Intelligent optimization method for message delay based on FPGA (Field Programmable Gate Array)

By connecting FPGA boards in network devices, calculating compatibility coefficients and standard deviations, and judging compatibility, the problem of message processing delay in short-term is solved, the ability to quickly process a large number of messages is realized, and the performance of network devices is reduced. Influence.

CN118945087AInactive Publication Date: 2024-11-12NANJING AOTONG INTELLIGENT TECH CO LTD

Patent Information

Application Number
CN202410976063.4
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-07-19
Publication Date
2024-11-12
Estimated Expiration
Not applicable · inactive patent

AI Technical Summary

Technical Problem

When a large number of messages need to be processed in the short term, the compatibility test of the FPGA-based message delay optimization technology and network equipment is not completed, resulting in delayed message processing and affecting network equipment performance.

Method used

By connecting the FPGA board to the network device, dividing the preset time interval to obtain throughput, identification and processing information, calculating compatibility coefficients, filtering out abnormal coefficients and standard deviations, judging compatibility, and deciding whether to use FPGA technology to process messages directly.

Benefits of technology

Reduce message processing delays, reduce the impact on overall network equipment performance, and allow the use of FPGA technology directly when processing large numbers of messages quickly in the short term, avoiding waiting for complete tests.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN118945087A_ABST
    Figure CN118945087A_ABST
Patent Text Reader

Abstract

The invention discloses an intelligent optimization method for message delay of an FPGA (Field Programmable Gate Array), which relates to the technical field of message processing, and comprises the following steps of: connecting an FPGA board card into a network environment of network equipment, and obtaining a throughput amplification coefficient, a data identification error coefficient and a data tampering coefficient in each subinterval to obtain a compatibility coefficient of the subinterval; the method comprises the following steps: selecting a plurality of sub-intervals, screening out abnormal coefficients in all the sub-intervals, calculating standard deviations in all the abnormal coefficients, comparing the standard deviations with a preset standard deviation value, and judging the compatibility of the FPGA-based message delay optimization technology and network equipment according to a comparison result, so that the compatibility between the FPGA-based message delay optimization technology and the network equipment is improved. When the compatibility between the FPGA-based message delay optimization technology and the network equipment is good, and the network equipment needs to quickly process a large number of messages in a short time, the network equipment directly uses the FPGA-based message delay optimization technology to carry out data processing on the messages, and compared with the existing FPGA-based message delay optimization, the message delay optimization method has the advantages that the message delay optimization efficiency is improved, and the message delay optimization efficiency is improved. In this way, more message processing delay can be reduced.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The invention relates to the technical field of message processing, and in particular to an intelligent optimization method for message delay based on FPGA. Background Art

[0002] FPGA-based message delay optimization is to design and implement specific hardware logic on the FPGA to reduce the delay in processing and forwarding packets within the network device. This optimization can significantly improve the performance and response speed of network equipment, especially for scenarios that require high-speed packet processing, such as routers and switches in data center networks and telecommunications networks. For example, a data center router uses FPGA-based message delay optimization technology. FPGA can implement customized packet processing logic, such as fast flow table matching and forwarding decisions, avoiding the additional delay caused by traditional software-based processing paths. As a result, the forwarding time of packets from the input port to the output port is greatly reduced, thereby significantly improving the throughput and response speed of the overall network.

[0003] Generally speaking, in order to make FPGA-based message delay optimization technology better used in existing network equipment, when introducing FPGA-based message delay optimization technology into network equipment, when determining the compatibility between FPGA-based message delay optimization technology and network equipment, it is also necessary to test FPGA-based message delay optimization technology and network equipment for a long time to ensure that network equipment can use FPGA-based message delay optimization technology for a long time; however, if the network equipment needs to process a large number of messages quickly in a short period of time, waiting for the complete FPGA and network equipment test to be completed may cause message processing delays, thereby affecting the performance of the overall network equipment. Summary of the invention

[0004] The purpose of the present invention is to solve the problem and propose an intelligent optimization method for message delay based on FPGA.

