Transmission control method and system based on cloud native network identifier

Abstract

The invention belongs to the technical field of big data services, and particularly discloses a transmission control method and system based on a cloud native network identifier, and the method comprises the steps: for a data layer, reducing the data volume of repeated services through a label reference algorithm, and repeating a message header through a virtual splicing technology; for a transmission layer, perception analysis is carried out on different channel types, different transmission control algorithms are adopted, and multi-channel dynamic switching is realized according to channel attribute changes in the transmission process; for a channel layer, path optimization, backup path remapping and multi-channel comprehensive transmission control are carried out for different access means; in a cloud native network, efficient transmission of information between a transmission network and an information network is realized through layered optimization of the data layer, the transmission layer and the channel layer.

Description

Technical Field

[0001] The present invention belongs to the technical field of big data services, and in particular, relates to a transmission control method and system based on cloud native network identification. Background Art

[0002] The current Internet OTT (over the top) model consists of the underlying transmission network and the upper-layer service information network. The transmission network and the information network are relatively independent. It is difficult for the transmission network to perceive the service quality requirements of the upper-layer service, resulting in low transmission efficiency and no service quality guarantee. At the same time, the existing transmission control method mainly uses three transmission services of the integrated platform, namely, long message, short message, and real-time message, and two modes of TCP and UDP transport layer protocols. The three transmission services of the integrated platform, namely, long message, short message, and real-time message, mainly use VMF for application identification mapping and addressing. One application corresponds to a global VMF number. Applications of different network nodes interact with each other through transmission services for signaling and data transmission and reception control. There is a lack of a perception mechanism for communication means and network status, and the scheduling of underlying network resources cannot be realized, resulting in low transmission reliability and extended transmission time. Especially for satellite, microwave and other means, the transmission efficiency is further reduced. TCP and UDP transport layer protocols mainly use port numbers for application identification mapping and addressing. TCP and UDP transport layer protocols are designed for wired networks, and data transmission and reception are controlled internally through congestion control algorithms. The original intention of designing the congestion control algorithm was to avoid congestion and packet loss in routing and switching equipment. However, this backoff principle is counterproductive for wireless channels (packet loss in wireless channels is not caused by congestion). Although many congestion control algorithms have been designed for wireless channels, they can only adapt to limited scenarios and cannot adapt to the dynamic changes of wireless networks, which also results in low transmission reliability and extended transmission time. Summary of the invention

[0003] The present invention provides a transmission control method and system based on cloud native network identification, which are used to solve the problem that it is difficult to break through the "barrier" between the transmission network and the information network in the cloud native network.

[0004] In a first aspect, a transmission control method based on a cloud native network identifier is provided, the method comprising:

[0005] For the data layer, the label reference algorithm is used to reduce the amount of repeated business data, and the virtual splicing technology is used to repeat the message header;

[0006] For the transmission layer, different channel types are perceived and analyzed, different transmission control algorithms are used, and dynamic switching of multiple channels is achieved according to changes in channel attributes during the transmission process;

[0007] At the channel layer, path optimization, backup path remapping and multi-channel integrated transmission control are performed for different access methods;

[0008] In a cloud-native network, efficient transmission of information between the transmission network and the information network is achieved through layered optimization of the data layer, transmission layer, and channel layer.

[0009] In a second aspect, a transmission control system based on a cloud native network identifier is provided, the system comprising:

[0010] The data layer optimization module is used to reduce the amount of duplicate business data through label reference algorithm and duplicate message headers through virtual splicing technology for the data layer;

[0011] The transport layer optimization module is used to perceive and analyze different channel types at the transport layer, adopt different transmission control algorithms, and realize dynamic switching of multiple channels according to changes in channel attributes during the transmission process;

[0012] The channel layer optimization module is used to perform path optimization, backup path remapping and multi-channel integrated transmission control for different access methods at the channel layer;

[0013] The information transmission module is used to achieve efficient transmission of information between the transmission network and the information network in the cloud native network through layered optimization of the data layer, the transmission layer and the channel layer.

[0014] The embodiments of the present invention provide a transmission control method and system based on cloud native network identification, which adopts a transmission control method based on cloud native network identification and adopts a layered design from the data level, transmission level, and channel level to solve the problems of low transmission efficiency and no service quality guarantee.

[0015] Other features and advantages of the present invention will be described in the following description, and partly become apparent from the description, or be understood by practicing the present invention. The purpose and other advantages of the present invention can be realized and obtained by the structures particularly pointed out in the written description, claims, and drawings. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] The drawings described herein are used to provide a further understanding of the present invention and constitute a part of the present invention. The exemplary embodiments of the present invention and their descriptions are used to explain the present invention and do not constitute an improper limitation of the present invention. In the drawings:

[0017] Figure 1 A schematic diagram of an implementation flow of a transmission control method based on a cloud native network identifier according to an embodiment of the present invention;

[0018] Figure 2A schematic diagram of an example of the working principle of the probability-free forward error correction provided according to an embodiment of the present invention;

[0019] Figure 3 A schematic diagram of a wireless link adaptive transmission optimization closed-loop control model provided according to an embodiment of the present invention;

[0020] Figure 4 A schematic diagram of transmission path calculation with multiple input conditions provided according to an embodiment of the present invention;

[0021] Figure 5 The figure is a schematic diagram of a backup link provided according to an embodiment of the present invention. DETAILED DESCRIPTION

[0022] In order to make the purpose, technical solutions and advantages of this specification more clear, the technical solutions of this specification will be clearly and completely described below in combination with the specific embodiments of this specification and the corresponding drawings. Obviously, the described embodiments are only part of the embodiments of this specification, not all of the embodiments. Based on the embodiments in this specification, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of this specification.

