Intelligent scheduling method of network traffic based on multi-dimensional data

Abstract

This invention discloses a method for intelligently scheduling network traffic based on multi-dimensional data. The method includes: after all nodes are started, they report the multi-dimensional node data to a scheduling cluster; the scheduling cluster calculates node scores based on the multi-dimensional node data, filters and distributes available node information to each node; and after a user sends a data packet and it arrives at the source node (i.e., the starting node), an intelligent scheduling algorithm uses a path evaluation function to determine the optimal routing path, forwarding the routing information between nodes until it reaches the target node and the target user. This method accelerates network transmission, improves the overall availability of network transmission services, including service quality and disaster recovery, and reduces the overall cost of network transmission services.

Description

Technical Field

[0001] The present invention belongs to the field of network communication technology, and in particular relates to a method for intelligently scheduling network traffic based on multi-dimensional data. Background Art

[0002] In recent years, the development of the internet and streaming media has garnered increasing attention. Companies are widely implementing their own network acceleration services, and network routing algorithms are becoming increasingly diverse. Commonly used methods include fixed routing, the Dijkstra algorithm, and various AI algorithms. These scheduling methods simply implement fixed scheduling based on information from different provinces and carriers, or calculate the minimum latency path based on latency. They fail to comprehensively calculate paths based on multidimensional data. Furthermore, the following issues exist: 1. They fail to fully consider multidimensional data, such as jitter, packet loss, and machine load; 2. They focus solely on algorithms, ignoring node availability, such as excessive pressure on a particular node; 3. Their implementation is complex, impractical, and difficult to implement. Summary of the Invention

[0003] The technical problem to be solved by the present invention is to address the deficiencies of the above-mentioned existing technologies and provide a network traffic intelligent scheduling method based on multi-dimensional data, which acts on network transmission acceleration services, improves the overall availability of network transmission services: service quality, disaster recovery, and reduces the overall cost of network transmission services.

[0004] In order to achieve the above technical objectives, the technical solution adopted by the present invention is:

[0005] A network traffic intelligent scheduling method based on multi-dimensional data is used to accelerate network transmission, including:

[0006] Step 1: After all nodes are started, they report multi-dimensional node data to the scheduling cluster;

[0007] Step 2: The scheduling cluster calculates node scores based on multi-dimensional node data, filters and sends available node information to each node;

[0008] Step 3: After the user sends a data packet to the source node, that is, the starting node, the intelligent scheduling algorithm obtains the best routing path based on the path evaluation function to forward the routing information between nodes, and finally reaches the target node and the target user.

[0009] To optimize the above technical solutions, specific measures taken also include:

[0010] The above multi-dimensional node data includes processor usage, memory usage, inter-node round-trip delay, inter-node jitter, and inter-node packet loss rate.

[0011] In step 2 above, the node scores S of all nodes are calculated, overloaded nodes are filtered out, and available nodes are retained.

[0012] The above node score calculation formula is:

[0013] S=CPU*W[1]+MEM*W[2]+RTT*W[3]+Jitter*W[4]+Loss*W[5]

[0014] Among them, CPU, MEM, RTT, Jitter, and Loss represent the node processor utilization, memory utilization, inter-node round-trip delay, inter-node jitter, and inter-node packet loss rate, respectively;

[0015] W[1], W[2], W[3], W[4], and W[5] are the processor weight, memory weight, RTT weight, jitter weight, and packet loss rate weight, respectively.

[0016] In step 3 above, the source node calculates the best routing path based on the node information and forwards it to the next node. The next node obtains the routing information and forwards it to the next node; eventually, it reaches the target node and the target user.

[0017] The process of the above intelligent scheduling algorithm includes:

[0018] S1. Maintain an open list openList and a closed list closeList;

[0019] Among them, the nodes with the F value of the path evaluation function to be checked are stored in openList, which represents the set of reachable nodes;

[0020] closeList stores nodes that are no longer checked, indicating a set of unreachable nodes;

[0021] S2. In each main loop, the node with the smallest F value is searched from the openList as the current node;

[0022] S3. Add the neighboring nodes of the current node to the openList as nodes to be checked, and add the current node to the closeList to prevent re-checking;

[0023] S4. Process the neighbor nodes of the current node (reachable but not in the closeList). If the neighbor node is not in the openList, set F and parentNode and add it to the openList for selection.

