case 1 / Case 2 / Case 3 is empty, proceed to different branch steps:
[0223] a. If there is a dominant path belonging to case 1, the path with the best latency is selected as the optimal path among the dominant paths in case 1. If there are multiple paths with the best latency at the same time, the path with the lowest jitter is selected as the optimal path among the multiple paths with the best latency and recorded.
[0224] b. If the optimal path in case 1 is empty but there is an optimal path in case 2, all the optimal paths in case 2 are calculated using the formula "metric_total = packet loss rate + delay / 5". The packet loss rate and delay are converted into a comprehensive factor metric_total. The path with the smallest metric_total is recorded as the optimal path.
[0225] c. If the dominant paths in Case 1 and Case 2 are both empty, it means that all dominant paths belong to Case 3. In this case, the path with the lowest packet loss rate is selected as the dominant path among all the dominant paths. If there are multiple paths with the lowest packet loss rate at the same time, the path with the lowest latency is selected as the optimal path and recorded.
[0226] 3.4.5.3. The optimal path information and the advantageous path information (i.e., the target path) are decentralized to the RTN nodes of each hop along the way to form a local routing table. It should be noted that in some embodiments of the present application, the local routing table may only include advantageous path information, and then each RTN node calculates the optimal path based on the network performance requirements of the service; the local routing table may also be a control node that calculates the optimal path based on the network performance requirements of the service, and then directly sends the optimal path information to the RTN nodes of each hop to form a local routing table. It should also be noted that the operation of forming the local routing table can be completed by the control node or by each RTN node itself, and this application does not limit this.
[0227] 3.5. Each RTN node along the route receives the optimal path information and / or dominant path information sent by the control node and forms a local routing table (i.e., the multi-objective factor routing table described above). Alternatively, each RTN node along the route receives the local routing table sent by the control node.
[0228] Step 4: The source device initiates service data transmission, and the RTN node forwards the data packet. This step is completed by the source device and all RTN nodes along the optimal path, and continues throughout the entire service data transmission process. The specific action flow is as follows:
[0229] 4.1. After receiving the routing table sent by the control node (or forming its own local routing table), the RTN node replies with confirmation information to the control node.
[0230] 4.2. After receiving confirmation information from all RTN nodes on the optimal path, the control node sends a data transmission permission signal to the source RTN node.
[0231] 4.3. After receiving the data transmission permission signal, the source RTN node forwards the signal to host 1.
[0232] 4.4. After receiving the signal allowing data transmission, host 1 sends the service data to the source RTN node.
[0233] 4.5. The source RTN node and subsequent RTN nodes on the optimal path forward the data along the optimal path to the destination RTN node based on the information in the routing table, and the destination RTN node forwards the data to host 2.
[0234] 4.6. After the service data transmission is completed, host 1 terminates the data transmission, and the RTN nodes on the optimal path stop forwarding data packets.
[0235] It should be noted that among the four main steps of the above solution, steps 1, 2, and 4 are unique data transmission processes for RTN networks. These three steps will change with the changes in application scenarios. For example, the data transmission process in the underlay physical device router will be different from the steps in this embodiment (but the overall method is similar). Step 3 is the core implementation of this solution.
[0236] From the process of step 3, it can be seen that it is the dominant path, not the optimal path, that determines the calculation results of subsequent iterations. Therefore, the network performance requirements of the service are only needed when calculating the optimal path in step 3.4.5. In the previous iterations of steps 3.1 to 3.4, it was mainly the dominant path set, whose performance arrays were incomparable to each other, that determined the information recorded in the routing table entry. Therefore, if the network performance requirements of the service change but the various target factors remain unchanged (for example, the optimal path calculation is still based on the three factors of delay, packet loss, and jitter, but the network performance requirements are different from those of this solution), in step 3, only the optimal path screening method needs to be updated in step 3.4.5, and no modifications or adjustments are required to the previous steps. Therefore, this solution has strong adaptability to changes in the network performance requirements of the service.
[0237] To further illustrate the technical effects of the embodiments of the present application, the present application obtained live network data of the consumer cloud RTN network and a solution for a multi-objective factor scenario of the current RTN network. The performance and accuracy of the multi-objective factor scenario of the present solution and the existing RTN network solution were further compared in RTN networks of different sizes. The comparison results are shown in FIG17 .
[0238] Unlike the solution in this application, the current calculation method used by the consumer cloud's RTN network for multi-objective factor scenarios is the aforementioned approximate optimal path algorithm based on the single-objective factor KSP. Specifically, the existing RTN network method uses the KSP algorithm to calculate the 10 paths with the shortest end-to-end delay, forming a set of candidate paths. Then, based on the network performance requirement of "case 1 is better than case 2, which is better than case 3", the optimal path among the 10 candidate paths is selected and recorded as the end-to-end optimal path.