[0005] The present invention proposes an intelligent optimization method for message delay based on FPGA, the method comprising:

[0006] Connect the FPGA board to a network environment of a network device, obtain the compatibility between the FPGA and the network device within a preset time interval, and divide the preset time interval into several sub-intervals according to a preset time window;

[0007] Acquire data throughput information of network devices at different times in each sub-interval, and process the data throughput information to obtain a throughput increase coefficient of the sub-interval;

[0008] Acquire data identification information of network devices at different times in each sub-interval, and process the data identification information to obtain a data identification error coefficient of the sub-interval;

[0009] Acquire data processing information of network devices at different times in each sub-interval, and process the data processing information to obtain a data tampering coefficient of the sub-interval;

[0010] The compatibility coefficient of each sub-interval is obtained according to the throughput increase coefficient, data recognition error coefficient and data tampering coefficient in each sub-interval;

[0011] A compatible coefficient set is established based on the compatibility coefficients in all sub-intervals, and all abnormal coefficients in the compatible coefficient set are screened out, and an abnormal set is established for all the abnormal coefficients. The standard deviation in the abnormal set is calculated, and the standard deviation is compared with the preset standard deviation value. The compatibility of the FPGA-based message delay optimization technology and the network equipment is determined based on the comparison results.

[0012] Optionally, obtaining data throughput information of the network device at different times in each sub-interval, and processing the data throughput information to obtain the throughput increase coefficient of the sub-interval includes:

[0013] Obtain the actual data throughput number before the FPGA board is connected and the actual data throughput number after the FPGA board is connected of the network device at different times in each sub-interval, and mark the actual data throughput number before the FPGA board is connected and the actual data throughput number after the FPGA board is connected as Q respectively. 前 y and Q 后 y , y represents the number of actual data throughputs before the FPGA board is connected and the number of actual data throughputs after the FPGA board is connected of the network device at different times, y=1, 2, 3, 4, ..., m, m is the number of numbers and is a positive integer;

[0014] Q 前 y and Q 后 y After processing, the throughput increase coefficient is obtained. The specific calculation expression is: Where RF is the throughput increase factor.

[0015] Optionally, acquiring data identification information of a network device at different times in each sub-interval, and processing the data identification information to obtain a data identification error coefficient of the sub-interval includes:

[0016] The data recognition error coefficient includes a data recognition missing coefficient and a data recognition delay coefficient;

[0017] The data identification missing coefficients include:

[0018] Get the number of data actually recognized by the network device at different times in each sub-interval and mark it as S 实 g , g represents the number of data actually recognized by the network device at different times, g = 1, 2, 3, 4, ..., n, n is the number of numbers and is a positive integer;

[0019] Get the number of preset recognized data of the network device at different times in each sub-interval and mark it as S 预 f , f represents the number of preset identified data of the network device at different times, f=1, 2, 3, 4, ..., l, l is the number of numbers and is a positive integer;

[0020] Calculate the data recognition missing coefficient, the calculation formula is: UY = ln (l / n + 1), where UY is the data recognition missing coefficient.

[0021] Optionally, the data identification delay coefficient includes:

[0022] Obtain the highest threshold of the time interval of the effective identification data of the network device at different times in each sub-interval, and mark the highest threshold as F G , and 0 <F G ;

[0023] Obtain the time interval of the data actually recognized by the network device at different times in each sub-interval, and mark the time interval of the data actually recognized as F 实 s , s represents the number of the time interval of the data actually recognized by the network device at different times, s = 1, 2, 3, 4, ..., k, k is the number of numbers and is a positive integer;

[0024] will be greater than F G The time interval of the actual recognition data of the network device is recorded as F 大 j ,j is greater than F G The number of the time interval of the data actually recognized by the network device, j = 1, 2, 3, 4, ..., b, b is the number of numbers and is a positive integer;

[0025] Calculate the data recognition delay coefficient, the calculation expression is: DH is the data recognition delay factor;

[0026] The data recognition omission coefficient and the data recognition delay coefficient are processed to obtain the data recognition error coefficient, and the calculation expression is: KU = α×RF+β×GM; where KU is the data recognition error coefficient, α and β are the preset proportional coefficients of the data recognition omission coefficient and the data recognition delay coefficient, respectively, and α and β are both greater than 0.

