TOE comprehensive performance tuning test method and system based on multi-layer protocol stack

Abstract

The invention discloses a TOE comprehensive performance tuning test method and system based on a multi-level protocol stack, and the method comprises the steps: building a comprehensive performance calculation formula related to each performance index based on each performance index of the multi-level protocol stack, and a link layer, a network layer and a transmission layer which act on TOE, and obtaining a TOE comprehensive performance model; providing a plurality of application scenes corresponding to TOE for a plurality of indexes in the TOE comprehensive performance model; and for performance bottlenecks in different application scenes, the value of the TOE comprehensive performance is improved by adjusting influence factors, so that the optimal TOE comprehensive performance is obtained. According to the method provided by the invention, the comprehensive performance index of the TOE can be maximized while the network data transmission stability is ensured.

Description

Technical Field

[0001] The present invention relates to the fields of TOE, network performance optimization and testing, and in particular to a TOE comprehensive performance tuning test method and system based on a multi-layer protocol stack. Background Art

[0002] With the surge in network data traffic and the bottleneck of server processing capacity, TCP Offload Engine (TOE) technology has been widely used to reduce CPU burden and improve network data processing efficiency. However, in actual applications, the performance of TOE at different network levels varies, and the interaction of different protocol levels may lead to performance bottlenecks. Therefore, how to conduct effective comprehensive performance testing and optimization of TOE technology in an actual network environment has become a key issue in improving network performance.

[0003] Currently, existing performance testing methods mainly focus on performance optimization and bottleneck identification of a single protocol layer, lack comprehensive performance evaluation for multi-layer protocol stacks, and fail to fully test the interaction between layers.

[0004] In view of this, this application is hereby filed. Summary of the invention

[0005] The technical problem to be solved by the present invention is that the existing performance testing methods mainly focus on the performance optimization and bottleneck identification of a single protocol layer, lack of comprehensive performance evaluation for multi-layer protocol stacks, and the interaction between layers has not been fully tested. The purpose of the present invention is to provide a TOE comprehensive performance tuning test method and system based on a multi-layer protocol stack. This method uses the test results of the multi-layer protocol stack and combines the performance of TOE at each layer to comprehensively evaluate the overall performance, that is, comprehensively evaluate the test results of the multi-layer protocol stack to obtain the overall optimal performance. This method can maximize the comprehensive performance indicators of TOE while ensuring the stability of network data transmission.

[0006] The present invention is achieved through the following technical solutions:

[0007] In a first aspect, the present invention provides a TOE comprehensive performance tuning test method based on a multi-layer protocol stack, the method comprising:

[0008] Based on the performance indicators of the multi-layer protocol stack, a comprehensive performance calculation formula associated with each performance indicator is constructed for the link layer, network layer and transport layer where TOE acts, and a TOE comprehensive performance model is obtained;

[0009] For multiple indicators in the TOE comprehensive performance model, multiple application scenarios corresponding to TOE are proposed; and for performance bottlenecks in different application scenarios, the value of TOE comprehensive performance is improved by adjusting the influencing factors to obtain the optimal TOE comprehensive performance, so as to achieve TOE comprehensive performance improvement.

[0010] Furthermore, the steps to obtain the TOE comprehensive performance model are:

[0011] Based on the performance indicators of the multi-layer protocol stack, a comprehensive performance calculation formula related to each performance indicator is constructed;

[0012] The TOE is tested, and various performance indicators are obtained by combining the test results of the link layer, network layer and transport layer;

[0013] Substitute each performance index into the comprehensive performance calculation formula to obtain the TOE comprehensive performance model;

[0014] Among them, the performance indicators include throughput indicators, connection indicators and CPU consumption indicators.

[0015] Furthermore, the comprehensive performance calculation formula is:

[0016] total=A*n1+B*n2+C*n3

[0017] In the formula, total is the total score of TOE comprehensive performance; A is the score of the throughput indicator; B is the score of the connection indicator; C is the score of the CPU consumption indicator; n1 is the weight of the throughput indicator; n2 is the weight of the connection indicator; n3 is the weight of the CPU consumption indicator.