[0023] It should be noted that the terms "first", "second", etc. in the specification and claims of the present invention and the above-mentioned drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequence. It should be understood that the data used in this way can be interchanged where appropriate, so that the embodiments of the present invention described herein can be implemented in an order other than those illustrated or described herein. In addition, the terms "including" and "having" and any variations thereof are intended to cover non-exclusive inclusions, for example, a process, method, device, product or equipment that includes a series of steps or units is not necessarily limited to those steps or units clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products or equipment.

[0024] The technical solutions provided by the embodiments of this specification are described in detail below in conjunction with the accompanying drawings.

[0025] In order to improve the transmission capability of a cloud native network, an embodiment of the present invention provides a transmission control method and system based on a cloud native network identifier.

[0026] Figure 1 A schematic diagram of an implementation flow of a transmission control method based on a cloud native network identifier according to an embodiment of the present invention is shown in FIG. Figure 1 , the method comprising:

[0027] S100: For the data layer, the label reference algorithm is used to reduce the amount of repeated business data, and the virtual splicing technology is used to repeat the message header;

[0028] S102: For the transmission layer, perceive and analyze different channel types, adopt different transmission control algorithms, and implement dynamic switching of multiple channels according to changes in channel attributes during the transmission process;

[0029] S104: For the channel layer, path optimization, backup path remapping and multi-channel integrated transmission control are performed for different access methods;

[0030] S106: In the cloud native network, efficient transmission of information between the transmission network and the information network is achieved through layered optimization of the data layer, the transmission layer, and the channel layer.

[0031] It should be noted that in recent years, with the rise of cloud computing, big data, the Internet of Things, and artificial intelligence technologies, the traditional Internet's OTT model has become increasingly overwhelmed, and transmission networks and information networks have become increasingly difficult to adapt to business needs. As business demand grows, cloud service providers at the edge of the network are becoming larger and more concentrated, which makes the network structure increasingly unreasonable. A cloud-native network is a network that is born from the cloud, exists for the cloud, and is built on the cloud. Specifically, if the network is built based on the cloud rather than routers, the network mainly solves cloud service problems, and the main function of the network changes from the original end-to-end transmission connection to intelligent collaboration between the cloud and the edge.

[0032] In a specific implementation, for the data layer, using a label reference algorithm to reduce the amount of repeated service data, and using a virtual splicing technology to repeat the message header, specifically includes:

[0033] The data transmitted in the cloud-native network is repetitive (for example, different users at the same access point request the same resource). For large blocks of data, deduplication technology is used to reduce network resource consumption. The sender and receiver of the data cache the original data at the same time and replace the longer data with a shorter reference tag. When the duplicate data is transmitted again, only the reference tag needs to be transmitted. The receiver can restore the original data through the reference tag to avoid repeated transmission of large amounts of data.

[0034] For high-reliability small data, small data virtual splicing technology is adopted. The source node pre-processes a large number of high-reliability small data messages, splices messages with similar data volume and small time interval into virtual files, and packages the virtual files; the destination node unpacks the received virtual files and reorders them according to the address, sequence number and timestamp of the original message to ensure that the message order of the sender and the receiver is consistent.

[0035] It should be noted that with the expansion of interactive data types, the amount of data has increased exponentially, and a large amount of highly reliable small data has been generated. In order to optimize the transmission business process, reduce resource utilization, and improve transmission efficiency, these large amounts of highly reliable small data with low repetition frequency have become the focus of mining transmission optimization capabilities in efficient data transmission mechanisms.

[0036] In a specific implementation, the data transmitted in the cloud native network is repetitive. For large blocks of data, deduplication technology is used to reduce network resource consumption. The sender and receiver of the data cache the original data at the same time, and replace the longer data with a shorter reference tag. When the repeated data is transmitted again, only the reference tag needs to be transmitted, and the receiver can restore the original data through the reference tag to avoid repeated transmission of a large amount of data. The specific steps include:

[0037] Step 1: When sending for the first time, the tag database storing reference tags and original data is empty, and the original data is sent;

[0038] Step 2: The sender and the receiver use the same method to segment the original large block of data. The segment length is 48 to 96 bytes, and a reference tag is calculated for each segment. The reference tag length is 4 bytes. The reference tag and the original data are stored in the tag database. If the reference tags are the same but the original data are different, they are not stored.

[0039] Step 3: When sending the same data again, segment the data;

[0040] Step 4: For each segment, first search the label database. If the corresponding reference label and original data exist in the label database, send the label; otherwise, send the original data.

[0041] Step 5: When the receiver receives the reference tag, it searches for the corresponding original data from the tag database and merges all the original data into a large block of data before segmentation.

[0042] In a specific implementation, for high-reliability small data, a small data virtual splicing technology is used. The source node pre-processes a large number of high-reliability small data messages, splices messages with similar data volumes and small time intervals into virtual files, and packages the virtual files; the destination node unpacks the received virtual files and reorders them according to the address, sequence number and timestamp of the original message to ensure that the message order of the sending end and the receiving end is consistent. The specific steps include:

[0043] Step 1: When sending a small data request, the sender does not send it immediately, but caches the small data and sets a timer;

[0044] Step 2: If a request to send small data is received again within the scheduled time, the small data will continue to be cached;

[0045] Step 3: Repeat Step 2: until the timer times out or the total size of all cached small data exceeds the limit, then all cached small data are taken out and numbered in sequence;

[0046] Step 4: Extract the common part from the message header of each small data and keep only one common message header;

[0047] Step 5: Integrate the small data with the message header deleted into a large data, and attach a public message header to form a virtual file for sending;

[0048] Step 6: After receiving the virtual file, the receiver parses the original small data, adds a public message header to each small data, and submits it to the upper-level application in sequence according to the sequence number.