[0239] The shortcomings of this method, as discussed previously, are that in highly heterogeneous networks, inaccurate calculations are very likely to occur. This solution can accurately calculate the optimal path based on the application network performance requirements within the top 10 latency paths. However, the optimal performance path for multiple objective factors may not necessarily appear within the top 10 latency paths. For example, the top 10 latency paths may have poor packet loss, bandwidth, cost, or other factors. In this case, the top 10 latency paths may miss the globally optimal path.
[0240] In order to further illustrate the problem of inaccurate calculation of this solution, the calculation results of this solution and the existing solution of the RTN network are listed in Figure 18. As can be seen from Figure 18, in a 50-node RTN network, when calculating the optimal path of multiple target factors from node 0 as the source node to other nodes, the calculation results of this application solution and the existing RTN solution are different: from the calculation results of node 0 reaching node 1, node 10, node 11, node 17 and node 20, it can be seen that the optimal path calculated by this application solution belongs to case 1 according to the network performance requirements of the business of this embodiment; and when the existing RTN solution calculates the optimal path of the same source / destination node, the calculated path belongs to case 2 and case 3. According to the calculation logic of "case 1 is better than case 2, which is better than case 3", it is obvious that the end-to-end optimal path calculated by this application solution is significantly better than the calculation result of the existing RTN solution. The reason for the inaccurate calculations in existing RTN solutions is that the top 10 paths do not include the path with the best overall performance. The latency of the paths calculated by existing RTN solutions is better than that of the proposed solution, but the overall performance (i.e., considering latency, jitter, and packet loss) is weaker than that of the proposed solution. Therefore, compared with existing solutions, the proposed solution has improved calculation accuracy. As previously demonstrated, the proposed solution can guarantee absolute accuracy of calculation results under any network conditions and service network performance requirements.
[0241] In addition, by comparing the computational efficiency of the present application solution and the existing consumer cloud RTN network solution in Figure 17, it can be found that the present application solution improves computational efficiency by a factor of 100 in large-scale network scenarios (more than 50 nodes). Therefore, the present application solution not only ensures the absolute accuracy of multi-objective factor calculations, but also significantly improves computational efficiency.
[0242] Example 2: Scenario where network performance indicators and conditional constraints exist simultaneously in multiple objective factors.
[0243] In the application scenario shown in Example 1, while the network performance requirements of the service encompass multiple factors, the calculation logic is primarily based on the "optimization" dimension, such as finding the optimal path for latency in Case 1 and the optimal path for packet loss in Case 3. However, in some application scenarios, not all metrics need to be optimal. Sometimes, optimizing certain performance metrics to within certain thresholds is sufficient, such as latency <300ms, bandwidth >500Mbps, and hop count no more than 4. We refer to these optimization objectives, which do not require an optimal solution, as conditionally constrained objectives.
[0244] If the network performance requirements of the business in actual applications are more complex than those in Example 1, when network performance indicators and conditional constraints exist simultaneously in the multi-objective factors, relying solely on the partial order comparison method of optimizing the target performance array (that is, the performance array only includes network performance indicators) to screen the dominant path may still result in calculation errors.
[0245] Figure 19 shows a scenario where an error in the optimal path calculation occurs when only the network performance indicators are partially filtered when the network performance indicators and the conditional constraint targets are both present in the network performance requirements of the business. In this scenario, the optimization targets are bandwidth and delay, and the constraint is the number of hops. The specific business requirements are: among the paths with no more than 4 hops, the path with the largest bandwidth is given priority; when the bandwidth of multiple paths is simultaneously the largest, the path with the lowest delay is given priority. In essence, it is a dual-objective optimal path calculation of bandwidth + delay, and the number of hops is used as a constraint, which only needs to be no more than 4 hops, and there is no need to optimize to the extreme value. Assuming that the application scenario of this embodiment is still the centralized calculation of the control node of the RTN network, according to the implementation method in Example 1, the advantageous path screening and iterative calculation are performed based on the two-dimensional performance array (bandwidth, delay), and the hop count is filtered and screened for the reachable paths with a threshold of 4. As shown in Figure 19, there are 5 reachable paths with node A as the destination node after iterative calculation, namely: P0: (bandwidth 20, delay 5), hop count 4; P1: (bandwidth 10, delay 2), hop count 2; P2: (bandwidth 20, delay 10), hop count 4; P3: (bandwidth 10, delay 2), hop count 2; P4: (bandwidth 18, delay 10), hop count 3. Since the hop count of all paths does not exceed the threshold of 4 hops, they are all valid paths. At this time, based on the partial order comparison of the two attributes (bandwidth, delay), the preferred paths selected are P0 and P1 (P0's two attributes are both better than or equal to P2 and P4, and P1's two attributes are both better than or equal to P3; P0 and P1 cannot be compared, P0 has better bandwidth, and P1 has better delay). Continue iterative calculation, assuming that there is a directly connected node B behind node A, the performance of the direct path between node A and node B is P AB : (Bandwidth 19, latency 3). At this time, in the control node's records, the dominant paths to node A are P0 and P1. Therefore, the control node calculates that there are only two possible paths to node B via node A, namely P0+P AB : (bandwidth 19, latency 8), hop count 5, and P1+P AB : (bandwidth 10, delay 5), hop count 3; due to the path P0+P AB , the number of hops exceeds the threshold limit and is therefore filtered out. Finally, according to the technical solution of Example 1, the optimal path to node B is calculated as: P1+P AB The performance of the path is: bandwidth 10, delay 5.