[0018] Furthermore, the expression of TOE comprehensive performance model is:

[0019]

[0020] In the formula, total is the total score of TOE comprehensive performance; T 链路 is the link layer throughput indicator; T 网络 is the network layer throughput indicator; T 传输 is the transport layer throughput indicator; n1 is the weight of the throughput indicator; n2 is the weight of the connection indicator; n3 is the weight of the CPU consumption indicator; T 连接建立时间 TCP connection establishment time; T 连接延迟 is the TCP connection delay; T 重传率 is the retransmission rate of the TCP connection; T 丢包率 is the packet loss rate of the TCP connection; w1, w2, w3, and w4 are T 连接建立时间 , T 连接延迟 , T 重传率 , T 丢包率The corresponding weight; cpu consumption is the cpu consumption indicator.

[0021] Furthermore, the throughput index is based on the bucket effect, and the throughput index value is ultimately the minimum value among the link layer throughput index, the network layer throughput index, and the transport layer throughput index;

[0022] The connection index test references connection establishment time, connection delay, retransmission rate and stability;

[0023] The CPU consumption indicator is to use the top or htop tool to directly monitor the system's CPU usage and view the percentage of CPU occupied by TOE driver-related processes.

[0024] Furthermore, TOE corresponds to multiple application scenarios and adjustments, including:

[0025] The first TOE application scenario is a high-concurrency or bandwidth-constrained network, and the corresponding adjustment factor is the network bandwidth limitation;

[0026] The second TOE application scenario is an unstable network, and the corresponding adjustment factor is the packet loss rate;

[0027] The third TOE application scenario is a high-latency network, and the corresponding adjustment factor is latency;

[0028] The fourth TOE application scenario is the flow control mechanism, and the corresponding adjustment factor is the sliding window size;

[0029] The fifth TOE application scenario is small packet intensive requests or large packet traffic, and the corresponding adjustment factor is the data packet size;

[0030] The sixth TOE application scenario is a complex multi-path network, and the corresponding adjustment factor is the routing strategy;

[0031] The seventh TOE application scenario is the coexistence of short connections and long connections, and the corresponding adjustment factor is the TCP request type;

[0032] The eighth TOE application scenario is conventional high-concurrency connection, and the corresponding adjustment influencing factor is normal TCP connection;

[0033] The ninth TOE application scenario is sudden connection disconnection or half-open connection, and the corresponding adjustment influencing factor is TCP abnormal connection.

[0034] Furthermore, for the performance in different application scenarios, the value of TOE comprehensive performance is improved by adjusting the influencing factors to obtain the optimal TOE comprehensive performance, including:

[0035] According to the performance in different application scenarios, the weight parameters of each indicator in the TOE comprehensive performance model are adjusted accordingly;

[0036] In a single application scenario, limit the restricted factors of the application scenario and adjust other unrestricted factors to maximize the total performance score to obtain the optimal TOE comprehensive performance.

[0037] In a second aspect, the present invention further provides a TOE comprehensive performance tuning test system based on a multi-layer protocol stack, the system comprising:

[0038] The performance model building unit is used to build a comprehensive performance calculation formula associated with each performance indicator for the link layer, network layer and transport layer of the TOE based on each performance indicator of the multi-layer protocol stack, so as to obtain a TOE comprehensive performance model;

[0039] A TOE scenario proposal unit is used to propose multiple application scenarios corresponding to TOE for multiple indicators in the TOE comprehensive performance model;

[0040] The TOE performance tuning unit is used to improve the value of TOE comprehensive performance by adjusting the influencing factors of performance bottlenecks in different application scenarios, so as to obtain the optimal TOE comprehensive performance and achieve TOE comprehensive performance improvement.

[0041] Furthermore, the execution process of the performance model building unit is:

[0042] Based on the performance indicators of the multi-layer protocol stack, a comprehensive performance calculation formula related to each performance indicator is constructed;

[0043] The TOE is tested, and various performance indicators are obtained by combining the test results of the link layer, network layer and transport layer;

[0044] Substitute each performance index into the comprehensive performance calculation formula to obtain the TOE comprehensive performance model;

[0045] Among them, the performance indicators include throughput indicators, connection indicators and CPU consumption indicators.