[0049] In a specific embodiment,

[0050] The different channel types include long-delay channels;

[0051] In view of the characteristics of the long-delay channel, based on the traditional congestion control algorithm, the window W of the new congestion control algorithm is calculated according to the following formula: t renew:

[0052]

[0053] In the formula, Where RTT is the link delay, RTT0 is the predetermined round trip delay; ρ*γ is the slow start threshold; t γ,0 is the time to exit the slow start phase; SS represents the slow start phase; CA represents the congestion avoidance phase.

[0054] It should be noted that the traditional congestion control algorithm has the disadvantages of slow transmission speed growth rate and low throughput on long-delay links. Therefore, it is necessary to adjust the slow start and congestion avoidance mechanisms of the traditional congestion control algorithm to achieve the same transmission speed growth rate and throughput for links with different delays but the same other parameters.

[0055] The window update algorithm formula of the traditional congestion control algorithm is as follows:

[0056]

[0057] SS (Slow Start) indicates the slow start phase, and CA (Congestion Avoidance) indicates the congestion avoidance phase. The window size update mechanism is to update the window size once each ACK is received, so that links with small delays update the window more frequently than links with large delays. The longer the link delay (RTT) is, the longer it stays in the slow start phase, and the congestion window of the link with small delay also has a great advantage in the congestion avoidance phase, which results in the throughput of the link with small delay being greater than that of the link with large delay.

[0058] In a specific embodiment,

[0059] The different channel types also include high error code channels;

[0060] In view of the characteristics of the high-error channel, a probability-free forward error correction code working mode is adopted, so that the receiving end does not need to receive specific data blocks, and only needs to receive a certain amount of data blocks for decoding. The sending end does not need to retransmit the lost data blocks. When there are multiple receiving ends, a large amount of retransmitted data can be reduced; the sending end continuously generates and sends coded blocks, and the receiving end only needs to notify the sending end after successful decoding, thereby reducing the amount of data for protocol interaction between the two ends.

[0061] How probability-free forward error correction codes work, see Figure 2 , the original data is divided into three equal parts: d1, d2 and d3, and the sender encodes them into s1 to s6, where the connection relationship represents XOR. For example, the data of s1 is d1, and the data of s5 is obtained by XORing the data of d2 and d3.

[0062] Assume that errors occur in s2, s3, and s5 during the transmission process. The receiving end receives s1, s4, and s6. Through s1, data block d1 is obtained. d2 can be obtained by XORing s4 and d1, and d3 can be obtained by XORing s6, d1, and d2. At this point, the receiving end has received all the blocks of the original data and can form the original data.

[0063] The specific description is as follows (^ represents XOR):

[0064] ① The sender encodes d1, d2, d3 to obtain s1, s2, s3, s4, s5, s6: s1 = d1, s2 = d2, s3 = d3, s4 = d1^d2, s5 = d2^d3, s6 = d1^d2^d3.

[0065] ② The receiving end receives s1, s4, and s6, and obtains: d1=s1. According to s4=d1^d2, d2=s4^d1. According to s6=d1^d2^d3, d3=s6^d1^d2 is obtained, and the original data d1, d2, and d3 are decoded.

[0066] It can be seen that the receiving end does not need to receive specific data blocks, as long as it can receive a certain amount of data blocks, it can decode, and the sending end does not need to retransmit the lost data blocks. When there are multiple receiving ends, a large amount of retransmitted data can be reduced.

[0067] Because it is a probability-free encoding method, the sender can continuously generate and send coded blocks, and the receiver only needs to notify the sender after successful decoding, reducing the amount of data exchanged between the two ends of the protocol.

[0068] It should be noted that for the high bit error characteristics of the radio link, using the traditional ACK mechanism to ensure the reliability of the data packet requires additional bandwidth and transmission time. The optimization method for high bit error links is to use forward error correction code technology to encode the data to offset the data packets lost when the link has a high bit error rate.

[0069] The traditional ACK mechanism is that the receiving end will feedback the sequence number of the packet discarded due to bit errors and other reasons to the sending end, and the sending end will retransmit the lost packet. When the bit error rate is high, the amount of feedback data during transmission is also large. On the other hand, during broadcast or multicast transmission, if any receiving end does not correctly receive the sent information, the sending end needs to retransmit the message. When there are many receiving ends and the link is poor, the amount of information retransmitted by the sending end will increase greatly, resulting in a significant decrease in transmission efficiency.

[0070] In a specific embodiment,

[0071] The different channel types also include narrowband channels;

[0072] In view of the characteristics of the narrowband channel, a protocol interaction mechanism is adopted to reduce bandwidth occupancy; in the protocol interaction mechanism, all data sent are valid data, and the protocol does not need to perform the TCP three-way handshake process to establish a connection. When data needs to be sent, data can be sent directly; when the receiving end receives the first data, a session is automatically established until the data transmission is completed; the protocol uses a negative acknowledgment mechanism NACK to confirm the data. When the sending end receives NACK, only the lost data is retransmitted, and the receiving end can reorder the data according to the sequence number to ensure that the data is in order.

[0073] It should be noted that for narrowband links, the traditional TCP protocol is basically unable to meet the transmission requirements of applications. The ACK (Acknowledgement) confirmation mechanism and retransmission mechanism of the TCP protocol increase the amount of data on the link, occupying the bandwidth of the narrowband link; in addition, the traditional TCP congestion control mechanism will seriously increase the data transmission delay in the case of deep queues, and cannot fully utilize the bandwidth. The present invention mainly reduces bandwidth occupancy from the protocol interaction mechanism.

