Rollbackable flow table compression method, flow table compression system, device, and medium

Abstract

The application discloses a rollbackable flow table compression method, a flow table compression system, equipment and a medium. The method comprises the following steps: defining flow tables of network equipment with all flow table utilization rates greater than a preset utilization rate as high-risk flow tables; predicting subsequent traffic of the high-risk flow tables according to current flow table item information and historical flow table item information; determining a compression strategy of the high-risk flow tables according to the subsequent traffic, obtaining corresponding compressed flow tables, and performing flow table update processing on corresponding network equipment according to the compressed flow tables; in response to forwarding consultation information sent by the network equipment according to the updated compressed flow tables, calculating a negative influence value of a compressed flow table item in the compressed flow table; and performing flow table item rollback processing on the compressed flow table item with the negative influence value greater than a preset influence value. According to the application, the subsequent traffic is predicted to improve the flow table compression effect, and the rollback operation of the compressed flow table item can be performed to slow down the negative influence caused by the flow table compression.

Description

Technical Field

[0001] This application relates to the field of information technology, specifically to a rollback-capable flow table compression method, flow table compression system, device, and medium. Background Technology

[0002] In traditional networks, routers have the dual functions of data forwarding and path decision-making. However, Software-Defined Networking (SDN) introduces an innovative concept: the separation of the control plane and the data plane. In an SDN network, the forwarding paths of all data flows are managed by a centralized SDN controller. This controller is responsible for calculating routing tables and distributing them to OpenFlow switches in the network, enabling the switches to forward data flows based on the routing tables. The advantage of this approach is that it allows for the independent development of the data plane and the control plane. Furthermore, due to the involvement of high-level control programs, path planning is more rational and easier to meet complex requirements.

[0003] The above approach leads to a situation where, after the switch loses its independent decision-making ability, the forwarding of all packets can only be determined by querying the flow table. In complex network environments, as data traffic increases, the flow table entries in the OpenFlow switch will continue to grow, which may eventually cause the flow table to overflow, thus affecting data scheduling, causing SDN controller overload, or even the entire network to crash. Currently, there are some flow table compression schemes that can compress and redirect flow table entries. However, since current schemes generally rely only on real-time flow table information for compression, the compression effect is poor. Moreover, current schemes cannot roll back after flow table entries are compressed, which can lead to the SDN controller being unable to replan the routed data flow path after unreasonable scheduling decisions are made, thereby affecting the overall efficiency and reliability of the network. Summary of the Invention

[0004] This application provides a rollback-enabled flow table compression method, flow table compression system, device, and medium. At least, it can ensure that while improving the flow table compression effect by predicting subsequent flow rates, the solution can also perform rollback operations on compressed flow table entries to mitigate the negative impacts caused by flow table compression.

[0005] In a first aspect, embodiments of this application provide a fallback-enabled flow table compression method, comprising:

[0006] Retrieve information about the current flow table entries contained in the flow tables of multiple network devices;

[0007] Calculate the flow table utilization rate of multiple network devices based on the current flow table entry information;

[0008] Define the flow tables of all network devices whose flow table usage rate is greater than the preset usage rate as high-risk flow tables.

[0009] Obtain historical flow table entry information for the high-risk flow table;

[0010] The subsequent traffic of the high-risk flow table is predicted based on the current flow table entry information and the historical flow table entry information;

[0011] The compression strategy of the high-risk flow table is determined based on the subsequent traffic, the corresponding compressed flow table is obtained, and the flow table update process is performed on the corresponding network device based on the compressed flow table.

[0012] In response to a forwarding consultation message sent by the network device based on the updated compressed flow table, the negative impact value of the compressed flow table entry is calculated.

[0013] For compressed flow table entries whose negative impact value is greater than a preset impact value, flow table entry rollback processing is performed.

[0014] In some embodiments, predicting the subsequent traffic of the high-risk flow table based on the current flow table entry information and the historical flow table entry information includes:

[0015] If the current data flow corresponding to the current flow table entry is less than the prediction threshold, calculate the average data flow corresponding to the current flow table entry and the historical flow table entry, and determine the average data flow as the subsequent flow of the high-risk flow table.

[0016] If the current data flow is greater than or equal to the prediction threshold, the current flow table entry information and the historical flow table entry information are input into the LSTM model for prediction processing to obtain the subsequent flow of the high-risk flow table.

[0017] In some embodiments, determining the compression strategy of the high-risk flow table based on the subsequent traffic to obtain the corresponding compressed flow table includes:

[0018] Based on the subsequent traffic, the matching information, output information, and priority information of each flow table entry in the high-risk flow table are obtained;

[0019] The compression strategy for the high-risk flow table is determined based on the matching information;

[0020] Based on the corresponding compression strategy, the output information, and the priority information, the high-risk flow table is compressed to obtain a compressed flow table.

[0021] In some embodiments, determining the compression strategy for the high-risk flow table based on the matching item information includes:

[0022] All matching items in the high-risk flow table are determined based on the matching item information;

[0023] By selecting combinations of matching items from all the aforementioned matching items, various compression and combination methods can be obtained;

[0024] The compression efficiency of all the compression combinations is calculated based on a preset performance function;

[0025] The compression combination method that yields the greatest compression benefit is determined to be the compression strategy for the high-risk flow table.

[0026] In some embodiments, the step of compressing the high-risk flow table according to the corresponding compression strategy, the output information, and the priority information to obtain a compressed flow table includes:

[0027] Wildcard replacement processing is performed on all matching items corresponding to the compression strategy in the high-risk flow table, and the output port of the flow table item after wildcard replacement is adjusted according to the output information to obtain the wildcard replacement flow table.

[0028] Adjust the priority of the wildcard replacement entries in the flow table according to the priority information to obtain a compressed flow table.

[0029] In some embodiments, calculating the compression efficiency of all compression combinations based on a preset performance function includes:

[0030] Obtain the number of flow table entries before compression and the number of flow table entries after compression of the compression combination method;

[0031] The flow table compression ratio of the compression combination method is calculated based on the number of flow table entries before compression and the number of flow table entries after compression.

[0032] Obtain the flow table entry matching traffic and flow table entry error traffic after compression according to the described compression combination method;

[0033] The compression accuracy of the compression combination method is calculated based on the flow table entry matching traffic and the flow table entry error traffic;

[0034] The compression efficiency of the compression combination method is calculated based on the flow table compression ratio and the compression accuracy.

[0035] In some embodiments, the step of calculating the negative impact value of a compressed flow table entry in the compressed flow table in response to a forwarding consultation message sent by the network device based on the updated compressed flow table includes:

[0036] In response to the forwarding consultation information sent by the network device based on the updated compressed flow table, the current traffic path and the optimal traffic path corresponding to the compressed flow table entry in the compressed flow table are calculated;

[0037] Calculate the current path cost corresponding to the current traffic path and the optimal path cost corresponding to the optimal traffic path based on the corresponding path hop count.