[0046] Compared with the prior art, the present invention has the following advantages and beneficial effects:

[0047] The present invention is based on a TOE comprehensive performance tuning test method and system of a multi-layer protocol stack. The test results of the multi-layer protocol stack are combined with the performance of TOE at each layer to comprehensively evaluate the overall performance, that is, the overall optimal performance is obtained by comprehensively evaluating the test results of the multi-layer protocol stack. This method can maximize the comprehensive performance indicators of TOE while ensuring the stability of network data transmission. BRIEF DESCRIPTION OF THE DRAWINGS

[0048] The drawings described herein are used to provide a further understanding of the embodiments of the present invention, constitute a part of this application, and do not constitute a limitation of the embodiments of the present invention. In the drawings:

[0049] Figure 1 It is a flow chart of the TOE comprehensive performance tuning test method based on a multi-layer protocol stack of the present invention;

[0050] Figure 2 This is a flow chart of the TOE comprehensive performance tuning test method according to Embodiment 1 of the present invention;

[0051] Figure 3 It is a structural block diagram of the TOE comprehensive performance tuning test system based on a multi-layer protocol stack of the present invention. DETAILED DESCRIPTION

[0052] In order to make the objectives, technical solutions and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with embodiments and drawings. The exemplary embodiments of the present invention and their description are only used to explain the present invention and are not intended to limit the present invention.

[0053] The present invention provides a TOE comprehensive performance tuning test method and system based on a multi-layer protocol stack. Through layered analysis and comprehensive evaluation of the performance of each protocol layer, it can accurately identify and quantify the various factors that affect the TOE (TCP Offload Engine) performance, and combine the barrel effect principle to accurately calculate the system's comprehensive throughput and performance bottleneck. Compared with traditional technologies, the present invention has the following significant advantages:

[0054] (1) Comprehensive performance scoring system

[0055] The present invention proposes a comprehensive evaluation method based on performance scores. The system generates a unified comprehensive performance score by setting multiple performance indicators such as throughput, latency, and CPU occupancy, and combining a weighted scoring method. This comprehensive scoring system can help engineers quickly determine the overall performance of the TOE system, identify which layer has a performance bottleneck, and then adjust the configuration in a targeted manner to optimize performance.

[0056] (2) Automated testing and optimization suggestions

[0057] The test method of the present invention not only provides diagnosis of system performance bottlenecks, but also automatically generates optimization suggestions based on test data to help users quickly find the best configuration. Compared with traditional manual debugging and optimization methods, this function greatly improves test efficiency, reduces manual intervention, and can perform performance optimization automatically on a large scale.

[0058] (3) Improve equipment stability and adaptability

[0059] Since the present invention considers the synergy of multi-layer protocol stacks and evaluates and optimizes the performance of each layer through fine-grained parameter adjustment, it can ensure the stability and adaptability of TOE devices in different network environments. Compared with traditional methods that rely on single parameter optimization, it avoids the problem of uneven performance or local optimization failing to bring about global performance improvement.

[0060] (4) Wide range of application scenarios

[0061] The test method of the present invention is not only applicable to traditional TOE device performance testing, but can also be widely used in emerging hardware acceleration network devices, such as smart network cards, network processing units (NPUs), etc. The method can help device developers, network engineers, and operators test and optimize device performance in different environments, especially in demanding scenarios such as high concurrency, large traffic, and low latency, and has important practical application value.

[0062] In summary, the present invention provides a comprehensive and accurate TOE performance testing solution by introducing a multi-layer protocol stack testing method, combining the barrel effect principle and the comprehensive performance scoring system, which can effectively identify and optimize the performance bottlenecks of each protocol layer and improve the throughput, latency and stability of the overall system. Compared with traditional performance testing methods, the present invention has higher flexibility and adaptability, and can provide strong support for the development and optimization of TOE equipment.

[0063] Example 1

[0064] like Figure 1 As shown, the TOE comprehensive performance tuning test method based on the multi-level protocol stack of the present invention comprises:

[0065] Based on the performance indicators of the multi-layer protocol stack, a comprehensive performance calculation formula associated with each performance indicator is constructed for the link layer, network layer and transport layer where TOE acts, and a TOE comprehensive performance model is obtained;

[0066] For multiple indicators in the TOE comprehensive performance model, multiple application scenarios corresponding to TOE are proposed; and for performance bottlenecks in different application scenarios, the value of TOE comprehensive performance is improved by adjusting the influencing factors to obtain the optimal TOE comprehensive performance, so as to achieve TOE comprehensive performance improvement.