[0024] S5. If the neighbor node is already in the openList and the F value from the current node to the neighbor node becomes smaller, update the data of the neighbor node;

[0025] S6. Check the next neighbor node and repeat S2-S5.

[0026] If S7 and openList are empty, the search is complete and no path is found. If the current node is the destination node, the path is found.

[0027] The above evaluation function is:

[0028]

[0029] Where P[i] is the current node i; n is the number of nodes from the starting node to the current node;

[0030] F(P[i]) accumulates the node scores S of the path from the starting node P[src] to P[i].

[0031] The present invention has the following beneficial effects:

[0032] 1. Comprehensively consider multi-dimensional data, such as jitter, packet loss, and machine load, rather than just latency;

[0033] 2. Dynamically calculate the optimal nodes and paths, rather than just a fixed scheduling path, to avoid excessive pressure on certain nodes;

[0034] 3. Simple implementation, practicality and easy implementation;

[0035] 4. It is scalable and can be implanted with updated algorithms and more dimensional data. BRIEF DESCRIPTION OF THE DRAWINGS

[0036] Figure 1 This is a flowchart of the intelligent scheduling method of the present invention;

[0037] Figure 2 This is the logical flow of the intelligent scheduling algorithm of the present invention. DETAILED DESCRIPTION

[0038] The embodiments of the present invention are described in further detail below with reference to the accompanying drawings.

[0039] like Figure 1 and Figure 2 As shown, the network traffic intelligent scheduling method based on multi-dimensional data of the present invention includes:

[0040] Step 1: After all nodes are started, they report multi-dimensional node data to the scheduling cluster;

[0041] Step 2: The scheduling cluster calculates node scores based on multi-dimensional node data, filters and sends available node information to each node;

[0042] Step 3: After the user sends a data packet to the source node, that is, the starting node, the intelligent scheduling algorithm obtains the best routing path based on the path evaluation function to forward the routing information between nodes, and finally reaches the target node and the target user.

[0043] In an embodiment, the multi-dimensional node data includes processor usage, memory usage, round-trip delay between nodes, jitter between nodes, and packet loss rate between nodes.

[0044] Step 2 calculates the node scores S of all nodes, filters out overloaded nodes, and retains available nodes.

[0045] The node score calculation formula is:

[0046] S=CPU*W[1]+MEM*W[2]+RTT*W[3]+Jitter*W[4]+Loss*W[5]

[0047] Among them, CPU, MEM, RTT, Jitter, and Loss represent the node processor utilization, memory utilization, inter-node round-trip delay, inter-node jitter, and inter-node packet loss rate, respectively;

[0048] W[1], W[2], W[3], W[4], and W[5] are the processor weight, memory weight, RTT weight, jitter weight, and packet loss rate weight, respectively.

[0049] In step 3, the source node calculates the best routing path based on the node information and forwards it to the next node. The next node obtains the routing information and forwards it to the next node; eventually, it reaches the target node and the target user.

[0050] The process of the intelligent scheduling algorithm includes:

[0051] S1. Maintain an open list openList and a closed list closeList;

[0052] Among them, the nodes with the F value of the path evaluation function to be checked are stored in openList, which represents the set of reachable nodes;

[0053] closeList stores nodes that are no longer checked, indicating a set of unreachable nodes;

[0054] S2. In each main loop, the node with the smallest F value is searched from the openList as the current node;

[0055] S3. Add the neighboring nodes of the current node to the openList as nodes to be checked, and add the current node to the closeList to prevent re-checking;

[0056] S4. Process the neighbor nodes of the current node (reachable but not in the closeList). If the neighbor node is not in the openList, set F and parentNode and add it to the openList for selection.

[0057] S5. If the neighbor node is already in the openList and the F value from the current node to the neighbor node becomes smaller, update the data of the neighbor node;

[0058] S6. Check the next neighbor node and repeat S2-S5.

[0059] If S7 and openList are empty, the search is complete and no path is found. If the current node is the destination node, the path is found.