[0038] The negative impact value of the compressed flow table entry is calculated based on the current path cost and the optimal path cost.

[0039] In some embodiments, calculating the negative impact value of the compressed flow table entry based on the current path cost and the optimal path cost includes:

[0040] Obtain the entry transmission rate of the compressed flow table entry and the error traffic transmission rate corresponding to the forwarding advisory information;

[0041] Calculate the first product of the current path cost and the erroneous traffic transmission rate, and the second product of the optimal path cost and the entry transmission rate. Divide the first product by the second product to obtain the negative impact value of the compressed flow entry.

[0042] In some embodiments, the flow table entry rollback process for compressed flow table entries whose negative impact value is greater than a preset impact value includes:

[0043] Obtain the optimal flow path corresponding to the compressed flow table entry whose negative impact value is greater than a preset impact value;

[0044] The fallback flow table entry corresponding to the optimal traffic path is sent to the network device so that the network device can replace the compressed flow table entry with the fallback flow table entry.

[0045] Secondly, embodiments of this application provide a fallback flow table compression system. The system is applied to an SDN controller in an SDN network. The SDN network further includes multiple network devices that are communicatively connected to the SDN controller and mutually connected. The system includes:

[0046] The traffic prediction module is used to obtain current flow table entry information contained in the flow tables of multiple network devices; calculate the flow table utilization rate of the multiple network devices based on the current flow table entry information; define the flow tables of all network devices whose flow table utilization rate is greater than a preset utilization rate as high-risk flow tables; obtain historical flow table entry information of the high-risk flow tables; and predict the subsequent traffic of the high-risk flow tables based on the current flow table entry information and the historical flow table entry information.

[0047] The flow table compression module is used to determine the compression strategy of the high-risk flow table based on the subsequent traffic, obtain the corresponding compressed flow table, and perform flow table update processing on the corresponding network device based on the compressed flow table.

[0048] The flow entry rollback module is used to respond to the forwarding consultation information sent by the network device based on the updated compressed flow table, calculate the negative impact value of the compressed flow table entry, and perform flow entry rollback processing on the compressed flow table entry whose negative impact value is greater than a preset impact value.

[0049] Thirdly, embodiments of this application provide an electronic device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the rollbackable flow table compression method as described in any embodiment of the first aspect.

[0050] Fourthly, embodiments of this application provide a computer-readable storage medium storing computer-executable instructions for performing a rollback flow table compression method as described in any of the embodiments of the first aspect.

[0051] This application has at least the following beneficial effects: First, by obtaining the current flow table entry information contained in the flow tables of multiple network devices and calculating the flow table utilization rate of multiple network devices based on the current flow table entry information, this application defines the flow tables of all network devices with a flow table utilization rate greater than a preset utilization rate as high-risk flow tables. This process enables the determination of whether the flow tables of network devices are in a high-risk state, thereby allowing for targeted flow table management of high-risk flow tables and effectively preventing flow table overflow. Furthermore, by obtaining historical flow table entry information of high-risk flow tables to predict their subsequent traffic, this application can formulate compression strategies based on a thorough understanding of the network conditions. It selects the most suitable compression strategy based on future data traffic trends and obtains the corresponding compressed flow tables. Based on the compressed flow tables, it updates the flow tables of the corresponding network devices, replacing some flow table entries in the high-risk flow tables with compressed flow table entries in the compressed flow tables. This ensures both the efficiency and effectiveness of flow table compression. This application reduces the impact of flow table compression on data forwarding efficiency, thereby effectively improving the flow table compression effect. Furthermore, after updating the flow tables of corresponding network devices based on the compressed flow tables, it also calculates the negative impact value of compressed flow table entries in response to forwarding consultation information sent by the devices. This allows for real-time monitoring of the performance of the compressed flow tables and assessment of potential negative impacts, providing a selection basis for subsequent flow table entry rollback processing. Furthermore, by performing flow table entry rollback processing on compressed flow table entries with negative impact values ​​exceeding a preset value, this application enables the rollback operation of compressed flow table entries. This allows for the replanning of data flow routes when flow table compression causes negative impacts, preventing serious consequences due to unreasonable compression decisions. In summary, this application can improve the flow table compression effect, solve the flow table overflow problem, and mitigate the negative impact of flow table compression on data flow scheduling, effectively improving the overall network reliability and stability. Attached Figure Description

[0052] Figure 1 This is a flowchart of a fallback flow table compression method proposed in an embodiment of this application;

[0053] Figure 2 A flowchart of a fallback flow table compression method proposed in another embodiment of this application;

[0054] Figure 3 This is a flowchart illustrating an example of predictive processing according to another embodiment of this application;

[0055] Figure 4 A flowchart illustrating a fallback flow table compression method proposed in another embodiment of this application;

[0056] Figure 5 A flowchart illustrating a fallback flow table compression method proposed in another embodiment of this application;

[0057] Figure 6 A flowchart illustrating a fallback flow table compression method proposed in another embodiment of this application;

[0058] Figure 7 A flowchart illustrating a fallback flow table compression method proposed in another embodiment of this application;

[0059] Figure 8 Here is an example flowchart of a flow table entry rollback process proposed in another embodiment of this application;

[0060] Figure 9 This is a schematic diagram of the structure of an SDN network proposed in another embodiment of this application;

[0061] Figure 10 This is a schematic diagram of the structure of an electronic device according to another embodiment of this application. Detailed Implementation

[0062] To make the objectives, technical solutions, and advantages of this application clearer, the following detailed description is provided in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the scope of this application.

[0063] In some embodiments, although functional modules are divided in the system diagram and a logical order is shown in the flowchart, in some cases, the steps shown or described may be performed in a different order than the module division in the system or the order in the flowchart. The terms "first," "second," etc., in the specification, claims, and the foregoing drawings are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence.

[0064] Furthermore, unless otherwise explicitly specified and limited, the term "connection / linkage" should be interpreted broadly, for example, it can be a fixed connection or a movable connection, a detachable connection or a non-detachable connection, or an integral connection; it can be a mechanical connection or an electrical connection or a connection that can communicate with each other; it can be a direct connection or an indirect connection through an intermediate medium.

[0065] In the description of the embodiments in this application, the terms "one embodiment / implementation," "another embodiment / implementation," or "some embodiments / implementations," "in the above embodiments / implementations," etc., refer to specific features, structures, materials, or characteristics described in conjunction with embodiments or examples that are included in at least two embodiments or implementations disclosed in this application. In this disclosure, the illustrative expressions of the above terms do not necessarily refer to the same embodiment or implementation. It should be noted that although a logical order is shown in the flowcharts, in some cases, the steps shown or described may be performed in a different order than that shown in the flowcharts.