[0067] like Figure 2 As shown, the specific implementation steps of the present invention are as follows:

[0068] Step 1: Construct a comprehensive performance calculation formula associated with each performance indicator. The comprehensive performance calculation formula is as follows:

[0069] total=A*n1+B*n2+C*n3

[0070] In the formula, total is the total score of TOE comprehensive performance; A is the score of the throughput indicator; B is the score of the connection indicator; C is the score of the CPU consumption indicator; n1 is the weight of the throughput indicator; n2 is the weight of the connection indicator; n3 is the weight of the CPU consumption indicator.

[0071] The total performance score is obtained by multiplying each performance indicator and its weight. Each weight parameter is adjusted according to the focus of the object to be tested. It is necessary to ensure that each weight parameter is a number between 0 and 1.

[0072] Step 2: Test the TOE. The performance indicators are obtained by combining the test results of the link layer, network layer and transport layer.

[0073] Step 3: In step 1, the throughput index in the comprehensive performance calculation formula is obtained by the following formula:

[0074] A=min(T 链路 ,T 网络 ,T 传输 )

[0075] Based on the barrel effect, the throughput index value finally takes the minimum value among the link layer throughput index, network layer throughput index, and transport layer throughput index.

[0076] The link layer throughput test focuses on the physical transmission medium, the quality of the link, and the transmission efficiency of the protocol frame. The link layer throughput is usually affected by factors such as link bandwidth, link delay and jitter, and packet loss rate. To test the link layer throughput, the test process simulates different link qualities. Use the tc tool to control and simulate network conditions to simulate different bandwidths, delays, and packet loss rates, and use the iperf3 tool to send traffic to record throughput.

[0077] The calculation formula of the link layer throughput indicator is as follows:

[0078]

[0079] The network layer throughput mainly focuses on the packet forwarding efficiency and network routing efficiency of the IP layer, which is mainly reflected in the transmission of IP packets, network delay, packet loss rate, and routing performance. Use the traceroute tool to simulate different network routes and congestion conditions, and use the iperf3 tool to start the client and server to record the throughput.

[0080] The calculation formula for the network layer throughput indicator is as follows:

[0081]

[0082] The transport layer focuses on end-to-end reliability and flow control mechanisms, which are mainly reflected in TCP throughput, latency, retransmission, etc. Use the iperf3 tool to start the client and server to record the throughput.

[0083] The calculation formula of the transport layer throughput index is as follows:

[0084]

[0085] In summary, the minimum value of the three-layer throughput can be taken as the system throughput indicator.

[0086] Step 4: The connection indicator is to ensure whether TOE can correctly handle various TCP connection situations, including connection establishment, disconnection, connection retry, connection stability, etc.

[0087] The test connection indicators need to refer to: connection establishment time, connection delay, retransmission rate, and stability.

[0088] The connection establishment time is measured by testing the time difference between the SYN, SYN-ACK, and ACK requests of the TCP three-way handshake. The process uses the iperf3 tool to simulate a large number of connection requests and check the time it takes to establish a connection.

[0089] The connection delay is calculated by calculating the time difference from SYN to ACK request in the TCP three-way handshake and the time difference from FIN to the last ACK request in the TCP four-way handshake. The process uses the netcat tool to simulate the client and server to test the connection disconnection and establishment process.

[0090] The retransmission rate is used to observe the proportion of TCP retransmissions by simulating packet loss. This process uses the tc tool or scapy to verify the retransmission rate by sending a large number of ACK packets.

[0091] The packet loss rate is measured by simulating the packet loss rate under high concurrency and high load conditions. This process uses the ab tool or wrk tool to initiate a large number of requests and set the number of concurrent connections to simulate high concurrency and high load conditions.

[0092] The calculation formula of the connection index is as follows:

[0093]

[0094] Where T 连接建立时间 TCP connection establishment time; T 连接延迟 is the TCP connection delay; T 重传率 is the retransmission rate of the TCP connection; T 丢包率 is the packet loss rate of the TCP connection; w1, w2, w3, and w4 are T 连接建立时间 , T 连接延迟 , T 重传率 , T 丢包率Corresponding weights; the weights of various parameters are set independently according to system requirements, and the weights must be between 0 and 1.