[0027] Optionally, obtaining data processing information of a network device at different times in each sub-interval, and processing the data processing information to obtain a data tampering coefficient of the sub-interval includes:

[0028] Convert the data not processed by the FPGA-based message delay optimization technology into binary, and add a certain number of zero bits at the end of each data packet, so that the data not processed by the FPGA-based message delay optimization technology and the data processed by the FPGA-based message delay optimization technology are received in the same format;

[0029] Performing a modulo-2 division operation on the binary value of the data that has not been processed by the FPGA-based message delay optimization technology to obtain a remainder;

[0030] The remainder is added to the end of the data that is not processed by the message delay optimization technology based on the FPGA to obtain a check code of the data that is not processed by the message delay optimization technology based on the FPGA;

[0031] The data processed by the FPGA-based message delay optimization technology is also converted into the same binary as above, and the steps as described above are performed to obtain the check code of the data processed by the FPGA-based message delay optimization technology;

[0032] The check codes of the data not processed by the FPGA-based message delay optimization technology and the check codes of the data processed by the FPGA-based message delay optimization technology are compared, the number of consistent comparisons is recorded as GH, the number of inconsistent comparisons is recorded as GS, and the data tampering coefficient is calculated. The calculation expression is: KH=GS / GH+GS, where KH is the data tampering coefficient.

[0033] Optionally, obtaining the compatibility coefficient of each sub-interval according to the throughput increase coefficient, the data recognition error coefficient, and the data tampering coefficient in the sub-interval includes:

[0034] The throughput increase coefficient, data recognition error coefficient and data tampering coefficient of each sub-interval are weighted and summed to obtain the compatibility coefficient of the sub-interval. The calculation formula is:

[0035]

[0036] Wherein, Lgt is the compatibility coefficient, RF, KU, and KH are the throughput increase coefficient, data recognition error coefficient, and data tampering coefficient, respectively; a1, a2, and a3 are the preset proportional coefficients of the throughput increase coefficient, data recognition error coefficient, and data tampering coefficient, respectively; and a1, a2, and a3 are all greater than 0.

[0037] Optionally, a compatible coefficient set is established based on the compatible coefficients in all subintervals, and all abnormal coefficients in the compatible coefficient set are screened out, including:

[0038] A compatible coefficient set is established according to the compatible coefficients in all sub-intervals, and all abnormal coefficients in the compatible coefficient set are screened out;

[0039] When the compatibility coefficient is not less than the preset compatibility coefficient reference threshold, the compatibility coefficient is marked as an abnormal coefficient;

[0040] When the compatibility coefficient is less than a preset compatibility coefficient reference threshold, the compatibility coefficient is marked as a normal coefficient.

[0041] Optionally, all abnormal coefficients are used to establish an abnormal set, and the standard deviation in the abnormal set is calculated, and the standard deviation is compared with a preset standard deviation value. The compatibility of the FPGA-based message delay optimization technology and the network device is determined according to the comparison result, including:

[0042] All abnormal coefficients are used to establish an abnormal set, and the standard deviation in the abnormal set is calculated. The standard deviation is compared with the preset standard deviation value, and the accuracy of the analysis results output by the system is determined based on the comparison results;

[0043] A data set is established by using several abnormal coefficients and calibrated as W, then W = {TG d}, where h = {1, 2, 3…o}, o is a positive integer;