[0246] However, as shown in Figure 19, in the process of selecting the optimal path for node A, path P4 is eliminated due to its poor performance in terms of both bandwidth and latency. If this path P4 can be recorded, when calculating the optimal path to node B, there will be an additional candidate path P4+P in the candidate path set. AB : (bandwidth 18, latency 13), hop count 4. New path P4+P AB The number of hops still does not exceed the threshold, but compared with the optimal path P1+P calculated based on the partial order screening of the optimization target performance array AB , the new path has better performance (larger bandwidth). Therefore, when the business requirements are more complex and there are network performance indicators and conditional constraints, only optimizing the partial order comparison of network performance indicators may still result in inaccurate calculations.
[0247] In response to the limitations of the solution in Example 1, this Example 2 improves the solution. When business requirements include both network performance indicators and constraint conditions, the accuracy of multi-objective factor calculations is still guaranteed, making this solution applicable to the calculation of network paths in any multi-objective factor scenario.
[0248] Based on this, the technical solution of this Example 2 improves the calculation method of the dominant path and the optimal path in step 3 on the basis of the implementation process of Example 1. The implementation process of the specific new calculation method can be simply summarized in combination with Figure 20 as follows: when network performance indicators and conditional restriction targets exist in the business requirements at the same time, if the conditional restriction target of the current path does not reach the threshold, the conditional restriction target and the network performance indicator are combined to form a unified performance array, and partial order comparison and dominant path screening are performed based on the unified performance array in the entire path to the destination node; when the path's restriction condition target exceeds the threshold, it will no longer participate in the subsequent partial order comparison of the unified performance array.
[0249] Still taking the dual-factor optimization of "bandwidth + latency" and the hop limit of no more than 4 hops as shown in Figure 19 as an example, in this example, the specific technical solution implementation adjustments are as follows:
[0250] When the calculation iteration reaches node A as the destination, there are five reachable paths to node A: P0 (bandwidth 20, latency 5), hop count 4; P1 (bandwidth 10, latency 2), hop count 2; P2 (bandwidth 20, latency 10), hop count 4; P3 (bandwidth 10, latency 2), hop count 2; and P4 (bandwidth 18, latency 10), hop count 3. Since the hop counts of all paths do not exceed the threshold of 4 hops, the five paths are compared in a partial order using a unified performance array of the three factors (bandwidth, latency, and hop count). The resulting optimal paths are P0, P1, and P4. (When the three-factor array is partially compared, P0 and P4 cannot be compared because, although P0 has better bandwidth and latency than P4, P4 has a higher hop count than P0.) At this point, the control node records P0, P1, and P4 as optimal paths to node A and includes them in subsequent iterations.
[0251] Furthermore, when calculating the path to node B, based on the previous advantage record, there will be 3 candidate paths: P0+P AB (Bandwidth 19, latency 8, hop count 5), P1+P AB (Bandwidth 10, latency 5, hop count 3) and P4+P AB (Bandwidth 18, delay 13, hop count 4); due to the path P0+P AB The number of hops exceeds the threshold limit and is therefore filtered out. The ultimate advantageous path to node B is: P1+P AB and P4+P AB Afterwards, when calculating the optimal path, it can be found that P4+P AB The performance is better (larger bandwidth), so the optimal path to the destination node B is finally recorded as P4+P AB The solution in this example introduces the constraint condition target into the partial order comparison of the performance array, thereby successfully avoiding the error caused by the threshold limit in the candidate path screening.
[0252] In summary, the embodiment of the present application addresses the limitation that the classic IGP distributed algorithm cannot calculate the optimal path under conditional constraints, and designs a processing method for converting conditional constraint targets into optimization indicators. The parameters corresponding to the conditional constraint targets are uniformly processed according to the optimization indicators before reaching the threshold, and participate in the partial order comparison of the performance array together with other network performance targets, thereby ensuring that the optimal solution is not missed during the dominant path screening process. The embodiment of the present application can ensure the accuracy of network path calculation in multi-objective factor scenarios under any business requirements, that is, in any complex multi-objective factor scenario, the end-to-end optimal path can be correctly calculated.