[0095] Step 5: CPU consumption indicators Use the top or htop tool to directly monitor the system's CPU usage and view the percentage of CPU occupied by TOE driver-related processes.

[0096] Step 6: Combining the formulas from step 1 to step 5, substituting various performance indicators into the comprehensive performance calculation formula, and obtaining the TOE comprehensive performance model; the expression of the TOE comprehensive performance model is:

[0097]

[0098] In the formula, total is the total score of TOE comprehensive performance; T 链路 is the link layer throughput indicator; T 网络 is the network layer throughput indicator; T 传输 is the transport layer throughput indicator; n1 is the weight of the throughput indicator; n2 is the weight of the connection indicator; n3 is the weight of the CPU consumption indicator; T 连接建立时间 TCP connection establishment time; T 连接延迟 is the TCP connection delay; T 重传率 is the retransmission rate of the TCP connection; T 丢包率 is the packet loss rate of the TCP connection; w1, w2, w3, and w4 are T 连接建立时间 , T 连接延迟 , T 重传率 , T 丢包率 The corresponding weight; cpu consumption is the cpu consumption indicator.

[0099] In summary, steps 1 to 6 explain the formula and the indicator testing process required for the formula.

[0100] Step 7: Summarizing steps 1 to 6, the factors that affect the overall TOE performance score include: network bandwidth limitation, packet loss rate, latency, sliding window size, data packet size, routing strategy, TCP single request types, TCP complete connection request types, TCP abnormal connection request types. In view of the above influencing factors, the following TOE application scenarios are proposed, and the TOE corresponding application scenarios and adjustments include:

[0101] (1) First TOE application scenario: In high-concurrency or bandwidth-constrained networks, it is necessary to measure the throughput stability under different bandwidths; the influencing factor for the corresponding adjustment is the network bandwidth limitation;

[0102] (2) Second TOE application scenario: In unstable networks such as wireless and long-distance links, the retransmission rate and performance degradation need to be observed; the corresponding adjustment factor is the packet loss rate;

[0103] (3) Third TOE application scenario: In high-latency networks such as satellite communications, it is necessary to measure the impact of latency on throughput and connection quality; the corresponding adjustment factor is latency;

[0104] (4) Fourth TOE application scenario: The flow control mechanism is tested, such as large file transfer, and the matching between window size and throughput performance needs to be analyzed; the corresponding adjustment factor is the sliding window size;

[0105] (5) Fifth TOE application scenario: small packet intensive requests or large packet traffic such as IOT devices or high-definition video, it is necessary to verify the throughput capacity under different packet sizes; the corresponding adjustment factor is the data packet size;

[0106] (6) The sixth TOE application scenario: complex multi-path networks, it is necessary to analyze the impact of routing changes on connection stability and throughput; the corresponding adjustment factor is the routing strategy;

[0107] (7) Seventh TOE application scenario: short connections and long connections coexist, and the connection management efficiency needs to be evaluated; the corresponding adjustment influencing factor is the TCP request type;

[0108] (8) The eighth TOE application scenario: conventional high-concurrency connections, the performance of establishing and maintaining connections needs to be tested; the corresponding adjustment influencing factor is the normal TCP connection;

[0109] (9) The ninth TOE application scenario: sudden connection disconnection or half-open connection, the efficiency of exception handling and resource recovery capability need to be verified; the corresponding adjustment influencing factor is TCP abnormal connection.

[0110] Step 8: For the TOE application scenarios proposed in step 7, the weight parameters of the performance score indicators in step 6 can be adjusted accordingly to meet the focus requirements of each TOE application scenario. In a single application scenario, the restricted factors of the application scenario are restricted, and other unrestricted factors are adjusted to maximize the total performance score;

[0111] Step 9: Referring to Step 7 and Step 8, other TOE application scenarios can be proposed as needed, and based on the TOE comprehensive performance model in Step 6, a single indicator that needs to be restricted can be proposed. By restricting this indicator and adjusting other indicators to ensure that the formula score reaches the highest, the TOE comprehensive performance can be ensured to be optimal.