[0060]

[0061] Where P[i] is the current node i; n is the number of nodes from the starting node to the current node;

[0062] F(P[i]) accumulates the node scores S of the path from the starting node P[src] to P[i].

[0063] In summary, the present invention provides an intelligent algorithm for finding the best routing path for network packets.

[0064] Table 1 Symbol table

[0065] symbol illustrate src Starting Node dst Target Node allNodes All nodes S[i] Indicates the score of node i X[j] Indicates the jth scoring factor, such as CPU, MEM, etc. W[j] Indicates the weight of the j-th rating openList Represents the set of reachable nodes closeList Represents a set of unreachable nodes P[i] Represents the current node i F(P[i]) S accumulation of the path from the starting node P[src] to P[i]

[0066] Combined with Table 1, the calculation function of S[i] is:

[0067]

[0068] In this algorithm, after the S[i] algorithm is concretized:

[0069] S=CPU*W[1]+MEM*W[2]+RTT*W[3]+Jitter*W[4]+Loss*W[5]

[0070] The evaluation function F(P[i]) of P[i] is defined as:

[0071]

[0072] The overall implementation logic of the algorithm is:

[0073] 1. First calculate the node score S for all nodes, filter out overloaded nodes, and retain available nodes

[0074] 2 Maintains an open list openList and a closed list closeList;

[0075] The openList stores the nodes whose path evaluation function F values ​​are to be checked, and the closeList stores the nodes that are no longer checked.

[0076] 3. In each main loop, the node with the smallest F value is found from openList as the current node;

[0077] 4. Add the neighboring nodes of the current node to the openList as nodes to be checked. Add the current node to the closeList to prevent re-checking.

[0078] 5. Process the neighbor nodes of the current node (reachable but not in closeList). If the neighbor node is not in openList, then you need to set F and parentNode and add it to openList for selection.

[0079] 6 If this node is already in the openList, and the F value from the current node to this node becomes smaller, update the data of this neighbor node.

[0080] 7 Check the next neighbor node and repeat 3-6

[0081] 8 If openList is empty, the search is complete but no path is found. If the current node is the destination node, the path is found.

[0082] The present invention also provides a method for intelligently scheduling network traffic based on multi-dimensional data, including:

[0083] The data reporting module is used to report multi-dimensional node data to the scheduling cluster after all nodes are started;

[0084] The scoring calculation module is used to schedule the cluster to calculate node scores based on multi-dimensional node data, filter and send available node information to each node;

[0085] The scheduling module is used when the user sends a data packet to the source node, that is, the starting node. The intelligent scheduling algorithm uses the path evaluation function to derive the best routing path to forward routing information between nodes, and finally reaches the target node and target user.

[0086] The present invention has the following beneficial effects:

[0087] 1. Comprehensively consider multi-dimensional data, such as jitter, packet loss, and machine load, rather than just latency;

[0088] 2. Dynamically calculate the optimal nodes and paths, rather than just a fixed scheduling path, to avoid excessive pressure on certain nodes;

[0089] 3. Simple implementation, practicality and easy implementation;

[0090] 4. It is scalable and can be implanted with updated algorithms and more dimensional data.

[0091] In this embodiment, it should be understood that the disclosed method and system can be implemented in other ways. For example, the system implementation described above is merely illustrative. For example, the module division is merely a logical functional division. In actual implementation, other division methods may be used. For example, multiple modules or components may be combined or integrated into another system, or some features may be ignored or not implemented.

[0092] The modules described above as separate components may or may not be physically separate, and some or all of them may be selected based on actual needs to achieve the purpose of the present embodiment. Furthermore, the functional modules in various embodiments of the present invention may be integrated into a single processing module, each module may exist physically separately, or two or more modules may be integrated into a single module.

[0093] If the functions are implemented in the form of software modules and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, or the portion that contributes to the prior art, or the portion of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions for causing a computer device (which can be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in various embodiments of the present invention.

[0094] The aforementioned storage media include: U disk, mobile hard disk, read-only memory (ROM), random access memory (RAM), magnetic disk or optical disk and other media that can store program code.