[0066] In traditional networks, routers have the dual functions of data forwarding and path decision-making. However, Software-Defined Networking (SDN) introduces an innovative concept: the separation of the control plane and the data plane. In an SDN network, the forwarding paths of all data flows are managed by a centralized SDN controller. This controller is responsible for calculating routing tables and distributing them to OpenFlow switches in the network, enabling the switches to forward data flows based on the routing flow tables. The advantage of this approach is that it allows for the independent development of the data plane and control plane. Furthermore, due to the involvement of high-level control programs, path planning is more rational and better able to meet complex requirements. However, this approach also leads to a situation where, after a switch loses its independent decision-making ability, the forwarding of all packets can only be determined by querying the flow table. In complex network environments, as data traffic increases, the flow table entries within the OpenFlow switches will continue to grow, potentially causing the flow table to overflow. This can affect data scheduling, leading to SDN controller overload or even network collapse.

[0067] Among related technologies, there are several flow table compression schemes that can compress and redirect flow table entries. Specifically, there are currently memory access technologies such as TCAM, QDR, and SRAM. Although these technologies can alleviate the problem, the increase in capacity still lags far behind the growth rate of flow table entries. In addition to developing memory technologies, there are also solutions such as FTS and Q-Flag, which include flow table entry compression and redirection. However, these solutions generally rely solely on real-time flow table information for compression. That is, these solutions can only perform compression based on existing flow table entries and cannot predict subsequent traffic, resulting in poor compression performance. Furthermore, current solutions cannot roll back after flow table entries are compressed. This can lead to unreasonable scheduling decisions, making it impossible for the SDN controller to replan the routed data flow. As a result, an unreasonable compression decision may prevent the entire network from using the optimal scheduling scheme, thereby affecting the overall efficiency and reliability of the network.

[0068] To at least address the aforementioned issues, this application discloses a rollback-enabled flow table compression method, system, device, and medium. The method includes defining flow tables of all network devices with flow table utilization rates exceeding a preset utilization rate as high-risk flow tables; predicting subsequent traffic for high-risk flow tables based on current and historical flow table entry information; determining a compression strategy for the high-risk flow tables based on subsequent traffic to obtain corresponding compressed flow tables, and updating the flow tables of the corresponding network devices based on the compressed flow tables; calculating the negative impact value of compressed flow table entries in the compressed flow tables in response to forwarding consultation information sent by the network devices based on the updated compressed flow tables; and performing flow table entry rollback processing on compressed flow table entries with negative impact values ​​exceeding a preset impact value. This solution improves flow table compression effectiveness by predicting subsequent data flows through the switch and continuously monitors affected data flows. When the negative impact value exceeds a certain threshold, the flow table entry is reset, enabling rollback of compressed flow table entries. This avoids suboptimal scheduling issues caused by compression and mitigates the negative impact of flow table compression.

[0069] The embodiments of this application will be further described below with reference to other accompanying drawings.

[0070] Please refer to Figure 1 , Figure 1 This is a flowchart of a fallback flow table compression method according to an embodiment of this application; in a first aspect, this application provides a fallback flow table compression method, which includes at least the following steps S110 and S180:

[0071] Step S110: Obtain the current flow table entry information contained in the flow tables of multiple network devices;

[0072] Step S120: Calculate the flow table utilization rate of multiple network devices based on the current flow table entry information;

[0073] Step S130: Define the flow tables of all network devices whose flow table usage rate is greater than the preset usage rate as high-risk flow tables;

[0074] Step S140: Obtain historical flow table entry information for high-risk flow tables;

[0075] Step S150: Predict the subsequent flow of high-risk flow tables based on the current flow table entry information and historical flow table entry information;

[0076] Step S160: Determine the compression strategy for high-risk flow tables based on subsequent traffic, obtain the corresponding compressed flow tables, and update the flow tables of the corresponding network devices based on the compressed flow tables.

[0077] Step S170: In response to the forwarding consultation information sent by the network device based on the updated compressed flow table, calculate the negative impact value of the compressed flow table entry;

[0078] Step S180: Perform flow table rollback processing on compressed flow table entries whose negative impact value is greater than the preset impact value.

[0079] In some embodiments, based on the above steps S110 and S180, this application obtains the current flow table entry information contained in the flow tables of multiple network devices, calculates the flow table utilization rate of multiple network devices based on the current flow table entry information, and defines the flow tables of all network devices with a flow table utilization rate greater than a preset utilization rate as high-risk flow tables. This process enables the determination of whether the flow tables of network devices are in a high-risk state, thereby enabling targeted flow table management of high-risk flow tables and effectively preventing flow table overflow. Furthermore, this application obtains historical flow table entry information of high-risk flow tables to predict their subsequent traffic, thereby enabling the formulation of compression strategies based on a full understanding of the network conditions. It selects the most suitable compression strategy based on future data traffic trends and obtains the corresponding compressed flow tables. Based on the compressed flow tables, it updates the flow tables of the corresponding network devices, replacing some flow table entries in the high-risk flow tables with compressed flow table entries in the compressed flow tables, ensuring the efficiency and effectiveness of flow table compression. Simultaneously, this application reduces the impact of flow table compression on data forwarding efficiency, thereby effectively improving the flow table compression effect. Furthermore, after updating the flow tables of the corresponding network devices based on the compressed flow tables, this application also calculates the negative impact value of the compressed flow table entries in response to the forwarding consultation information sent by the devices, so as to monitor the performance of the compressed flow table in real time and evaluate the negative impact value of possible negative impacts, providing a selection basis for subsequent flow table entry rollback processing. Furthermore, this application implements the rollback operation of compressed flow table entries by performing flow table entry rollback processing on compressed flow table entries with negative impact values ​​greater than the preset impact value. This allows for the replanning of the data flow routes when flow table compression causes negative impacts, preventing serious consequences caused by unreasonable compression decisions. In summary, this application can improve the flow table compression effect, solve the flow table overflow problem, and mitigate the negative impact of flow table compression on data flow scheduling, effectively improving the overall network reliability and stability.

[0080] Specifically, in some embodiments, the fallback flow table compression system corresponding to the method of this application is applied to the SDN controller in the SDN network. The SDN network also includes multiple network devices that are communicatively connected to the SDN controller and have communication connections with each other. The network devices include OpenFlow switches. Corresponding to steps S110 to S130, flow table entries in the SDN network are all generated and distributed by the SDN controller. Therefore, the method of this application can record the number of flow table entries distributed to each OpenFlow switch through the SDN controller to basically understand the usage of flow tables of each OpenFlow switch. When the usage rate of flow tables exceeds a certain threshold, corresponding to steps S140 to S150, this application triggers the process to enter the traffic prediction module. The SDN controller can use the dump-flows command to obtain the traffic information of each flow table entry in the high-risk flow table and use the LSTM model to predict the size of the subsequent traffic of the flow table entry. Long Short-Term Memory (LSTM) The Memory (LSTM) model is a recurrent neural network (RNN) that can be used for traffic prediction. In step S160, after the traffic size prediction is completed, the application will pass the prediction result to the flow table compression module. Here, the application will use heuristic decision-making to select the optimal compression strategy. Then, the application will compress the flow table entries in high-risk flow tables according to the compression strategy, and use the OpenFlow protocol to distribute new flow table entries and delete old flow table entries to the OpenFlow switch. In steps S170 to S180, after the flow table compression is completed, the application will continue to observe the impact of the flow table compression. If the negative impact caused by the compression exceeds a threshold, a rollback operation will be performed on the affected flow table entries to mitigate the losses caused by the flow table compression.