[0112] Example 2

[0113] like Figure 3 As shown, the difference between this embodiment and embodiment 1 is that this embodiment further provides a TOE comprehensive performance tuning test system based on a multi-layer protocol stack, and the system corresponds to the TOE comprehensive performance tuning test method based on a multi-layer protocol stack in embodiment 1 in a one-to-one manner; the system includes:

[0114] The performance model building unit is used to build a comprehensive performance calculation formula associated with each performance indicator for the link layer, network layer and transport layer of the TOE based on each performance indicator of the multi-layer protocol stack, so as to obtain a TOE comprehensive performance model;

[0115] The TOE scenario proposal unit is used to propose multiple application scenarios corresponding to TOE for multiple indicators in the TOE comprehensive performance model;

[0116] The TOE performance tuning unit is used to improve the value of TOE comprehensive performance by adjusting the influencing factors of performance bottlenecks in different application scenarios, so as to obtain the optimal TOE comprehensive performance and achieve TOE comprehensive performance improvement.

[0117] As a further implementation, the execution process of the performance model building unit is:

[0118] Based on the performance indicators of the multi-layer protocol stack, a comprehensive performance calculation formula related to each performance indicator is constructed;

[0119] The TOE is tested, and various performance indicators are obtained by combining the test results of the link layer, network layer and transport layer;

[0120] Substitute each performance index into the comprehensive performance calculation formula to obtain the TOE comprehensive performance model;

[0121] Among them, the performance indicators include throughput indicators, connection indicators and CPU consumption indicators.

[0122] Among them, the execution process of each unit can be executed according to the process steps of the TOE comprehensive performance tuning test method based on the multi-layer protocol stack in Example 1, and will not be repeated one by one in this embodiment.

[0123] Those skilled in the art will appreciate that the embodiments of the present application may be provided as methods, systems, or computer program products. Therefore, the present application may adopt the form of a complete hardware embodiment, a complete software embodiment, or an embodiment in combination with software and hardware. Moreover, the present application may adopt the form of a computer program product implemented in one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) that include computer-usable program code.

[0124] The present application is described with reference to the flowcharts and / or block diagrams of the methods, devices (systems), and computer program products according to the embodiments of the present application. It should be understood that each process and / or box in the flowchart and / or block diagram, as well as the combination of the processes and / or boxes in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to generate a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0125] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0126] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process. Figure 1 A process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.

[0127] The specific implementation methods described above further illustrate the objectives, technical solutions and beneficial effects of the present invention in detail. It should be understood that the above description is only a specific implementation method of the present invention and is not intended to limit the scope of protection of the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the scope of protection of the present invention.

Claims

1. A TOE comprehensive performance tuning test method based on a multi-layer protocol stack, characterized in that: The method includes: Based on the performance indicators of the multi-layer protocol stack, a comprehensive performance calculation formula associated with each performance indicator is constructed for the link layer, network layer and transport layer where TOE acts, and a TOE comprehensive performance model is obtained; For multiple indicators in the TOE comprehensive performance model, multiple application scenarios corresponding to TOE are proposed; and for performance bottlenecks in different application scenarios, the value of TOE comprehensive performance is improved by adjusting influencing factors to obtain the optimal TOE comprehensive performance.

2. The TOE comprehensive performance tuning test method based on a multi-layer protocol stack according to claim 1 is characterized in that: The steps to obtain the TOE comprehensive performance model are: Based on the performance indicators of the multi-layer protocol stack, a comprehensive performance calculation formula related to each performance indicator is constructed; The TOE is tested, and various performance indicators are obtained by combining the test results of the link layer, network layer and transport layer; Substituting each performance index into the comprehensive performance calculation formula to obtain a TOE comprehensive performance model; Among them, the performance indicators include throughput indicators, connection indicators and CPU consumption indicators.

3. The TOE comprehensive performance tuning test method based on a multi-layer protocol stack according to claim 2 is characterized in that: The comprehensive performance calculation formula is: total=A*n1+B*n2+C*n3 In the formula, total is the total score of TOE comprehensive performance; A is the score of the throughput indicator; B is the score of the connection indicator; C is the score of the CPU consumption indicator; n1 is the weight of the throughput indicator; n2 is the weight of the connection indicator; n3 is the weight of the CPU consumption indicator.