[0253] It should be understood that Example 1 and Example 2 introduced above are only two specific scenarios in which the network path calculation method provided in the embodiment of the present application is applied. The network path calculation method provided in the embodiment of the present application is not limited to the above scenarios when applied, and this application will not elaborate on this.
[0254] On the basis of the above embodiment, in order to better implement the above scheme of the embodiment of the present application, the following also provides related devices for implementing the above scheme. Please refer to Figure 21 specifically, Figure 21 is a structural diagram of a network device provided in the embodiment of the present application, and the network device 2100 may specifically include: a determination module 2101, used to determine a first reachable network path with a source node as the starting point and a directly connected node of the source node as the end point, the source node being the network node to which the target device in the target network is connected; a selection module 2102, used to select a first node from the set of nodes to be added, and obtain a second reachable network path with the source node as the starting point and the first node as the relay, the set of nodes to be added being the set of other network nodes in the target network except the source node; a comparison module 2103, used to compare the m network paths based on n target factors if there are m network paths to the same destination node in the target path, obtain a first advantageous network path, and update the first reachable network path based on the first advantageous network path to obtain an updated first reachable network path, the target path including the first reachable network path and the second reachable network path, m≥2, n≥2.
[0255] In one possible design, the network device 2100 also includes an update module 2104 for deleting the first node from the set of nodes to be added, and triggering the selection module 2102, the comparison module 2103 and the update module 2104 to execute repeatedly until the set of nodes to be added is an empty set.
[0256] In a possible design, the comparison module 2103 is specifically configured to perform a partial order comparison on the m network paths based on n target factors to obtain second dominant network paths, and each second dominant network path serves as the first dominant network path.
[0257] In one possible design, the selection module 2102 is also used to: when the set of nodes to be added is an empty set after deleting the first node, determine the first target network path with the source node as the starting point and the destination node corresponding to the data to be transmitted as the end point from the first reachable network path obtained by the last update based on the network performance requirements of the data to be transmitted on the target device, and the first target network path is used to transmit the data to be transmitted.
[0258] In one possible design, the comparison module 2103 is further specifically used to: perform a partial order comparison on the m network paths based on n target factors to obtain a second advantageous network path; when there are at least p second advantageous network paths, select q from the p second advantageous network paths as the first advantageous network paths, m≥p≥2, p≥q≥1.
[0259] In one possible design, the selection module 2102 is also used to: when the set of nodes to be added is an empty set after the first node is deleted, determine, based on the network performance requirements of the data to be transmitted on the target device, a second target network path with the source node as the starting point and the destination node corresponding to the data to be transmitted as the end point from the first reachable network path obtained by the last update, and the second target network path is used to transmit the data to be transmitted.
[0260] In one possible design, the network device 2100 also includes: a cyclic trigger module 2105, which is used to trigger the determination module 2101, the judgment module 2102, the selection module 2102, the comparison module 2103 and the update module 2104 to execute repeatedly until the target path obtained in the last round meets the convergence condition or reaches a preset number of rounds.
[0261] In one possible design, the loop triggering module 2105 can be used to: determine that when there are at least p second dominant network paths, the current set of nodes to be added is the target node set; use the target node set as the new set of nodes to be added, trigger the selection module 2102, the comparison module 2103 and the update module 2104 to execute repeatedly until the first reachable network path obtained in the last round meets the convergence condition or reaches the preset number of rounds.
[0262] In one possible design, the selection module 2102 is further specifically used to: switch the transmission path of the data to be transmitted from the second target network path to a third target network path, where the third target network path is one of the target paths obtained in multiple rounds that meets preset conditions, and the network performance of the third target network path is better than the network performance of the second target network path.
[0263] In one possible design, the selection module 2102 is further specifically used to: obtain a switching instruction, the switching instruction is used to instruct the source node to wait for a preset period of time before initiating the transmission of untransmitted data packets in the data to be transmitted on the third target network path.
[0264] In one possible design, the comparison module 2103 is further specifically used to: construct m performance arrays of the m network paths based on the n target factors, each performance array includes n elements arranged in order, one network path corresponds to one performance array, and one element corresponds to the specific value of a target factor on the corresponding network path; when the value of each element in the first performance array is greater than or equal to the value of the element at the corresponding position in the second performance array, determine that the network path corresponding to the first performance array is the second advantageous network path, and the first performance array and the second performance array are any two of the m performance arrays.
[0265] In one possible design, the comparison module 2103 is further specifically used to: when there is a first element in the first performance array whose value is greater than or equal to the value of the element at the corresponding position in the second performance array, and there is a second element in the first performance array whose value is less than or equal to the value of the element at the corresponding position in the second performance array, determine that the network path corresponding to the first performance array and the network path corresponding to the second performance array are both the second advantageous network path, and the first element and the second element are any two elements in the first performance array.