[0081] Please refer to Figure 2 , Figure 2 This is a flowchart of a fallback flow table compression method proposed in another embodiment of this application; in some embodiments, predicting the subsequent flow of high-risk flow tables based on current flow table entry information and historical flow table entry information includes at least the following steps S210 and S220:

[0082] Step S210: If the current data flow corresponding to the current flow table entry information is less than the prediction threshold, calculate the average data flow corresponding to the current flow table entry information and the historical flow table entry information, and determine the average data flow as the subsequent flow of the high-risk flow table.

[0083] Step S220: If the current data flow is greater than or equal to the prediction threshold, input the current flow table entry information and the historical flow table entry information into the LSTM model for prediction processing to obtain the subsequent flow of the high-risk flow table.

[0084] In some embodiments, corresponding to steps S210 and S220, in order to enable the flow table compression module to make the most favorable decision for the future, the flow table compression module will pass the flow table entry statistics information obtained from the monitoring module to the traffic prediction module, so that it can predict which flow table entries are more important in the future, so as to give priority when selecting compression strategies. Specifically, the solution of this application can use the LSTM algorithm, and use the LSTM model to combine historical traffic data to predict the changes in future traffic. The traffic prediction process can include two parts: data collection and preprocessing, and the result prediction process using the model.

[0085] Please refer to Figure 3 , Figure 3 Here is a flowchart of an example of predictive processing proposed in another embodiment of this application; specifically, the predictive processing process includes the following embodiments: In some embodiments, since the system corresponding to the method of this application is applied to the SDN controller, unlike the traditional network, the generation of flow table entries in the SDN network all comes from the SDN controller. As the neural hub of the SDN network, it can easily count the usage of flow tables in each switch to obtain the current flow table entry information corresponding to the current traffic information and the historical flow table entry information corresponding to the historical data, and store them in the form of a flow table statistical storage structure, as shown in Table (1).

[0086] Table (1) Statistical Structure of Flow Items

[0087] Fields illustrate Switch Flow table matching items Count Number of flow table entries Table_name Flow table name Flow_limit Upper limit of flow table entries

[0088] Furthermore, the flow table utilization rate of multiple network devices can be calculated based on the current flow table entry information. Flow tables of all network devices with a flow table utilization rate greater than a preset utilization rate are defined as high-risk flow tables. Specifically, the preset utilization rate can be set to 70%. In subsequent steps, this application can set switches that exceed the threshold as high-risk switches, and the flow tables of high-risk switches are high-risk flow tables.

[0089] In some embodiments, after identifying a high-risk flow table, the historical flow table entry information of the high-risk flow table is obtained, including: periodically probing and obtaining the existing flow table entries in the switch and the transmission status of the data flow in each flow table entry according to a preset time interval, and obtaining the historical flow table entry information of the high-risk switch. The preset time interval can be 5 minutes, and the information is saved to the flow table entry statistics storage structure shown in Table (2).

[0090] Table (2) Flow Table Entry Statistics Storage Structure

[0091]

[0092] The flow table entry statistics storage structure is a hash structure. The key value (flow table matching item, the switch to which the flow table belongs, and the flow table name) in the hash structure is used to uniquely represent the flow table entry, and the value value (statistical time and the amount of traffic flowing through within the statistical time) is the traffic statistics data of that flow table entry. When the statistical data of the corresponding historical flow table entry information exceeds a set limit, such as a limit of 10 statistical data entries (i.e., the historical flow table entry information includes 10 statistical data entries), the statistical information will be preprocessed to facilitate subsequent flow table entry traffic prediction. The preprocessing process may include obtaining the statistical information of the data items, classifying the statistical information by flow table entry, normalizing the predicted data, and saving it.

[0093] In some embodiments, based on the above-described flow entry statistical storage structure, in order to ensure that the selected compression strategy achieves maximum effectiveness in the future, this application needs to predict the subsequent traffic of each flow entry in the high-level switch. To meet the above requirements, this application can use the LSTM algorithm combined with the historical traffic data of each flow entry to predict the subsequent traffic size of the flow entry. In a large data center network, a high-risk switch often has a large number of flow entries. If a full prediction method is adopted, it will consume a lot of computing resources. In order to improve the prediction efficiency and reduce the prediction time, this application not only uses a concurrent method but also performs prediction of multiple flow entries simultaneously; it can also By setting a prediction threshold, if the current data flow is greater than or equal to the prediction threshold, the current flow entry information and historical flow entry information are normally input into the LSTM model for prediction processing to obtain the subsequent flow of high-risk flow entries. If the current data flow corresponding to the current flow entry information is less than the prediction threshold, the average data flow corresponding to the current flow entry information and historical flow entry information is calculated, and the average data flow is determined as the subsequent flow of high-risk flow entries. That is, when the data flow matched by a certain flow entry within the observation time is less than the prediction threshold, the prediction size will be given directly based on the average data flow of the corresponding historical average value, instead of using the prediction of the LSTM model, so as to improve the calculation speed of subsequent flow.

[0094] It is conceivable that combining current traffic information with historical data can effectively improve the accuracy of prediction results. Therefore, in this application, the traffic statistics information of the last 10 times, including current traffic information and historical data, can be passed to the LSTM traffic prediction model for prediction. After obtaining the traffic prediction result of the flow table item, this application will store it in the format of the prediction result storage structure in Table (3) and pass it to the flow table item compression module so that it can use this as a basis for subsequent compression strategy selection. Table (3) is shown below.

[0095] Table (3) Prediction result storage structure

[0096] Fields illustrate Flow_num Flow table entry ID Throughput Predicted traffic size (bytes)

[0097] Please refer to Figure 4 , Figure 4 This is a flowchart of a fallback flow table compression method proposed in another embodiment of this application; in some embodiments, a compression strategy for high-risk flow tables is determined based on subsequent traffic to obtain the corresponding compressed flow tables, including at least the following steps S410 and S430:

[0098] Step S410: Obtain the matching information, output information, and priority information of each flow table entry in the high-risk flow table based on the subsequent traffic.