[0074] In specific implementation, unlike the traditional reliable acknowledgment mechanism (ACK), this protocol uses a negative acknowledgment mechanism (NACK, Negative ACK) to confirm data. ACK needs to confirm most of the received data, while the negative acknowledgment mechanism only needs to confirm when data is lost, which can greatly reduce the amount of data in the reverse link. In addition, the ACK mechanism requires that when the sender receives the acknowledgment, if data loss is detected, in order to ensure the order of the data, it is necessary to retransmit from the lost data, even if the data received in the middle and not lost must be retransmitted. However, when the sender receives NACK, this protocol only retransmits the lost data, and the receiving end can reorder the data according to the sequence number to ensure that the data is in order. The negative acknowledgment mechanism ensures that each sent data is basically useful data, which saves the bandwidth of the narrowband link to a great extent.

[0075] When the sender sends a piece of data, it needs to fragment the data and number it, and then send it in sequence. At the same time, a query message is sent at regular intervals. The query message contains the maximum number QueryMax, that is, the query number is less than or equal to QueryMax. The loss of the fragments. Before receiving the complete data, the receiver needs to cache the received fragments. When receiving the query message, the receiver knows the lost fragments before QueryMax based on the received fragment numbers, and feeds back the lost fragment numbers to the sender through NACK for retransmission. When the sender has sent all the fragments, the receiver can receive most of the fragments through queries and NACKs. If there are still losses, retransmission will continue through regular queries. Finally, after the receiver receives all the fragments, it sorts them according to the fragment numbers and reassembles them into complete data content.

[0076] In a specific implementation, the multi-channel dynamic switching is implemented according to the change of channel attributes during the transmission process, specifically including:

[0077] According to the property changes of long-delay channels, high-error channels and narrowband channels during transmission, threshold intervals are set for the three protocols of long-delay, high-error and narrowband respectively. The threshold interval of each type of protocol includes multi-dimensional parameters of delay, bit error rate and bandwidth. During data transmission, the corresponding protocol feedbacks the comparison between the corresponding multi-dimensional parameters and the corresponding thresholds, which is used as the basis for decision-making protocol switching. If the parameters fed back by the protocol match the threshold interval of the currently used protocol, there is no need to switch the protocol. Otherwise, the parameters of the current channel are compared with the threshold intervals of different protocols, and the final protocol to be switched to is calculated according to the evaluation value. If the parameter is within the threshold interval, the evaluation value is positive, and the closer to the median of the interval, the higher the evaluation value. If it is outside the threshold interval, the evaluation value is negative, and the more deviated, the smaller the evaluation value. Finally, according to the comprehensive evaluation value, switch to the protocol with the largest evaluation value.

[0078] It should be noted that different transmission optimization protocols are designed for different channel characteristics. For example, a protocol is designed for long delay, high bit error, and narrowband channels respectively. However, the actual environment is often more complex and may have multiple characteristics at the same time.

[0079] During the transmission process, the protocol usually estimates the link bandwidth, delay and other parameters, and uses these parameters to adjust the transmission rate within the protocol. These parameters can reflect the changes in the underlying link more timely and accurately. However, due to the differences in the calculation methods and units of different protocol parameters, this information cannot be provided to the outside and is not fully utilized. By adding a protocol feedback mechanism, the internal parameters of different protocols can be converted into a unified unit, thereby providing the upper layer with more timely and accurate changes in link attributes. This information can be integrated into the network topology to provide a network resource situation that is closer to the actual situation, and can also serve as an important basis for adjusting transmission optimization strategies.

[0080] In the actual implementation, the link parameters are constantly changing, so the link parameters estimated within the protocol will also have certain fluctuations and errors. In order to avoid unnecessary impact of these fluctuations and errors on the upper layer, the protocol feedback parameters need to be smoothed and filtered. By counting and analyzing a large number of protocol feedback parameters, combined with the ideas of CUSUM algorithm and Kalman filtering, a protocol feedback outlier filtering mechanism is implemented. Through the internal feedback mechanism of the link protocol, when the link bandwidth is constantly changing, the algorithm can filter out the outliers generated in the protocol feedback process.

[0081] Link adaptive transmission optimization requires a mechanism to automatically adjust the system behavior to adapt to the dynamic characteristics of the network. Closed-loop control can solve the problem of adaptive adjustment. Loop control can basically be divided into two types: open-loop control and closed-loop control. Control theory only involves closed-loop control. In open-loop control, the control behavior of the controller is independent of the system output. In closed-loop control, the control behavior of the controller depends on the feedback formed by the system output.

[0082] See also Figure 3 In the closed-loop control model, the system output y(t) at a certain time t is fed back through the sensor F and then compared with the reference value r(t). The controller C obtains the difference e between the reference value and the output, and the control device P changes the system input u.

[0083] Wireless link adaptive transmission optimization uses closed-loop control functions in multiple locations to automatically adjust the system. Various transmission optimization protocols adaptively adjust the packet rate through feedback to improve transmission efficiency; data scheduling adaptively adjusts the transmission optimization protocol based on the link parameters fed back by the protocol, so that different protocols can work under adaptive conditions and achieve the best optimization effect.

[0084] In a specific implementation, the path optimization, backup path remapping and multi-channel integrated transmission control for different access means at the channel layer specifically include:

[0085] For the channel layer, communication resources are uniformly scheduled based on the transmission agent, and policy routing is used to provide smooth, flexible and reliable link transmission. The primary and backup link modes are pre-established, and fast switching can be completed through policy routing in case of link failure. The current network topology status, application routing strategy and service transmission requirements are used as input conditions for policy routing and provided to the routing algorithm. The input parameters are combined to calculate the end-to-end path through the routing algorithm, and the path scheduling of the routing switching equipment is carried out according to the routing results.