[0044] And calculate the standard deviation of several abnormal coefficients in the data set, and calibrate the standard deviation of several abnormal coefficients in the data set as tgf, and compare the abnormal coefficient standard deviation tgf with the preset abnormal coefficient standard deviation threshold tyu;

[0045] If the abnormal coefficient standard deviation tgf is not less than the preset abnormal coefficient standard deviation threshold tyu, a special alarm signal is issued, and the FPGA-based message delay optimization technology and network equipment are tested for a long time;

[0046] If the abnormal coefficient standard deviation tgf is not less than the preset abnormal coefficient standard deviation threshold tyu, no special alarm signal will be issued. At this time, there is no need to conduct long-term testing of FPGA-based message delay optimization technology and network equipment. The messages of the network equipment can be directly processed through the FPGA message delay optimization technology.

[0047] Beneficial effects of the present invention:

[0048] The present invention proposes an intelligent optimization method for message delay based on FPGA, which connects an FPGA board to a network environment of a network device, obtains a throughput increase coefficient, a data recognition error coefficient and a data tampering coefficient in each sub-interval, and obtains a compatibility coefficient of the sub-interval. A compatibility coefficient set is established according to the compatibility coefficients in all sub-intervals, and all abnormal coefficients in the compatibility coefficient set are screened out, and an abnormal set is established for all abnormal coefficients, and the standard deviation in the abnormal set is calculated, and the standard deviation is compared with a preset standard deviation value. The compatibility of the message delay optimization technology based on FPGA and the network device is judged according to the comparison result. In this way, when the compatibility of the message delay optimization technology based on FPGA and the network device is good, when the network device needs to quickly process a large number of messages in a short period of time, the network device directly uses the message delay optimization technology based on FPGA to process data of the message. Compared with the existing message delay optimization based on FPGA, this method can reduce more message processing delay and reduce the impact on the performance of the overall network device. BRIEF DESCRIPTION OF THE DRAWINGS

[0049] The present invention will be further described below in conjunction with the accompanying drawings.

[0050] Figure 1 The present invention provides a flow chart of an intelligent optimization method for message delay based on FPGA. DETAILED DESCRIPTION

[0051] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0052] Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in the field without making any creative work shall fall within the scope of protection of the present invention.

[0053] The embodiment of the present invention provides an intelligent optimization method for message delay based on FPGA. Figure 1 , Figure 1 A flowchart of an intelligent optimization method for message delay based on FPGA provided by an embodiment of the present invention. The method comprises the following steps:

[0054] Connect the FPGA board to a network environment of a network device, obtain the compatibility between the FPGA and the network device within a preset time interval, and divide the preset time interval into several sub-intervals according to a preset time window;

[0055] Acquire data throughput information of network devices at different times in each sub-interval, and process the data throughput information to obtain a throughput increase coefficient of the sub-interval;

[0056] Acquire data identification information of network devices at different times in each sub-interval, and process the data identification information to obtain a data identification error coefficient of the sub-interval;

[0057] Acquire data processing information of network devices at different times in each sub-interval, and process the data processing information to obtain a data tampering coefficient of the sub-interval;

[0058] The compatibility coefficient of each sub-interval is obtained according to the throughput increase coefficient, data recognition error coefficient and data tampering coefficient in each sub-interval;

[0059] A compatible coefficient set is established based on the compatibility coefficients in all sub-intervals, and all abnormal coefficients in the compatible coefficient set are screened out, and an abnormal set is established for all the abnormal coefficients. The standard deviation in the abnormal set is calculated, and the standard deviation is compared with the preset standard deviation value. The compatibility of the FPGA-based message delay optimization technology and the network equipment is determined based on the comparison results.