[0099] Step S420: Determine the compression strategy for high-risk flow tables based on the matching item information;

[0100] Step S430: Based on the corresponding compression strategy, output information and priority information, compress the high-risk flow table to obtain the compressed flow table.

[0101] In some embodiments, the flow table entry compression module corresponding to steps S410 to S430 can use a heuristic compression strategy. By analyzing the current situation and future data flow, the most favorable compression strategy for the future is found. The specific process includes parameter acquisition: before performing flow table compression, collecting and calculating flow table information and statistical information of each flow table entry within the flow table. Since a flow table is composed of flow table entries, this application can represent R as a flow table within an OpenFlow switch, and a flow table entry r∈R is defined as r = <M r O r ,P r >, where M r O represents the matching item in the flow table entry. r P represents the output of a flow table entry. r This indicates the priority of the flow table entry, while the matching entry M... r It is then defined as Where n represents the number of matching items, the flow table information is obtained by the dump-flows command in this application, and the statistical information of each flow table item can be obtained from the subsequent traffic data stored in the form of prediction result storage structure given by the previous traffic prediction module.

[0102] Please refer to Figure 5 , Figure 5 This is a flowchart of a fallback flow table compression method proposed in another embodiment of this application; in some embodiments, determining the compression strategy for high-risk flow tables based on matching information includes at least the following steps S510 and S540:

[0103] Step S510: Determine all matching items in the high-risk flow table based on the matching item information;

[0104] Step S520: By selecting combinations of matching items from all matching items, multiple compression combination methods are obtained;

[0105] Step S530: Calculate the compression efficiency of all compression combinations based on the preset performance function;

[0106] Step S540: Determine the compression combination method with the greatest compression benefit as the compression strategy of high-risk flow tables.

[0107] It is conceivable that, corresponding to steps S510 and S540, in order to enable the flow table entries to be compressed, this application adopts the method of replacing some matching fields in the flow table entries with wildcards * to reduce the number of flow table entries required. Specifically, if there are n types of matching items in the flow table entries, then when choosing to use wildcards to replace the original matching items for flow table compression, a total of 2^n combinations are obtained. At the same time, since the number of matching item types n supported by the OpenFlow switch is a constant, for a set of flow table entries, this application can perform an exhaustive search in polynomial time to determine the matching item combination used for wildcard replacement, and thus find the optimal strategy for flow rule compression.

[0108] Please refer to Figure 6 , Figure 6 The flowchart below shows a fallback flow table compression method proposed in another embodiment of this application. In some embodiments, high-risk flow tables are compressed according to a corresponding compression strategy, output information, and priority information to obtain compressed flow tables, including at least the following steps S610 and S620:

[0109] Step S610: Perform wildcard replacement processing on all matching items corresponding to the compression strategy in the high-risk flow table, and adjust the output port of the flow table item after wildcard replacement according to the output information to obtain the wildcard replacement flow table.

[0110] Step S620: Adjust the priority of wildcard replacement entries in the flow table according to the priority information to obtain a compressed flow table.

[0111] Please refer to Figure 7 , Figure 7 This is a flowchart of a fallback flow table compression method proposed in another embodiment of this application; in some embodiments, the compression efficiency of all compression combinations is calculated according to a preset performance function, including at least the following steps S710 and S750:

[0112] Step S710: Obtain the number of flow table entries before compression and the number of flow table entries after compression for the compression combination method;

[0113] Step S720: Calculate the flow table compression rate of the compression combination method based on the number of flow table entries before compression and the number of flow table entries after compression;

[0114] Step S730: Obtain the flow table entry matching flow and flow table entry error flow after compression according to the compression combination method;

[0115] Step S740: Calculate the compression accuracy of the compression combination method based on the flow table entry matching flow and the flow table entry error flow;

[0116] Step S750: Calculate the compression benefit of the compression combination method based on the flow table compression rate and compression accuracy.

[0117] In some embodiments, corresponding to steps S710 and S750, in order to select the most suitable compression strategy, this application needs to calculate the compression benefits generated by each compression strategy. For this purpose, this application defines the performance function as follows:

[0118] U a =ω1C a +ω2F a ;

[0119] Among them, U a This represents the compression benefit resulting from using compression strategy a (a∈A), where A represents the set of compression strategies, which has a total of 2... n Strategies; C a F represents the flow table compression effect brought about by the selected action 'a', i.e., the flow table compression ratio. a This indicates the case where the data stream is correctly routed, i.e., the compression accuracy. ω1 and ω2 are constant parameters. By modifying these two parameters, the weight between the flow table compression rate and the compression accuracy can be adjusted.

[0120] Specifically, this application defines C a The calculation formula is as follows: Where N represents the number of flow table entries before compression, N c This represents the number of flow table entries after compression.

[0121] Regarding the data stream scheduling effect, this application defines F a The calculation formula is as follows:

[0122]

[0123] Where Ne represents the number of data flows with routing errors caused by flow table compression, and the throughput corresponding to Ne represents the erroneous flow table entry traffic mentioned above. In contrast, the throughput corresponding to Nc represents the traffic size that matches each flow table entry, i.e., the traffic matched by each flow table entry mentioned above.

[0124] According to the above embodiments, by using the performance function U, this application can iteratively calculate the compression benefits brought by each strategy, and select the strategy with the greatest compression benefit to compress the flow table, thereby realizing the above steps of determining the compression strategy of high-risk flow table based on subsequent traffic, obtaining the corresponding compressed flow table, and updating the flow table of the corresponding network device based on the compressed flow table.

[0125] Specifically, in some embodiments, the pre-compression flow table entries are shown in Table (4):

[0126] Table (4) Items of Pre-compression Flow

[0127]

[0128] Based on the compression strategy determination process in the above embodiments, it can be concluded that the compression strategy selected by this application for Table (4) is to compress d and sp. The final compressed flow table entries are shown in Table (5):

[0129] Table (5) Flow table items after compression

[0130] Matching items Priority exit (s2, *, dp2) 4 1

[0131] It is conceivable that during flow table compression, in addition to changing the matching items, the priority and output port also need to be adjusted. Therefore, through steps S610 to S620, wildcard replacement processing is performed on all matching items corresponding to the compression strategy in the high-risk flow table, and the output port of the flow table item after wildcard replacement is adjusted according to the output information to obtain the wildcard replacement flow table. The priority of the flow table items in the wildcard replacement flow table is adjusted according to the priority information to obtain the compressed flow table. Specifically, when an exit conflict occurs, the exit value can be selected according to the subsequent traffic prediction result of the flow table item (corresponding to the above prediction result storage structure). At the same time, in order to avoid the compressed flow table item being affected by the original flow table item when replacing the old and new flow table items, the priority of the compressed flow table item will be set one level higher than the highest flow table item priority in the set of flow table items before compression. Then, after all flow table items are compressed, this application will distribute the compressed flow table items to the switch in batches through the OpenFlow protocol and delete the original corresponding flow table items in the switch.