[0086] The routing algorithm selects the K shortest path algorithm used in the virtual task network mapping. For each underlying link, there is a corresponding set of backup paths, which are represented by a backup link set. The underlying link and the backup path have the same starting node and ending node. When a failure occurs, a suitable backup link is selected from the backup link set for link remapping. When an underlying link fails, a backup path is selected for it. The path has the same starting node and ending node as the failed link. Due to the connectivity of the network, a link has multiple backup paths to choose from.

[0087] In specific implementation, generally speaking, the bandwidth of mobile channels is small, and the information communication capacity is insufficient to support the transmission needs of various application services. Even if redundant channels are deployed, it is difficult to transmit services with high bandwidth requirements due to the constraints of transmission mechanisms such as the isolation of trunk links and optimal routing. It is necessary to adopt new integrated transmission methods such as multiple channel load sharing and parallel use, and realize automatic switching between links to meet the communication guarantee needs.

[0088] 1) Unified scheduling of communication resources based on transmission agents

[0089] The integrated use of wireless communication means is realized by the transmission agent, which mainly realizes the unified scheduling and intensive use of communication resources, completes the integration and interoperability of multiple services and end-to-end transmission optimization, and improves the ability to guarantee refined business services.

[0090] In order to fully utilize the resource advantages of heterogeneous links, strategic routing is used to better provide smooth, flexible and reliable link transmission. Link load balancing and strategic link control functions are provided to support work under the policy scheduling of the transmission agent; path transmission strategies are pre-set between adjacent nodes with heterogeneous multi-links and between the edge cloud and the terminal subnet.

[0091] In order to meet the needs of reliable business transmission, the primary and backup links can be established in advance, and fast switching can be completed through policy routing when a link fails. The primary and backup links provide quality detection, and the link with the best quality and meeting the business transmission quality requirements is selected as the primary link based on comprehensive metrics such as link delay, jitter time, packet loss rate, link failure, and channel quality. When the quality of the primary link deteriorates and cannot meet the original business quality transmission requirements, the business will be switched to the backup link to ensure normal forwarding of the business.

[0092] In order to meet the demand for multi-channel concurrent transmission due to insufficient single-link transmission capacity, the transmission agent issues a transmission control strategy, establishes a multi-link concurrent transmission link group between two adjacent communication nodes, uses multiple redundant backup links simultaneously and performs load balancing, and extends the bandwidth of adjacent transmission links. Round-robin scheduling or weighted forwarding scheduling is implemented on multiple links.

[0093] 2) Transmission path calculation based on multiple input conditions

[0094] See also Figure 4 The current network topology status, application routing strategy, and service transmission requirements are provided as input conditions for routing to the routing algorithm. Based on the input parameters, the routing algorithm performs end-to-end path calculation and schedules the routing switching equipment based on the routing results. Path scheduling requires the provision of service data flow characteristics (e.g., quintuples), and service data matching the characteristics will be forwarded in a controlled manner according to the dispatched scheduling strategy. During service transmission, the topology status is sensed, the topology status data is updated, and the latest parameters are provided for other services.

[0095] The routing strategy is used as the overall routing principle to define the evaluation value of the routing result. Depending on the routing strategy, the routing result can be evaluated based on the load of the path, the number of path hops, the maximum bandwidth of the path, the remaining bandwidth of the path, etc.

[0096] Business requirements are rigid requirements for route selection, and the route selection results that do not meet the requirements are eliminated. Business requirements include the minimum bandwidth, maximum delay, maximum jitter, maximum packet loss rate, etc. of the path. Depending on the type of service, some services only need to meet some conditions, while others need to meet all conditions.

[0097] 3) Backup path remapping mechanism

[0098] Wireless communication resources are limited, and there is no condition to provide dedicated backup means. The idea of ​​backup routing can be used to analyze the fault tolerance of virtual mapping networks. Since the path selection algorithm has a high algorithm complexity in the process of establishing the mapping relationship, there will be a high delay when remapping the faulty link, which will cause the loss of user data. Before the virtual request arrives, a set of backup paths is calculated for each underlying link using the path selection algorithm, which can shorten the remapping delay. The path selection algorithm can select the K shortest path algorithm used in the virtual task network mapping. For each underlying link, there is a set of corresponding backup paths, which can be represented by a backup link set. The underlying link and the backup path have the same starting node and ending node. When a fault occurs, a suitable backup link can be selected from the backup link set for link remapping. When an underlying link fails, a backup path must be selected for it. The path has the same starting node and ending node as the failed link. Due to the connectivity of the network, a link has multiple backup paths to choose from. In the underlying network, each additional hop will cause more bandwidth overhead and introduce more delay, so the hop count can be used as an indicator to measure the quality of the backup path. In order to improve the quality of the backup route set, the paths that meet the hop count standard are included in the backup route set.

[0099] See also Figure 5 , define a function P(G, t, h) = {xhop(x) = h, h ≥ 1}, which returns the set Px of paths with the same hop count h as the starting node of link t. As shown in the figure, P(G, AB, 2) = {ADB, ACB}, the set of backup paths with hop count 2 of link AB is the two paths ADB and ACB.

[0100] In addition, a function Cov(G, t, h) is defined, which returns the set of paths whose hop count of link t is less than or equal to h, where h is an integer greater than or equal to 1. For example, the set of backup paths for link AB whose hop count is less than or equal to 2 is Cov(G, AB, 2) = {AB, ADB, ACB}. Then, if the performance indicator h is defined, the set of backup links that meet the conditions can be expressed as C(G, t, h) = Cov(Gs, t, h) - P(G, t, 1). As shown in the figure, when h = 3, the backup link set of link AB is Cov(G, AB, 3), whose value is the set of paths with A and B as nodes and whose hop count is less than or equal to 3 minus the path AB itself. There are 3 groups in total, namely ACB, ADB and ACDB. When link AB fails, the traffic carried by AB can be carried by any path in the 3 groups.