[0095] It should be emphasized that the embodiments of the present invention are illustrative rather than restrictive, and therefore the present invention includes but is not limited to the embodiments described in the specific embodiments. Any other embodiments derived by those skilled in the art based on the technical solutions of the present invention also fall within the scope of protection of the present invention. The above are only preferred embodiments of the present invention, and the scope of protection of the present invention is not limited to the above embodiments. Any technical solutions under the ideas of the present invention fall within the scope of protection of the present invention. It should be pointed out that for those skilled in the art, several improvements and modifications without departing from the principles of the present invention should be regarded as the scope of protection of the present invention.

Claims

1. A network traffic intelligent scheduling method based on multi-dimensional data for achieving network transmission acceleration, characterized in that: include: Step 1: After all nodes are started, they report multi-dimensional node data to the scheduling cluster; Step 2: The scheduling cluster calculates node scores based on multi-dimensional node data, filters and sends available node information to each node; Step 3: After the user sends a data packet to the source node, i.e., the starting node, the intelligent scheduling algorithm uses the path evaluation function to determine the optimal routing path to forward the routing information between nodes, ultimately reaching the target node and the target user. The process of the intelligent scheduling algorithm includes: S1. Maintain an open list openList and a closed list closeList; Among them, the nodes with the F value of the path evaluation function to be checked are stored in openList, which represents the set of reachable nodes; closeList stores nodes that are no longer checked, indicating a set of unreachable nodes; S2. In each main loop, first determine whether openList is an empty list. If openList is not an empty list, find the node with the smallest F value from openList as the current node, and determine whether the current node is the destination node. If the current node is the destination node, find the path and backtrack along the parent node of the destination node. The algorithm ends. If the current node is not the destination node, enter S3; if openList is an empty list, it means that no path is found and the algorithm ends. S3, move the current node to closeList to prevent re-checking; S4. Process the neighbor node of the current node and determine whether the neighbor node is reachable and not in the closeList. If not, process the next neighbor node until the neighbor node traversal is completed and return to S2. If so, determine whether the neighbor node is in the openList. If the neighbor node is not in the openList, set the F and parentNode of the neighbor node and add it to the openList for selection and return to S2. S5. If the neighbor node is already in the openList and the F value from the current node to the neighbor node becomes smaller, then update the F data of the neighbor node and set the neighbor node as the current node. Then determine whether the current node is the destination node. If the current node is the destination node, find the path and backtrack along the parent node of the destination node. The algorithm ends. If the current node is not the destination node, enter S3. If the openList is an empty list, it means that no path is found and the algorithm ends.

2. The method for intelligent network traffic scheduling based on multi-dimensional data according to claim 1, characterized in that: The multi-dimensional node data includes processor usage, memory usage, inter-node round-trip delay, inter-node jitter, and inter-node packet loss rate.

3. The network traffic intelligent scheduling method based on multi-dimensional data according to claim 1 is characterized in that: Step 2 calculates the node scores S of all nodes, filters out overloaded nodes, and retains available nodes.

4. The method for intelligent network traffic scheduling based on multi-dimensional data according to claim 1, characterized in that: The node score calculation formula is: S=CPU*W[1]+MEM*W[2]+RTT*W[3]+Jitter*W[4]+Loss*W[5] Among them, CPU, MEM, RTT, Jitter, and Loss represent the node processor utilization, memory utilization, inter-node round-trip delay, inter-node jitter, and inter-node packet loss rate, respectively; W[1], W[2], W[3], W[4], and W[5] are the processor usage weight, memory usage weight, RTT weight, jitter weight, and packet loss rate weight, respectively.

5. The method for intelligent network traffic scheduling based on multi-dimensional data according to claim 1, characterized in that: In step 3, the source node calculates the best routing path based on the node information and forwards it to the next node. The next node obtains the routing information and forwards it to the next node; eventually, it reaches the target node and the target user.

6. The method for intelligent network traffic scheduling based on multi-dimensional data according to claim 1, characterized in that: The evaluation function is: Where P[i] is the current node i; n is the number of nodes from the starting node to the current node; F(P[i]) accumulates the node scores S[i] of the path from the starting node P[src] to P[i].

Images