[0266] In one possible design, the selection module 2102 is further specifically used to: based on the network performance requirements of the data to be transmitted on the target device, select the first node that meets the preset performance requirements from the set of nodes to be added.
[0267] In one possible design, the update module 2104 is specifically used to: in the case where there are m network paths to the same destination node, only retain the first advantageous network path, and use the target path of the retained first advantageous network path as the updated first reachable network path.
[0268] In one possible design, the target factor includes at least one or more of the following: a network performance indicator, or a conditional constraint target.
[0269] In one possible design, the network performance indicator includes at least one of the following: bandwidth, delay, packet loss rate, jitter, and cost.
[0270] In one possible design, the target network includes at least one of the following: an overlay network or an underlay network.
[0271] It should be noted that the information interaction, execution process, etc. between the modules / units in the network device 2100 provided in Figure 21 are based on the same concept as the method embodiment corresponding to Figure 7 in this application. For specific contents, please refer to the description in the method embodiment shown above in this application, and will not be repeated here.
[0272] The embodiment of the present application also provides a network device, please refer to Figure 22, Figure 22 is a structural diagram of the network device provided in the embodiment of the present application. For ease of explanation, only the parts related to the embodiment of the present application are shown. For specific technical details not disclosed, please refer to the method part of the embodiment of the present application. The network device 2200 can be deployed with the modules described in the embodiment corresponding to Figure 21 to implement the functions of the network device 2100 in the embodiment corresponding to Figure 21. Specifically, the network device 2200 is implemented by one or more servers. The network device 2200 may have relatively large differences due to different configurations or performances. It may include one or more central processing units (CPUs) 2222 and memories 2232, and one or more storage media 2230 (e.g., one or more mass storage devices) for storing application programs 2242 or data 2244. Among them, the memories 2232 and the storage media 2230 can be temporary storage or persistent storage. The program stored in the storage medium 2230 may include one or more modules (not shown), each of which may include a series of instruction operations on the network device 2200. Furthermore, the central processor 2222 can be configured to communicate with the storage medium 2230 and execute a series of instruction operations in the storage medium 2230 on the network device 2200.
[0273] The network device 2200 may also include one or more power supplies 2226, one or more wired or wireless network interfaces 2250, one or more input and output interfaces 2258, and / or one or more operating systems 2241, such as Windows Server™, Mac OS X™, Unix™, Linux™, FreeBSD™, etc.
[0274] In an embodiment of the present application, the central processing unit 2222 is used to execute the calculation method of the network path in the embodiment corresponding to Figure 7. For example, the central processing unit 2222 can be used to: first, determine a reachable network path with a source node as the starting point and a directly connected node of the source node as the end point. The reachable network path can be referred to as a first reachable network path (which can be one or more). It should be noted that the reachable network path includes not only the network nodes involved in the reachable network path, but also the overall network performance of the entire reachable network path (such as the latency, bandwidth, packet loss rate, etc. of the reachable network path). Among them, the source node is a network node in the network (which can be referred to as the target network) to which the target device (i.e., the device with data to be sent) is connected. Afterwards, it is determined whether the set of nodes to be added is an empty set, wherein the set of nodes to be added is a set of other network nodes in the target network except the source node. If it is determined that the set of nodes to be added is not an empty set (i.e., there are network nodes in the set of nodes to be added), a node (which can be called the first node) is selected from the set of nodes to be added to obtain a reachable network path with the source node as the starting point and the first node as the relay. The reachable network path can be called the second reachable network path. It should be noted that in some embodiments of the present application, the intermediate nodes of the second reachable network path do not include other network nodes in the set to be added except the first node. Moreover, the intermediate nodes of the second reachable network path do not include the first network node (i.e., the source node) and the last destination node (i.e., the destination node) of the second reachable network path, but refer to any network node in the "middle". For example, assuming that the second reachable network path is A→B→C→…→X, where A, B, C,…, X are all network nodes, then the intermediate nodes of the second reachable network path refer to any network node in B, C,…, but do not include the source node A and the destination node X. All reachable network paths included in the first reachable network path and the second reachable network path can be referred to as the target path (one or more). If there are m network paths reaching the same destination node in the target path, then these m network paths are compared based on n target factors to obtain a first dominant network path. The target path is then updated based on the first dominant network path to obtain an updated target path, where m ≥ 2 and n ≥ 2. After obtaining the updated target path, the first node can be deleted from the set of nodes to be added, and the updated target path can be used as the new first reachable network path. After using the updated target path as the new first reachable network path, the above-mentioned steps of determining whether the set of nodes to be added is empty and subsequent steps are repeated until the set of nodes to be added is empty.