[0101] The present invention provides a transmission control method based on cloud native network identification, which adopts a transmission control method based on cloud native network identification and is layeredly designed from the data level, transmission level, and channel level to solve the problems of low transmission efficiency and no service quality guarantee.

[0102] Based on the same inventive concept, the present invention also provides a transmission control system based on a cloud native network identifier, the system comprising:

[0103] The data layer optimization module is used to reduce the amount of duplicate business data through label reference algorithm and duplicate message headers through virtual splicing technology for the data layer;

[0104] The transport layer optimization module is used to perceive and analyze different channel types at the transport layer, adopt different transmission control algorithms, and realize dynamic switching of multiple channels according to changes in channel attributes during the transmission process;

[0105] The channel layer optimization module is used to perform path optimization, backup path remapping and multi-channel integrated transmission control for different access methods at the channel layer;

[0106] The information transmission module is used to achieve efficient transmission of information between the transmission network and the information network in the cloud native network through layered optimization of the data layer, the transmission layer and the channel layer.

[0107] In a specific implementation, the data layer optimization module is specifically used to:

[0108] The data transmitted in the cloud-native network is repetitive. For large blocks of data, deduplication technology is used to reduce network resource consumption. The sender and receiver of the data cache the original data at the same time and replace the longer data with a shorter reference tag. When the duplicate data is transmitted again, only the reference tag needs to be transmitted. The receiver can restore the original data through the reference tag to avoid repeated transmission of large amounts of data.

[0109] For high-reliability small data, small data virtual splicing technology is adopted. The source node pre-processes a large number of high-reliability small data messages, splices messages with similar data volume and small time interval into virtual files, and packages the virtual files; the destination node unpacks the received virtual files and reorders them according to the address, sequence number and timestamp of the original message to ensure that the message order of the sender and the receiver is consistent.

[0110] In a specific implementation, the data transmitted in the cloud native network is repetitive. For large blocks of data, deduplication technology is used to reduce network resource consumption. The sender and receiver of the data cache the original data at the same time, and replace the longer data with a shorter reference tag. When the repeated data is transmitted again, only the reference tag needs to be transmitted, and the receiver can restore the original data through the reference tag to avoid repeated transmission of a large amount of data. The specific steps include:

[0111] Step 1: When sending for the first time, the tag database storing reference tags and original data is empty, and the original data is sent;

[0112] Step 2: The sender and the receiver use the same method to segment the original large block of data. The segment length is 48 to 96 bytes, and a reference tag is calculated for each segment. The reference tag length is 4 bytes. The reference tag and the original data are stored in the tag database. If the reference tags are the same but the original data are different, they are not stored.

[0113] Step 3: When sending the same data again, segment the data;

[0114] Step 4: For each segment, first search the label database. If the corresponding reference label and original data exist in the label database, send the label; otherwise, send the original data.

[0115] Step 5: When the receiver receives the reference tag, it searches for the corresponding original data from the tag database and merges all the original data into a large block of data before segmentation.

[0116] In a specific implementation, for high-reliability small data, a small data virtual splicing technology is used. The source node pre-processes a large number of high-reliability small data messages, splices messages with similar data volumes and small time intervals into virtual files, and packages the virtual files; the destination node unpacks the received virtual files and reorders them according to the address, sequence number and timestamp of the original message to ensure that the message order of the sending end and the receiving end is consistent. The specific steps include:

[0117] Step 1: When sending a small data request, the sender does not send it immediately, but caches the small data and sets a timer;

[0118] Step 2: If a request to send small data is received again within the scheduled time, the small data will continue to be cached;

[0119] Step 3: Repeat Step 2: until the timer times out or the total size of all cached small data exceeds the limit, then all cached small data are taken out and numbered in sequence;

[0120] Step 4: Extract the common part from the message header of each small data and keep only one common message header;

[0121] Step 5: Integrate the small data with the message header deleted into a large data, and attach a public message header to form a virtual file for sending;

[0122] Step 6: After receiving the virtual file, the receiver parses the original small data, adds a public message header to each small data, and submits it to the upper-level application in sequence according to the sequence number.

[0123] In a specific embodiment,

[0124] The different channel types include long-delay channels;

[0125] In view of the characteristics of the long-delay channel, based on the traditional congestion control algorithm, the window W of the new congestion control algorithm is calculated according to the following formula: t renew:

[0126]

[0127] In the formula, Where RTT is the link delay, RTT0 is the predetermined round trip delay; ρ*γ is the slow start threshold; t γ,0 is the time to exit the slow start phase; SS represents the slow start phase; CA represents the congestion avoidance phase.

[0128] In a specific embodiment,

[0129] The different channel types also include high error code channels;

[0130] In view of the characteristics of the high-error channel, a probability-free forward error correction code working mode is adopted, so that the receiving end does not need to receive specific data blocks, and only needs to receive a certain amount of data blocks for decoding. The sending end does not need to retransmit the lost data blocks. When there are multiple receiving ends, a large amount of retransmitted data can be reduced; the sending end continuously generates and sends coded blocks, and the receiving end only needs to notify the sending end after successful decoding, thereby reducing the amount of data for protocol interaction between the two ends.

[0131] In a specific embodiment,

[0132] The different channel types also include narrowband channels;

[0133] In view of the characteristics of the narrowband channel, a protocol interaction mechanism is adopted to reduce bandwidth occupancy; in the protocol interaction mechanism, all data sent are valid data, and the protocol does not need to perform the TCP three-way handshake process to establish a connection. When data needs to be sent, data can be sent directly; when the receiving end receives the first data, a session is automatically established until the data transmission is completed; the protocol uses a negative acknowledgment mechanism NACK to confirm the data. When the sending end receives NACK, only the lost data is retransmitted, and the receiving end can reorder the data according to the sequence number to ensure that the data is in order.