[0060] Based on the FPGA-based intelligent optimization method for message delay provided by the embodiment of the present invention, the present invention proposes an FPGA-based intelligent optimization method for message delay, by connecting the FPGA board to the network environment of the network device, and obtaining the throughput increase coefficient, data recognition error coefficient and data tampering coefficient in each sub-interval to obtain the compatibility coefficient of the sub-interval, establishing a compatibility coefficient set according to the compatibility coefficients in all sub-intervals, and screening out all abnormal coefficients in the compatibility coefficient set, and establishing an abnormal set for all abnormal coefficients, and calculating the standard deviation in the abnormal set, and comparing the standard deviation with the preset standard deviation value, judging the compatibility of the FPGA-based message delay optimization technology and the network device according to the comparison result, so that when the compatibility of the FPGA-based message delay optimization technology and the network device is good, when the network device needs to quickly process a large number of messages in a short period of time, the network device directly uses the FPGA-based message delay optimization technology to process the message data, compared with the existing FPGA-based message delay optimization, this method can reduce more message processing delays and reduce the impact on the performance of the overall network device.

[0061] It should be noted that the preset time interval and the preset time window are set by professionals according to actual circumstances, and are not specifically limited or elaborated.

[0062] In one embodiment, data throughput information of the network device at different times in each sub-interval is obtained, and the data throughput information is processed to obtain the throughput increase coefficient of the sub-interval, including:

[0063] Obtain the actual data throughput number before the FPGA board is connected and the actual data throughput number after the FPGA board is connected of the network device at different times in each sub-interval, and mark the actual data throughput number before the FPGA board is connected and the actual data throughput number after the FPGA board is connected as Q respectively. 前 y and Q 后 y , y represents the number of actual data throughputs before the FPGA board is connected and the number of actual data throughputs after the FPGA board is connected of the network device at different times, y=1, 2, 3, 4, ..., m, m is the number of numbers and is a positive integer;

[0064] Q 前 y and Q 后 y After processing, the throughput increase coefficient is obtained. The specific calculation expression is: Where RF is the throughput increase factor.

[0065] It should be noted that the actual data throughput number before the FPGA board of the network device is connected and the actual data throughput number after the FPGA board is connected can be obtained through the monitoring system of the network device, or other methods, which are not specifically limited.

[0066] It should be noted that the throughput increase coefficient refers to the increase in data throughput of the network device before and after the FPGA board is connected; if the data throughput of the network device before and after the FPGA board is connected is large, it means that the compatibility between the FPGA-based message delay optimization technology and the network device is good. If the network device needs to process a large number of messages quickly in a short period of time, it does not need to wait for the complete FPGA and network device test to complete before processing the message. The message delay optimization technology based on FPGA can be directly used to process the message through the network device, thereby reducing the delay of message processing and reducing the impact on the performance of the overall network device. The reason is that when the throughput increase is detected to be significant, it indicates that FPGA technology has shown good adaptability and performance improvement in the network device. In this case, even if the test has not been fully completed, FPGA technology can be confidently introduced to process a large number of messages. If the throughput is significantly improved in a short period of time, even if a long-term stability test has not been conducted, FPGA technology can be considered for emergency situations where the network device needs to process a large number of messages.

[0067] In one embodiment, data identification information of network devices at different times in each sub-interval is obtained, and the data identification information is processed to obtain a data identification error coefficient of the sub-interval, including:

[0068] The data recognition error coefficient includes a data recognition missing coefficient and a data recognition delay coefficient;

[0069] The data identification missing coefficients include:

[0070] Get the number of data actually recognized by the network device at different times in each sub-interval and mark it as S 实 g , g represents the number of data actually recognized by the network device at different times, g = 1, 2, 3, 4, ..., n, n is the number of numbers and is a positive integer;

[0071] Get the number of preset recognized data of the network device at different times in each sub-interval and mark it as S 预 f , f represents the number of preset identified data of the network device at different times, f=1, 2, 3, 4, ..., l, l is the number of numbers and is a positive integer;

[0072] Calculate the data identification missing coefficient, the calculation formula is: UY = ln (l / n + 1), where UY is the data identification missing coefficient;

[0073] The data identification delay coefficient includes:

[0074] Obtain the highest threshold of the time interval of the effective identification data of the network device at different times in each sub-interval, and mark the highest threshold as F G , and 0 <F G ;

[0075] Obtain the time interval of the data actually recognized by the network device at different times in each sub-interval, and mark the time interval of the data actually recognized as F 实 s , s represents the number of the time interval of the data actually recognized by the network device at different times, s = 1, 2, 3, 4, ..., k, k is the number of numbers and is a positive integer;

[0076] will be greater than F G The time interval of the actual recognition data of the network device is recorded as F 大 j ,j is greater than F G The number of the time interval of the data actually recognized by the network device, j = 1, 2, 3, 4, ..., b, b is the number of numbers and is a positive integer;

[0077] Calculate the data recognition delay coefficient, the calculation expression is: DH is the data recognition delay factor;

[0078] The data recognition omission coefficient and the data recognition delay coefficient are processed to obtain the data recognition error coefficient, and the calculation expression is: KU = α×RF+β×GM; where KU is the data recognition error coefficient, α and β are the preset proportional coefficients of the data recognition omission coefficient and the data recognition delay coefficient, respectively, and α and β are both greater than 0.

[0079] It should be noted that the maximum threshold of the time interval for effectively identified data and the preset number of data to be identified are set by professionals based on actual circumstances and are not specifically limited. In addition, the actual number of data identified by network devices at different times and the actual time interval for data identified by network devices at different times can be obtained through the monitoring system of the network devices or other methods, which are not specifically limited.

[0080] It should be noted that the data recognition error coefficient refers to the degree of data recognition omissions and data recognition delays of network devices in FPGA. If the degree of data recognition omissions and data recognition delays is large, it means that the compatibility between the FPGA-based message delay optimization technology and the network device is not good. The reasons are:

[0081] Data identification omission degree: Data integrity problem: When the data identification omission degree is large, it means that the network device has lost some messages during the communication with the FPGA board. This data loss may be due to incompatibility of the interface protocol, errors in the data transmission process, or timing problems. This will directly affect the integrity and reliability of data transmission.

[0082] Signal mismatch: There may be a mismatch in signal level or timing between the FPGA and the network device, resulting in some data in the message not being correctly recognized. This indicates that there is a compatibility issue between the two parties at the hardware or protocol level.

[0083] Data recognition delay: Transmission delay: A large data recognition delay indicates that there is a large delay in data transmission between the FPGA and the network device. This may be due to the mismatch between the two parties in data transmission protocols, timing coordination, etc., resulting in additional time for data packets to be processed and transmitted.

[0084] Processing delays: There may also be large delays in message processing within the FPGA or within the network device. This may be because the FPGA design does not fully take into account the specific needs of the network device, or the network device does not properly optimize the FPGA's processing mechanism.

[0085] Summary: Compatibility issues: Data recognition omissions and delays indicate compatibility issues between FPGA and network devices, and data cannot be effectively exchanged. FPGA message delay optimization technology may not fully consider the needs of network devices, or network devices are not optimized for FPGA characteristics, resulting in unstable or inefficient data transmission.

[0086] System efficiency: High leakage rate and high latency will lead to reduced system efficiency, and the reliability and real-time performance of data transmission cannot be guaranteed, which has a negative impact on network performance and user experience. Therefore, ensuring good compatibility between FPGA boards and network devices is crucial to achieve efficient message delay optimization.

[0087] Therefore, the degree of data recognition omissions and delays is relatively large, which directly reflects the compatibility issues between FPGA and network equipment.