[0132] In some embodiments, in response to a forwarding consultation message sent by a network device based on an updated compressed flow table, calculating the negative impact value of a compressed flow table entry includes: in response to a forwarding consultation message sent by a network device based on an updated compressed flow table, calculating the current traffic path and the optimal traffic path corresponding to the compressed flow table entry; calculating the current path cost corresponding to the current traffic path and the optimal path cost corresponding to the optimal traffic path based on the corresponding path hop count; and calculating the negative impact value of the compressed flow table entry based on the current path cost and the optimal path cost.

[0133] In some embodiments, calculating the negative impact value of a compressed flow table entry based on the current path cost and the optimal path cost includes: obtaining the entry transmission rate of the compressed flow table entry and the error traffic transmission rate corresponding to the forwarding advisory information; calculating a first product of the current path cost and the error traffic transmission rate, and a second product of the optimal path cost and the entry transmission rate; and dividing the first product by the second product to obtain the negative impact value of the compressed flow table entry.

[0134] In some embodiments, flow table entry rollback processing is performed on compressed flow table entries whose negative impact value is greater than a preset impact value, including: obtaining the optimal traffic path corresponding to the compressed flow table entry whose negative impact value is greater than the preset impact value; and sending the rollback flow table entry corresponding to the optimal traffic path to the network device so that the network device replaces the compressed flow table entry with the rollback flow table entry.

[0135] Please refer to Figure 8 , Figure 8 This is an example flowchart of the flow entry rollback process proposed in another embodiment of this application. To alleviate the problem of suboptimal data flow scheduling caused by flow entry compression, this application adds a flow entry rollback module to perform flow entry rollback processing on compressed flow entries whose negative impact value exceeds a preset impact value. The specific flow entry rollback process is as follows. Regarding the situation where traffic scheduling is affected (for ease of explanation, this article only analyzes the case where only one switch completes flow table compression), it can be divided into the following two types: 1. The data flow already existed before flow table compression, but due to flow table compression, the output port of the corresponding flow changed, resulting in a change in the routing path, that is... When this happens, the next-hop OpenFlow switch for the affected data flow will change, and the switch on the new path may not have a matching flow table entry and will send a packet-in forwarding query to the SDN controller; 2. A data flow should have completed the optimal path selection through the SDN controller, but because it matched a flow table entry with a wildcard, it was directly scheduled, preventing it from being delivered along the optimal path. In this case, the next-hop switch, because it has not performed compression, is likely unable to find a flow table entry matching the new flow, and will still send a packet-in forwarding query.

[0136] Through the above comparative analysis, it can be found that in both of the above situations, the adjacent switches will be unable to find a flow table entry that matches the flow, thereby triggering them to send forwarding consultation information, including packet-in data packets, to consult the SDN controller to obtain the corresponding flow table entry. In other words, when the SDN controller receives a forwarding consultation from a switch on a non-optimal path, and the corresponding data flow has also flowed through a switch that has undergone flow table compression, this application can consider that the flow has caused a routing error due to flow table compression. In order to quantitatively calculate the negative impact of flow table compression, it is necessary to calculate the negative impact value of the compressed flow table entry in the compressed flow table.

[0137] Specifically, when this application receives forwarding consultation information from a switch with a non-optimal path, it records the corresponding switch, the newly generated flow table entry information, and the cost of the new path versus the optimal path based on the path hop count. It also periodically uses the `dump-flows` command to obtain the current flow table entry information for the corresponding new data traffic and calculates the negative impact value. When the negative impact value exceeds a preset value, the SDN controller enters the flow table fallback module. The controller calculates the optimal path for the data flow based on the global topology and sets the priority of the flow table entry corresponding to this path higher than the original compressed flow table entry. This is then distributed to the corresponding switch. Thus, when new traffic arrives at the switch, it will be forwarded according to the optimal path and will no longer be affected by the compressed flow table entry.

[0138] Specifically, the formula for calculating the impact value can be as follows:

[0139]

[0140] Among them, c compress c represents the cost incurred by the propagation path chosen due to compression, corresponding to the current path cost mentioned above. best This represents the cost incurred by the optimal path of the data flow, corresponding to the optimal path cost mentioned above, rate. error This indicates the transmission rate of the erroneous traffic corresponding to the forwarded consultation information. This represents the transmission rate after compression of flow table entries. It can be considered the average transmission rate of traffic after compression of flow table entries. Considering that 80% of the data flow in the network only contributes 20% of the traffic, simply using the traffic size to judge the impact of compression of the routing table would bias the results towards large data flows. Therefore, to improve the impact of short flows caused by routing errors, this application uses rate′ in the erroneous traffic transmission rate. This application defines a lower limit for the erroneous traffic transmission rate in erroneous throughput, aiming to improve the impact of small flows being incorrectly scheduled, as defined below:

[0141]

[0142] Among them, rate min The minimum transmission rate in the SDN network is denoted as 'rate', where 'rate' is the actual calculated transmission rate of erroneous traffic. When the transmission rate of erroneous traffic is less than the minimum transmission rate, the transmission rate of erroneous traffic is determined as the minimum transmission rate to mitigate the impact of erroneous scheduling of small flows and thus improve the accuracy of the impact value calculation.

[0143] In summary, this application proposes a heuristic flow table compression strategy based on future traffic transmission trends. This makes the flow table compression decision-making process not only dependent on current information but also referencing future trends, which is more conducive to making optimal decisions for the future. It can solve a series of problems caused by flow table overflow in SDN scenarios. Furthermore, it proposes a flow table rollback function to resolve the negative impact of flow table compression. This helps data flows that cannot be routed through the optimal path due to flow table entry compression to replan their routes. Through the overall solution of flow table compression and flow table rollback in this application, those skilled in the art can effectively complete the design of a series of interactive processes, including traffic prediction, flow table compression, and effect monitoring, and achieve the technical effects of this application.

[0144] Please refer to Figure 9 , Figure 9 This is a schematic diagram of the structure of an SDN network proposed in another embodiment of this application; Secondly, embodiments of this application provide a fallback flow table compression system. The system is applied to an SDN controller in an SDN network. The SDN network also includes multiple network devices that are communicatively connected to the SDN controller and have communication connections with each other. The system includes: a traffic prediction module, used to obtain current flow table entry information contained in the flow tables of the multiple network devices, such as... Figure 9 The system includes network devices 1 to 5; it calculates the flow table utilization rate of multiple network devices based on the current flow table entry information; it defines the flow tables of all network devices with a flow table utilization rate greater than a preset utilization rate as high-risk flow tables; it obtains historical flow table entry information of high-risk flow tables; it predicts the subsequent traffic of high-risk flow tables based on the current and historical flow table entry information; a flow table compression module is used to determine the compression strategy of high-risk flow tables based on subsequent traffic, obtain the corresponding compressed flow tables, and perform flow table update processing on the corresponding network devices based on the compressed flow tables; a flow table entry rollback module is used to respond to the forwarding consultation information sent by the network devices based on the updated compressed flow tables, calculate the negative impact value of the compressed flow table entries in the compressed flow tables; and it performs flow table entry rollback processing on compressed flow table entries with a negative impact value greater than a preset impact value.