[0134] In a specific implementation, the multi-channel dynamic switching is implemented according to the change of channel attributes during the transmission process, specifically including:

[0135] According to the property changes of long-delay channels, high-error channels and narrowband channels during transmission, threshold intervals are set for the three protocols of long-delay, high-error and narrowband respectively. The threshold interval of each type of protocol includes multi-dimensional parameters of delay, bit error rate and bandwidth. During data transmission, the corresponding protocol feedbacks the comparison between the corresponding multi-dimensional parameters and the corresponding thresholds, which is used as the basis for decision-making protocol switching. If the parameters fed back by the protocol match the threshold interval of the currently used protocol, there is no need to switch the protocol. Otherwise, the parameters of the current channel are compared with the threshold intervals of different protocols, and the final protocol to be switched to is calculated according to the evaluation value. If the parameter is within the threshold interval, the evaluation value is positive, and the closer to the median of the interval, the higher the evaluation value. If it is outside the threshold interval, the evaluation value is negative, and the more deviated, the smaller the evaluation value. Finally, according to the comprehensive evaluation value, switch to the protocol with the largest evaluation value.

[0136] In a specific implementation, the channel layer optimization module is specifically used to:

[0137] For the channel layer, communication resources are uniformly scheduled based on the transmission agent, and policy routing is used to provide smooth, flexible and reliable link transmission. The primary and backup link modes are pre-established, and fast switching can be completed through policy routing in case of link failure. The current network topology status, application routing strategy and service transmission requirements are used as input conditions for policy routing and provided to the routing algorithm. The input parameters are combined to calculate the end-to-end path through the routing algorithm, and the path scheduling of the routing switching equipment is carried out according to the routing results.

[0138] The routing algorithm selects the K shortest path algorithm used in the virtual task network mapping. For each underlying link, there is a corresponding set of backup paths, which are represented by a backup link set. The underlying link and the backup path have the same starting node and ending node. When a failure occurs, a suitable backup link is selected from the backup link set for link remapping. When an underlying link fails, a backup path is selected for it. The path has the same starting node and ending node as the failed link. Due to the connectivity of the network, a link has multiple backup paths to choose from.

[0139] The present invention provides a transmission control system based on cloud native network identification, which adopts a transmission control method based on cloud native network identification and is designed in layers from the data level, transmission level and channel level to solve the problems of low transmission efficiency and no guarantee of service quality.

[0140] Although the preferred embodiments of the present invention have been described, those skilled in the art may make additional changes and modifications to these embodiments once they have learned the basic creative concept. Therefore, the appended claims are intended to be interpreted as including the preferred embodiments and all changes and modifications that fall within the scope of the present invention.

[0141] Obviously, those skilled in the art can make various changes and modifications to the present invention without departing from the spirit and scope of the present invention. Thus, if these modifications and variations of the present invention fall within the scope of the claims of the present invention and their equivalents, the present invention is also intended to include these modifications and variations.

Claims

1. A transmission control method based on cloud native network identification, characterized in that: The method comprises: For the data layer, the label reference algorithm is used to reduce the amount of repeated business data, and the virtual splicing technology is used to repeat the message header; For the transmission layer, different channel types are perceived and analyzed, different transmission control algorithms are used, and dynamic switching of multiple channels is achieved according to changes in channel attributes during the transmission process; At the channel layer, path optimization, backup path remapping and multi-channel integrated transmission control are performed for different access methods; In a cloud-native network, efficient transmission of information between the transmission network and the information network is achieved through layered optimization of the data layer, transmission layer, and channel layer.

2. The method according to claim 1, characterized in that For the data layer, the label reference algorithm is used to reduce the amount of repeated business data, and the message header is repeated by virtual splicing technology, which specifically includes: The data transmitted in the cloud-native network is repetitive. For large blocks of data, deduplication technology is used to reduce network resource consumption. The sender and receiver of the data cache the original data at the same time and replace the longer data with a shorter reference tag. When the duplicate data is transmitted again, only the reference tag needs to be transmitted. The receiver can restore the original data through the reference tag to avoid repeated transmission of large amounts of data. For high-reliability small data, small data virtual splicing technology is adopted. The source node pre-processes a large number of high-reliability small data messages, splices messages with similar data volume and small time interval into virtual files, and packages the virtual files; the destination node unpacks the received virtual files and reorders them according to the address, sequence number and timestamp of the original message to ensure that the message order of the sender and the receiver is consistent.

3. The method according to claim 2, characterized in that The data transmitted in the cloud native network is repetitive. For large blocks of data, deduplication technology is used to reduce network resource consumption. The sender and receiver of the data cache the original data at the same time and replace the longer data with a shorter reference tag. When the repeated data is transmitted again, only the reference tag needs to be transmitted. The receiver can restore the original data through the reference tag to avoid repeated transmission of a large amount of data. The specific steps include: Step 1: When sending for the first time, the tag database storing reference tags and original data is empty, and the original data is sent; Step 2: The sender and the receiver use the same method to segment the original large block of data. The segment length is 48 to 96 bytes, and a reference tag is calculated for each segment. The reference tag length is 4 bytes. The reference tag and the original data are stored in the tag database. If the reference tags are the same but the original data are different, they are not stored. Step 3: When sending the same data again, segment the data; Step 4: For each segment, first search the label database. If the corresponding reference label and original data exist in the label database, send the label; otherwise, send the original data. Step 5: When the receiver receives the reference tag, it searches for the corresponding original data from the tag database and merges all the original data into a large block of data before segmentation.