[0275] It should be noted that the central processing unit 2222 can also be used to execute any step in the method embodiment corresponding to Figure 7 in this application. For specific details, please refer to the description in the method embodiment shown above in this application, and will not be repeated here.
[0276] A computer-readable storage medium is also provided in an embodiment of the present application. The computer-readable storage medium stores a program for performing signal processing. When the program is run on a computer, the computer executes the steps performed in the description of the aforementioned embodiment.
[0277] It should also be noted that the device embodiments described above are merely illustrative, wherein the units described as separate components may or may not be physically separate, and the components displayed as units may or may not be physical units, that is, they may be located in one place, or they may be distributed across multiple network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the present embodiment. In addition, in the drawings of the device embodiments provided in this application, the connection relationship between the modules indicates that there is a communication connection between them, which can be specifically implemented as one or more communication buses or signal lines.
[0278] Through the description of the above embodiments, those skilled in the art can clearly understand that the present application can be implemented by means of software plus necessary general-purpose hardware, and of course can also be implemented by dedicated hardware including application-specific integrated circuits, dedicated CPUs, dedicated memories, dedicated components, etc. In general, all functions performed by computer programs can be easily implemented by corresponding hardware, and the specific hardware structures used to implement the same function can also be diverse, such as analog circuits, digital circuits, or dedicated circuits. However, for the present application, software program implementation is a better implementation method in most cases. Based on this understanding, the technical solution of the present application is essentially or the part that contributes to the prior art can be embodied in the form of a software product, which is stored in a readable storage medium, such as a computer floppy disk, USB flash drive, mobile hard disk, read-only memory (ROM), random access memory (RAM), magnetic disk or optical disk, etc., and includes a number of instructions for causing a computer device (which can be a personal computer, training equipment, or network equipment, etc.) to execute the methods described in each embodiment of the present application.
[0279] In the above embodiments, all or part of the embodiments may be implemented by software, hardware, firmware, or any combination thereof. When implemented by software, all or part of the embodiments may be implemented in the form of a computer program product.
[0280] The computer program product includes one or more computer instructions. When the computer program instructions are loaded and executed on a computer, the process or function described in the embodiment of the present application is generated in whole or in part. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. The computer instructions can be stored in a computer-readable storage medium or transmitted from one computer-readable storage medium to another computer-readable storage medium. For example, the computer instructions can be transmitted from one website, computer, training equipment or data center to another website, computer, training equipment or data center via a wired (e.g., coaxial cable, optical fiber, digital subscriber line) or wireless (e.g., infrared, wireless, microwave, etc.) method. The computer-readable storage medium can be any available medium that a computer can store or a data storage device such as a training device, data center, etc. that includes one or more available media integrations. The available medium can be a magnetic medium (e.g., a floppy disk, a hard disk, a tape), an optical medium (e.g., a high-density digital video disc (DVD)), or a semiconductor medium (e.g., a solid-state drive (SSD)).
Claims
1. A method for calculating a network path, characterized in that: include: Determine a first reachable network path starting from a source node and ending at a node directly connected to the source node, wherein the source node is a network node connected to a target device in a target network; Selecting a first node from a set of nodes to be added, and obtaining a second reachable network path with the source node as a starting point and the first node as a relay, wherein the set of nodes to be added is a collection of other network nodes in the target network except the source node; If there are m network paths with the same destination node as the end point in the target path, the m network paths are compared based on n target factors to obtain a first dominant network path, and the first reachable network path is updated based on the first dominant network path to obtain an updated first reachable network path. The target path includes the first reachable network path and the second reachable network path, m≥2, n≥2.
2. The method according to claim 1, characterized in that After obtaining the updated first reachable network path, the method further includes: The first node is deleted from the set of nodes to be added, and the steps of selecting the first node from the set of nodes to be added and subsequent steps are repeated until the set of nodes to be added is empty.
3. The method according to claim 2, characterized in that The comparing the m network paths based on the n target factors to obtain a first advantageous network path includes: A partial order comparison is performed on the m network paths based on n target factors to obtain a second advantageous network path, and each second advantageous network path serves as the first advantageous network path.
4. The method according to claim 3, characterized in that The method further comprises: When the set of nodes to be added is an empty set after deleting the first node, based on the network performance requirements of the data to be transmitted on the target device, a first target network path with the source node as the starting point and the destination node corresponding to the data to be transmitted as the end point is determined from the first reachable network path obtained from the last update, and the first target network path is used to transmit the data to be transmitted.
5. The method according to claim 2, characterized in that The comparing the m network paths based on the n target factors to obtain a first advantageous network path includes: Performing a partial order comparison on the m network paths based on the n target factors to obtain a second dominant network path; When there are at least p second advantageous network paths, q second advantageous network paths are selected from the p second advantageous network paths as first advantageous network paths, where m≥p≥2 and p≥q≥1.