[0145] refer to Figure 9 In some embodiments, Figure 9The traffic information includes current flow table entries and historical flow table entries. Because SDN adopts a strategy of separating the control plane and the data plane, the scheduling of network devices such as OpenFlow switches depends entirely on the configuration of flow table entries, resulting in a large increase in flow table entries. When the number of flow table entries exceeds the flow table capacity, it may lead to a series of problems such as data flow not being able to be scheduled. In order to solve a series of problems caused by the flow table being filled, this application proposes a rollback flow table compression system. It can complete the flow table compression work in the OpenFlow switch, avoid the problems caused by flow table overflow, and solve the problem that some traffic may not be scheduled according to the optimal path due to flow table entry compression. In order to avoid unacceptable negative impacts, the rollback flow table compression system includes a flow table entry rollback module in addition to the traffic prediction module and the flow table compression module. The flow table entry rollback module is used to continuously track the impact of compression and roll back some flow table entries in a timely manner.

[0146] It is conceivable that, given the limited memory of the switch, in order to avoid a series of problems caused by flow table overflow, the system in this application can not only solve the risk of flow table overflow through compression, but also provide a detection rollback mechanism to help reduce the negative impact of flow table compression.

[0147] Specifically, the overall architecture of the system in this application is as follows: Figure 9 As shown, consistent with traditional SDN architecture, the overall scheme of this application is divided into a control plane and a data plane from top to bottom. The data plane is generally composed of OpenFlow switches, which only forward packets. The control layer makes decisions based on the packet information fed back from the data layer and sends the determined flow table entries to the corresponding switches so that they can complete forwarding according to the flow table entries. In order to better illustrate the workflow of this application, the flow table compression system scheme is decomposed into three functional modules, namely the traffic prediction module: In order to maximize the effectiveness of the compression decisions made by the flow table compression system, for flow tables that the controller considers high-risk, the SDN controller will use the dump-flows command to obtain traffic information. The system uses LSTM to predict future traffic patterns within the switch and transmits the predicted trends to the flow table compression module to assist in selecting a compression strategy. The flow table compression module combines the current switch status and predicted traffic trends, using a heuristic algorithm to select a compression strategy. Based on the compression strategy, the system distributes the compressed flow table entries via packet_out and deletes the original entries, ultimately completing the flow table compression. The flow table entry rollback module continuously monitors the compression effect after completion. When the affected data flow exceeds the rollback threshold, the system rolls back the flow table entries for that flow.

[0148] Please refer to Figure 10 , Figure 10 This is a schematic diagram of the structure of an electronic device provided in an embodiment of this application. In a third aspect, some embodiments of this application provide an electronic device, which includes a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, it implements the fallback flow table compression method of any of the above embodiments.

[0149] The electronic device 1000 of this application embodiment includes one or more processors 1010 and memory 1020. Figure 10 The example uses a processor 1010 and a memory 1020.

[0150] The processor 1010 and the memory 1020 can be connected via a bus or other means. Figure 10 Taking the example of a connection between China and Israel via a bus.

[0151] Memory 1020, as a non-transitory computer-readable storage medium, can be used to store non-transitory software programs and non-transitory computer-executable programs. Furthermore, memory 1020 may include high-speed random access memory, and may also include non-transitory memory, such as at least one disk storage device, flash memory device, or other non-transitory solid-state storage device. In some embodiments, memory 1020 may optionally include memory 1020 remotely located relative to processor 1010. These remote memories can be connected to electronic device 1000 via a network, and examples of such networks include, but are not limited to, the Internet, intranets, local area networks, mobile communication networks, and combinations thereof.

[0152] In some embodiments, when the processor executes a computer program, it performs the fallback flow table compression method of any of the above embodiments at preset intervals.

[0153] Those skilled in the art will understand that Figure 10 The device structure shown does not constitute a limitation on the electronic device 1000, and may include more or fewer components than shown, or combine certain components, or have different component arrangements.

[0154] exist Figure 10 In the illustrated electronic device 1000, the processor 1010 can be used to call a rollback flow table compression method stored in the memory 1020, thereby implementing the rollback flow table compression method.

[0155] Based on the hardware structure of the above-mentioned electronic device 1000, various embodiments of the fallback flow table compression system of this application are proposed. Meanwhile, the non-transient software program and instructions required to implement the fallback flow table compression method of the above embodiments are stored in memory. When executed by the processor, the fallback flow table compression method of the above embodiments is executed.

[0156] Furthermore, embodiments of this application also provide a rollback flow table compression system, which includes the aforementioned electronic device.

[0157] In some embodiments, since the fallback flow table compression system of this application has the electronic device of the above embodiments, and the electronic device of the above embodiments is capable of executing the fallback flow table compression method of the above embodiments, the specific implementation and technical effects of the fallback flow table compression system of this application can refer to the specific implementation and technical effects of the fallback flow table compression method of any of the above embodiments. While improving the flow table compression effect by predicting subsequent flow, it can also realize the fallback operation of compressed flow table entries to mitigate the negative impact caused by flow table compression.

[0158] Fourthly, embodiments of this application also provide a computer-readable storage medium storing computer-executable instructions for performing the above-described fallback flow table compression method. For example, the computer-executable instructions may cause one or more processors to perform the fallback flow table compression method in the above-described method embodiments.

[0159] The device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separate; that is, they may be located in one place or distributed across multiple network nodes. Some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs.

[0160] Those skilled in the art will understand that all or some of the steps and systems in the methods disclosed above can be implemented as software, firmware, hardware, and suitable combinations thereof. Some or all of the physical components can be implemented as software executed by a processor, such as a central processing unit, digital signal processor, or microprocessor, or as hardware, or as an integrated circuit, such as an application-specific integrated circuit. Such software can be distributed on a computer-readable medium, which can include computer-readable storage media (or non-transitory media) and communication media (or transient media). As is known to those skilled in the art, the term computer-readable storage media includes volatile and non-volatile, removable and non-removable media implemented in any method or technology for storing information (such as computer-readable instructions, data structures, program modules, or other data). Computer-readable storage media include, but are not limited to, RAM, ROM, EEPROM, flash memory or other memory technologies, CD-ROM, digital versatile disc (DVD) or other optical disc storage, magnetic cartridges, magnetic tape, disk storage or other magnetic storage devices, or any other medium that can be used to store desired information and is accessible to a computer. Furthermore, as is known to those skilled in the art, communication media typically contain computer-readable instructions, data structures, program modules, or other data in modulated data signals such as carrier waves or other transmission mechanisms, and may include any information delivery medium.