4. The method according to claim 3, characterized in that For high-reliability small data, a small data virtual splicing technology is used. The source node pre-processes a large number of high-reliability small data messages, splices messages with similar data volume and small time interval into virtual files, and packages the virtual files; the destination node unpacks the received virtual files and reorders them according to the address, sequence number and timestamp of the original message to ensure that the message order of the sending end and the receiving end is consistent. The specific steps include: Step 1: When sending a small data request, the sender does not send it immediately, but caches the small data and sets a timer; Step 2: If a request to send small data is received again within the scheduled time, the small data will continue to be cached; Step 3: Repeat Step 2: until the timer times out or the total size of all cached small data exceeds the limit, then all cached small data are taken out and numbered in sequence; Step 4: Extract the common part from the message header of each small data and keep only one common message header; Step 5: Integrate the small data with the message header deleted into a large data, and attach a public message header to form a virtual file for sending; Step 6: After receiving the virtual file, the receiver parses the original small data, adds a public message header to each small data, and submits it to the upper-level application in sequence according to the sequence number.

5. The method according to claim 4, characterized in that The different channel types include long-delay channels; In view of the characteristics of the long-delay channel, based on the traditional congestion control algorithm, the window W of the new congestion control algorithm is calculated according to the following formula: t renew: In the formula, Where RTT is the link delay, RTT0 is the predetermined round trip delay; ρ*γ is the slow start threshold; t γ,0 is the time to exit the slow start phase; SS represents the slow start phase; CA represents the congestion avoidance phase.

6. The method according to claim 5, characterized in that The different channel types also include high error code channels; In view of the characteristics of the high-error channel, a probability-free forward error correction code working mode is adopted, so that the receiving end does not need to receive specific data blocks, and only needs to receive a certain amount of data blocks for decoding. The sending end does not need to retransmit the lost data blocks. When there are multiple receiving ends, a large amount of retransmitted data can be reduced; the sending end continuously generates and sends coded blocks, and the receiving end only needs to notify the sending end after successful decoding, thereby reducing the amount of data for protocol interaction between the two ends.

7. The method according to claim 6, characterized in that The different channel types also include narrowband channels; In view of the characteristics of the narrowband channel, a protocol interaction mechanism is adopted to reduce bandwidth occupancy; in the protocol interaction mechanism, all data sent are valid data, and the protocol does not need to perform the TCP three-way handshake process to establish a connection. When data needs to be sent, data can be sent directly; when the receiving end receives the first piece of data, a session is automatically established until the data transmission is completed; The protocol uses a negative acknowledgement mechanism NACK to acknowledge data. When the sending end receives NACK, only the lost data is retransmitted, and the receiving end can reorder the data according to the sequence number to ensure that the data is in order.

8. The method according to claim 7, characterized in that The method of implementing dynamic multi-channel switching according to changes in channel attributes during transmission specifically includes: According to the property changes of long-delay channels, high-error channels and narrowband channels during transmission, threshold intervals are set for the three protocols of long-delay, high-error and narrowband respectively. The threshold interval of each type of protocol includes multi-dimensional parameters of delay, bit error rate and bandwidth. During data transmission, the corresponding protocol feedbacks the comparison between the corresponding multi-dimensional parameters and the corresponding thresholds, which is used as the basis for decision-making protocol switching. If the parameters fed back by the protocol match the threshold interval of the currently used protocol, there is no need to switch the protocol. Otherwise, the parameters of the current channel are compared with the threshold intervals of different protocols, and the final protocol to be switched to is calculated according to the evaluation value. If the parameter is within the threshold interval, the evaluation value is positive, and the closer to the median of the interval, the higher the evaluation value. If it is outside the threshold interval, the evaluation value is negative, and the more deviated, the smaller the evaluation value. Finally, according to the comprehensive evaluation value, switch to the protocol with the largest evaluation value.

9. The method according to claim 8, characterized in that The channel layer performs path optimization, backup path remapping and multi-channel integrated transmission control for different access methods, specifically including: For the channel layer, communication resources are uniformly scheduled based on the transmission agent, and policy routing is used to provide smooth, flexible and reliable link transmission. The primary and backup link modes are pre-established, and fast switching can be completed through policy routing in case of link failure. The current network topology status, application routing strategy and service transmission requirements are used as input conditions for policy routing and provided to the routing algorithm. The input parameters are combined to calculate the end-to-end path through the routing algorithm, and the routing switching equipment is scheduled according to the routing results. The routing algorithm selects the K shortest path algorithm used in the virtual task network mapping. For each underlying link, there is a corresponding set of backup paths, which are represented by a backup link set. The underlying link and the backup path have the same starting node and ending node. When a failure occurs, a suitable backup link is selected from the backup link set for link remapping. When an underlying link fails, a backup path is selected for it. The path has the same starting node and ending node as the failed link. Due to the connectivity of the network, a link has multiple backup paths to choose from.

10. A transmission control system based on cloud native network identification, characterized in that: The transmission control method based on cloud native network identification according to any one of claims 1 to 9 is adopted, and the system includes: The data layer optimization module is used to reduce the amount of duplicate business data through label reference algorithm and duplicate message headers through virtual splicing technology for the data layer; The transport layer optimization module is used to perceive and analyze different channel types at the transport layer, adopt different transmission control algorithms, and realize dynamic switching of multiple channels according to changes in channel attributes during the transmission process; The channel layer optimization module is used to perform path optimization, backup path remapping and multi-channel integrated transmission control for different access methods at the channel layer; The information transmission module is used to achieve efficient transmission of information between the transmission network and the information network in the cloud native network through layered optimization of the data layer, the transmission layer and the channel layer.

Images