6. The method according to claim 5, characterized in that The method further comprises: When the set of nodes to be added is an empty set after deleting the first node, based on the network performance requirements of the data to be transmitted on the target device, a second target network path with the source node as the starting point and the destination node corresponding to the data to be transmitted as the end point is determined from the first reachable network path obtained by the last update, and the second target network path is used to transmit the data to be transmitted.
7. The method according to claim 6, characterized in that The method further comprises: Determining that when there are at least p second advantageous network paths, the current set of nodes to be added is the target node set; The target node set is used as a new set of nodes to be added, and the steps of selecting the first node from the set of nodes to be added, obtaining a second reachable network path with the source node as the starting point and the first node as the relay, and subsequent steps are repeated until the first reachable network path obtained in the last round meets the convergence condition or reaches a preset number of rounds, wherein selecting the first node from the set of nodes to be added, obtaining a second reachable network path with the source node as the starting point and the first node as the relay, and subsequent steps are performed once as one round.
8. The method according to claim 7, characterized in that The method further comprises: The transmission path of the data to be transmitted is switched from the second target network path to a third target network path, where the third target network path is one of the target paths obtained in multiple rounds that meets preset conditions, and the network performance of the third target network path is better than that of the second target network path.
9. The method according to claim 8, characterized in that The switching of the transmission path of the to-be-transmitted data from the second target network path to the third target network path comprises: A switching instruction is obtained, where the switching instruction is used to instruct the source node to wait for a preset time period before initiating transmission of untransmitted data packets in the data to be transmitted on the third target network path.
10. The method according to any one of claims 3 to 9, characterized in that The performing partial order comparison on the m network paths based on the n target factors to obtain the second dominant network path includes: Based on the n target factors, construct m performance arrays for the m network paths, each performance array including n elements arranged in order, one performance array corresponding to one network path, and one element corresponding to a value of a target factor on the corresponding network path; When the value of each element in the first performance array is greater than or equal to the value of the element at the corresponding position in the second performance array, the network path corresponding to the first performance array is determined to be the second advantageous network path, and the first performance array and the second performance array are two of the m performance arrays.
11. The method according to claim 10, characterized in that The method further comprises: When there is a first element in the first performance array whose value is greater than or equal to the value of the element at the corresponding position in the second performance array, and there is a second element in the first performance array whose value is less than or equal to the value of the element at the corresponding position in the second performance array, it is determined that the network path corresponding to the first performance array and the network path corresponding to the second performance array are both the second advantageous network paths, and the first element and the second element are two elements in the first performance array.
12. The method according to any one of claims 1 to 11, characterized in that The selecting a first node from the set of nodes to be added comprises: Based on the network performance requirements of the data to be transmitted on the target device, a first node that meets the preset performance requirements is selected from the set of nodes to be added.
13. The method according to any one of claims 1 to 12, characterized in that Updating the first reachable network path based on the first advantageous network path to obtain an updated first reachable network path includes: Only the first advantageous network path is retained, and the target path of the retained first advantageous network path is used as the updated first reachable network path.
14. The method according to any one of claims 1 to 13, characterized in that The target factors include at least one or more of the following: Network performance indicators, or,conditional constraints targets.
15. The method according to claim 14, characterized in that The network performance indicator includes at least one of the following: Bandwidth, latency, packet loss rate, jitter, and cost.
16. The method according to any one of claims 1 to 15, characterized in that The target network includes at least one of the following: Overlay network, or underlay network.
17. A network device, characterized in that: include: a determination module, configured to determine a first reachable network path starting from a source node and ending at a node directly connected to the source node, wherein the source node is a network node connected to a target device in a target network; A selection module is configured to select a first node from a set of nodes to be added, and obtain a second reachable network path with the source node as a starting point and the first node as a relay, wherein the set of nodes to be added is a collection of other network nodes in the target network except the source node; a comparison module, configured to compare, if there are m network paths reaching the same destination node in the target path, the m network paths based on n target factors to obtain a first dominant network path, and update the first reachable network path based on the first dominant network path to obtain an updated first reachable network path, wherein the target path includes the first reachable network path and the second reachable network path, and m ≥ 2, n ≥ 2.
18. The network device according to claim 17, wherein: The network device further includes an update module, configured to: The first node is deleted from the set of nodes to be added, and the selection module, the comparison module and the update module are triggered to execute repeatedly until the set of nodes to be added is empty.
19. The network device according to claim 18, wherein: The comparison module is specifically used to: A partial order comparison is performed on the m network paths based on n target factors to obtain a second advantageous network path, and each second advantageous network path serves as the first advantageous network path.