[0161] The above is a detailed description of the preferred embodiments of this application. However, this application is not limited to the above embodiments. Those skilled in the art can make various equivalent modifications or substitutions without departing from the spirit of this application. All such equivalent modifications or substitutions are included within the scope defined by the claims of this application.

Claims

1. A rollbackable flow table compression method, characterized in that, include: Retrieve information about the current flow table entries contained in the flow tables of multiple network devices; Calculate the flow table utilization rate of multiple network devices based on the current flow table entry information; Define the flow tables of all network devices whose flow table usage rate is greater than the preset usage rate as high-risk flow tables. Obtain historical flow table entry information for the high-risk flow table; The subsequent traffic of the high-risk flow table is predicted based on the current flow table entry information and the historical flow table entry information; The compression strategy of the high-risk flow table is determined based on the subsequent traffic, the corresponding compressed flow table is obtained, and the flow table update process is performed on the corresponding network device based on the compressed flow table. In response to a forwarding consultation message sent by the network device based on the updated compressed flow table, the negative impact value of the compressed flow table entry is calculated. For compressed flow table entries whose negative impact value is greater than a preset impact value, flow table entry rollback processing is performed.

2. The rollbackable flow table compression method according to claim 1, characterized in that, The step of predicting the subsequent traffic of the high-risk flow table based on the current flow table entry information and the historical flow table entry information includes: If the current data flow corresponding to the current flow table entry is less than the prediction threshold, calculate the average data flow corresponding to the current flow table entry and the historical flow table entry, and determine the average data flow as the subsequent flow of the high-risk flow table. If the current data flow is greater than or equal to the prediction threshold, the current flow table entry information and the historical flow table entry information are input into the LSTM model for prediction processing to obtain the subsequent flow of the high-risk flow table.

3. The rollbackable flow table compression method according to claim 1, characterized in that, The step of determining the compression strategy for the high-risk flow table based on the subsequent traffic to obtain the corresponding compressed flow table includes: Based on the subsequent traffic, the matching information, output information, and priority information of each flow table entry in the high-risk flow table are obtained; The compression strategy for the high-risk flow table is determined based on the matching information; Based on the corresponding compression strategy, the output information, and the priority information, the high-risk flow table is compressed to obtain a compressed flow table.

4. The rollbackable flow table compression method according to claim 3, characterized in that, The step of determining the compression strategy for the high-risk flow table based on the matching item information includes: All matching items in the high-risk flow table are determined based on the matching item information; By selecting combinations of matching items from all the aforementioned matching items, various compression and combination methods can be obtained; The compression efficiency of all the compression combinations is calculated based on a preset performance function; The compression combination method that yields the greatest compression benefit is determined to be the compression strategy for the high-risk flow table.

5. The rollbackable flow table compression method according to claim 3, characterized in that, The step of compressing the high-risk flow table according to the corresponding compression strategy, the output information, and the priority information to obtain a compressed flow table includes: Wildcard replacement processing is performed on all matching items corresponding to the compression strategy in the high-risk flow table, and the output port of the flow table item after wildcard replacement is adjusted according to the output information to obtain the wildcard replacement flow table. Adjust the priority of the wildcard replacement entries in the flow table according to the priority information to obtain a compressed flow table.

6. The rollbackable flow table compression method according to claim 4, characterized in that, The step of calculating the compression efficiency of all compression combinations based on a preset performance function includes: Obtain the number of flow table entries before compression and the number of flow table entries after compression of the compression combination method; The flow table compression ratio of the compression combination method is calculated based on the number of flow table entries before compression and the number of flow table entries after compression. Obtain the flow table entry matching traffic and flow table entry error traffic after compression according to the described compression combination method; The compression accuracy of the compression combination method is calculated based on the flow table entry matching traffic and the flow table entry error traffic; The compression efficiency of the compression combination method is calculated based on the flow table compression ratio and the compression accuracy.

7. The rollbackable flow table compression method according to claim 1, characterized in that, The step of responding to the network device sending forwarding advice information based on the updated compressed flow table, and calculating the negative impact value of the compressed flow table entry, includes: In response to the forwarding consultation information sent by the network device based on the updated compressed flow table, the current traffic path and the optimal traffic path corresponding to the compressed flow table entry in the compressed flow table are calculated; Calculate the current path cost corresponding to the current traffic path and the optimal path cost corresponding to the optimal traffic path based on the corresponding path hop count. The negative impact value of the compressed flow table entry is calculated based on the current path cost and the optimal path cost.

8. The reversible flow table compression method according to claim 7, characterized in that, The step of calculating the negative impact value of the compressed flow table entry based on the current path cost and the optimal path cost includes: Obtain the entry transmission rate of the compressed flow table entry and the error traffic transmission rate corresponding to the forwarding advisory information; Calculate the first product of the current path cost and the erroneous traffic transmission rate, and the second product of the optimal path cost and the entry transmission rate. Divide the first product by the second product to obtain the negative impact value of the compressed flow entry.

9. The rollbackable flow table compression method according to claim 1, characterized in that, The process of rolling back compressed flow table entries whose negative impact value is greater than a preset impact value includes: Obtain the optimal flow path corresponding to the compressed flow table entry whose negative impact value is greater than a preset impact value; The fallback flow table entry corresponding to the optimal traffic path is sent to the network device so that the network device can replace the compressed flow table entry with the fallback flow table entry.

10. A rollbackable flow table compression system, characterized in that, The system is applied to an SDN controller in an SDN network, which also includes multiple network devices that are communicatively connected to the SDN controller and have communication connections with each other. The system includes: The traffic prediction module is used to obtain current flow table entry information contained in the flow tables of multiple network devices; calculate the flow table utilization rate of the multiple network devices based on the current flow table entry information; define the flow tables of all network devices whose flow table utilization rate is greater than a preset utilization rate as high-risk flow tables; obtain historical flow table entry information of the high-risk flow tables; and predict the subsequent traffic of the high-risk flow tables based on the current flow table entry information and the historical flow table entry information. The flow table compression module is used to determine the compression strategy of the high-risk flow table based on the subsequent traffic, obtain the corresponding compressed flow table, and perform flow table update processing on the corresponding network device based on the compressed flow table. The flow entry rollback module is used to respond to the forwarding consultation information sent by the network device based on the updated compressed flow table, calculate the negative impact value of the compressed flow table entry, and perform flow entry rollback processing on the compressed flow table entry whose negative impact value is greater than a preset impact value.

11. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the computer program, it implements the fallback flow table compression method as described in any one of claims 1 to 9.

12. A computer-readable storage medium, characterized in that, The device contains a computer program that, when executed by a processor, implements the rollbackable flow table compression method as described in any one of claims 1 to 9.

Images