4. The TOE comprehensive performance tuning test method based on a multi-layer protocol stack according to claim 3 is characterized in that: The expression of the TOE comprehensive performance model is: In the formula, total is the total score of TOE comprehensive performance; T 链路 is the link layer throughput indicator; T 网络 is the network layer throughput indicator; T 传输 is the transport layer throughput indicator; n1 is the weight of the throughput indicator; n2 is the weight of the connection indicator; n3 is the weight of the CPU consumption indicator; T 连接建立时间 TCP connection establishment time; T 连接延迟 is the TCP connection delay; T 重传率 is the retransmission rate of the TCP connection; T 丢包率 is the packet loss rate of the TCP connection; w1, w2, w3, and w4 are T 连接建立时间 , T 连接延迟 , T 重传率 , T 丢包率 The corresponding weight; cpu consumption is the cpu consumption indicator.

5. The TOE comprehensive performance tuning test method based on a multi-layer protocol stack according to claim 2 is characterized in that: The throughput index is based on the barrel effect, and the throughput index value finally takes the minimum value among the link layer throughput index, the network layer throughput index, and the transport layer throughput index; The test of the connection indicators refers to connection establishment time, connection delay, retransmission rate and stability.

6. The TOE comprehensive performance tuning test method based on a multi-layer protocol stack according to claim 2 is characterized in that: The CPU consumption indicator is to use the top or htop tool to directly monitor the CPU usage of the system and check the percentage of CPU occupied by TOE driver-related processes.

7. The TOE comprehensive performance tuning test method based on a multi-layer protocol stack according to claim 1 is characterized in that: TOE corresponds to multiple application scenarios and adjustments, including: The first TOE application scenario is a high-concurrency or bandwidth-constrained network, and the corresponding adjustment factor is the network bandwidth limitation; The second TOE application scenario is an unstable network, and the corresponding adjustment factor is the packet loss rate; The third TOE application scenario is a high-latency network, and the corresponding adjustment factor is latency; The fourth TOE application scenario is the flow control mechanism, and the corresponding adjustment factor is the sliding window size; The fifth TOE application scenario is small packet intensive requests or large packet traffic, and the corresponding adjustment factor is the data packet size; The sixth TOE application scenario is a complex multi-path network, and the corresponding adjustment factor is the routing strategy; The seventh TOE application scenario is the coexistence of short connections and long connections, and the corresponding adjustment factor is the TCP request type; The eighth TOE application scenario is conventional high-concurrency connection, and the corresponding adjustment influencing factor is normal TCP connection; The ninth TOE application scenario is sudden connection disconnection or half-open connection, and the corresponding adjustment influencing factor is TCP abnormal connection.

8. The TOE comprehensive performance tuning test method based on a multi-layer protocol stack according to claim 7 is characterized in that: For the performance in different application scenarios, the value of TOE comprehensive performance is improved by adjusting the influencing factors to obtain the optimal TOE comprehensive performance, including: According to the performance in different application scenarios, the weight parameters of various indicators in the TOE comprehensive performance model are adjusted accordingly; In a single application scenario, limit the restricted factors of the application scenario and adjust other unrestricted factors to maximize the total performance score to obtain the optimal TOE comprehensive performance. 9.TOE comprehensive performance tuning test system based on multi-layer protocol stack, characterized in that: The system includes: The performance model building unit is used to build a comprehensive performance calculation formula associated with each performance indicator for the link layer, network layer and transport layer of the TOE based on each performance indicator of the multi-layer protocol stack, so as to obtain a TOE comprehensive performance model; A TOE scenario proposing unit, used to propose multiple application scenarios corresponding to TOE for multiple indicators in the TOE comprehensive performance model; The TOE performance tuning unit is used to improve the value of TOE comprehensive performance by adjusting the influencing factors of performance bottlenecks in different application scenarios to obtain the optimal TOE comprehensive performance.

10. The TOE comprehensive performance tuning test system based on a multi-layer protocol stack according to claim 9, characterized in that: The execution process of the performance model building unit is: Based on the performance indicators of the multi-layer protocol stack, a comprehensive performance calculation formula related to each performance indicator is constructed; The TOE is tested, and various performance indicators are obtained by combining the test results of the link layer, network layer and transport layer; Substituting each performance index into the comprehensive performance calculation formula to obtain a TOE comprehensive performance model; Among them, the performance indicators include throughput indicators, connection indicators and CPU consumption indicators.

Images