20. The network device according to claim 19, wherein: The selection module is further configured to: When the set of nodes to be added is an empty set after deleting the first node, based on the network performance requirements of the data to be transmitted on the target device, a first target network path with the source node as the starting point and the destination node corresponding to the data to be transmitted as the end point is determined from the first reachable network path obtained from the last update, and the first target network path is used to transmit the data to be transmitted.
21. The network device according to claim 18, wherein: The comparison module is further configured to: Performing a partial order comparison on the m network paths based on the n target factors to obtain a second dominant network path; When there are at least p second advantageous network paths, q second advantageous network paths are selected from the p second advantageous network paths as first advantageous network paths, where m≥p≥2 and p≥q≥1.
22. The network device according to claim 21, wherein: The selection module is further configured to: When the set of nodes to be added is an empty set after deleting the first node, based on the network performance requirements of the data to be transmitted on the target device, a second target network path with the source node as the starting point and the destination node corresponding to the data to be transmitted as the end point is determined from the first reachable network path obtained by the last update, and the second target network path is used to transmit the data to be transmitted.
23. The network device according to claim 22, wherein: The network device further includes a cycle triggering module, configured to: Determining that when there are at least p second advantageous network paths, the current set of nodes to be added is the target node set; The target node set is used as a new set of nodes to be added, and the selection module, the comparison module, and the update module are triggered to repeatedly execute the above steps until the first reachable network path obtained in the last round reaches the convergence condition or the preset number of rounds is reached, wherein one round consists of selecting the first node from the set of nodes to be added, obtaining the second reachable network path with the source node as the starting point and the first node as the relay, and the subsequent steps.
24. The network device according to claim 23, wherein: The selection module is further configured to: The transmission path of the data to be transmitted is switched from the second target network path to a third target network path, where the third target network path is one of the target paths obtained in multiple rounds that meets preset conditions, and the network performance of the third target network path is better than that of the second target network path.
25. The network device according to claim 24, characterized in that The selection module is further configured to: A switching instruction is obtained, where the switching instruction is used to instruct the source node to wait for a preset time period before initiating transmission of untransmitted data packets in the data to be transmitted on the third target network path.
26. The network device according to any one of claims 19 to 25, characterized in that: The comparison module is further configured to: Based on the n target factors, construct m performance arrays for the m network paths, each performance array including n elements arranged in order, one performance array corresponding to one network path, and one element corresponding to a value of a target factor on the corresponding network path; When the value of each element in the first performance array is greater than or equal to the value of the element at the corresponding position in the second performance array, the network path corresponding to the first performance array is determined to be the second advantageous network path, and the first performance array and the second performance array are any two of the m performance arrays.
27. The network device according to claim 26, wherein: The comparison module is further configured to: When there is a first element in the first performance array whose value is greater than or equal to the value of the element at the corresponding position in the second performance array, and there is a second element in the first performance array whose value is less than or equal to the value of the element at the corresponding position in the second performance array, it is determined that the network path corresponding to the first performance array and the network path corresponding to the second performance array are both the second advantageous network paths, and the first element and the second element are any two elements in the first performance array.
28. The network device according to any one of claims 17 to 27, characterized in that: The selection module is further configured to: Based on the network performance requirements of the data to be transmitted on the target device, a first node that meets the preset performance requirements is selected from the set of nodes to be added.
29. The network device according to any one of claims 17 to 28, characterized in that: The comparison module is specifically used to: Only the first advantageous network path is retained, and the target path of the retained first advantageous network path is used as the updated first reachable network path.
30. The network device according to any one of claims 17 to 29, characterized in that: The target factors include at least one or more of the following: Network performance indicators, or,conditional constraints targets.
31. The network device according to claim 30, wherein: The network performance indicator includes at least one of the following: Bandwidth, latency, packet loss rate, jitter, and cost.
32. The network device according to any one of claims 17 to 31, characterized in that: The target network includes at least one of the following: Overlay network, or underlay network.
33. A network device comprising a processor and a memory, wherein the processor is coupled to the memory, wherein: The memory is used to store programs; The processor is configured to execute the program in the memory, so that the network device executes the method according to any one of claims 1 to 16.
34. A computer-readable storage medium, characterized in that The invention comprises a program which, when being run on a computer, causes the computer to execute the method according to any one of claims 1 to 16.
35. A computer program product comprising instructions, characterized in that When the method is executed on a computer, the computer is enabled to execute the method according to any one of claims 1 to 16.
36. A chip, characterized in that: The chip includes a processor and a data interface, and the processor reads instructions stored in a memory through the data interface to execute the method according to any one of claims 1 to 16.
Citation Information
Cited By
Intelligent global cloud live broadcast system based on AI and multi-operator and multi-cloud optimization
CN120640020A