[0088] In one embodiment, data processing information of the network device at different times in each sub-interval is obtained, and the data processing information is processed to obtain the data tampering coefficient of the sub-interval, including:

[0089] Convert the data not processed by the FPGA-based message delay optimization technology into binary, and add a certain number of zero bits at the end of each data packet, so that the data not processed by the FPGA-based message delay optimization technology and the data processed by the FPGA-based message delay optimization technology are received in the same format;

[0090] Performing a modulo-2 division operation on the binary value of the data that has not been processed by the FPGA-based message delay optimization technology to obtain a remainder;

[0091] The remainder is added to the end of the data that is not processed by the message delay optimization technology based on the FPGA to obtain a check code of the data that is not processed by the message delay optimization technology based on the FPGA;

[0092] The data processed by the FPGA-based message delay optimization technology is also converted into the same binary as above, and the steps as described above are performed to obtain the check code of the data processed by the FPGA-based message delay optimization technology;

[0093] The check codes of the data not processed by the FPGA-based message delay optimization technology and the check codes of the data processed by the FPGA-based message delay optimization technology are compared, the number of consistent comparisons is recorded as GH, the number of inconsistent comparisons is recorded as GS, and the data tampering coefficient is calculated. The calculation expression is: KH=GS / GH+GS, where KH is the data tampering coefficient.

[0094] It should be noted that the data tampering coefficient refers to the degree of data identification tampering by the network device in the FPGA; if the degree of data identification tampering is large, it means that the compatibility between the FPGA-based message delay optimization technology and the network device is not good. The reason is that when data tampering occurs during the data transmission between the FPGA and the network device, it means that the transmitted data is modified or damaged to some extent. This may be caused by noise in signal transmission, incorrect protocol implementation, or hardware failure. These tampering directly affects the integrity of the data, indicating that there is a problem with the compatibility between the two.

[0095] In one embodiment, the compatibility coefficient of each sub-interval is obtained according to the throughput increase coefficient, the data recognition error coefficient and the data tampering coefficient in the sub-interval, including:

[0096] The throughput increase coefficient, data recognition error coefficient and data tampering coefficient of each sub-interval are weighted and summed to obtain the compatibility coefficient of the sub-interval. The calculation formula is:

[0097]

[0098] Wherein, Lgt is the compatibility coefficient, RF, KU, and KH are the throughput increase coefficient, data recognition error coefficient, and data tampering coefficient, respectively; a1, a2, and a3 are the preset proportional coefficients of the throughput increase coefficient, data recognition error coefficient, and data tampering coefficient, respectively; and a1, a2, and a3 are all greater than 0.

[0099] In one embodiment, a compatible coefficient set is established based on the compatible coefficients in all sub-intervals, and all abnormal coefficients in the compatible coefficient set are screened out, including:

[0100] A compatible coefficient set is established according to the compatible coefficients in all sub-intervals, and all abnormal coefficients in the compatible coefficient set are screened out;

[0101] When the compatibility coefficient is not less than the preset compatibility coefficient reference threshold, it indicates that the possibility of compatibility problems between the FPGA-based message delay optimization technology and the network device is greater, and the compatibility coefficient is marked as an abnormal coefficient;

[0102] When the compatibility coefficient is less than the preset compatibility coefficient reference threshold, it means that the compatibility possibility between the FPGA-based message delay optimization technology and the network device is smaller, and the compatibility coefficient is marked as a normal coefficient.

[0103] In one implementation, the above-mentioned preset compatibility coefficient reference threshold is set by professionals according to actual conditions, and the details are not repeated here.

[0104] In one embodiment, all abnormal coefficients are used to establish an abnormal set, and the standard deviation in the abnormal set is calculated, and the standard deviation is compared with a preset standard deviation value. The compatibility of the FPGA-based message delay optimization technology and the network device is determined according to the comparison result, including:

[0105] All abnormal coefficients are used to establish an abnormal set, and the standard deviation in the abnormal set is calculated. The standard deviation is compared with the preset standard deviation value, and the accuracy of the analysis results output by the system is determined based on the comparison results;

[0106] A data set is established by using several abnormal coefficients and calibrated as W, then W = {TG d}, where h = {1, 2, 3…o}, o is a positive integer;

[0107] And calculate the standard deviation of several abnormal coefficients in the data set, and calibrate the standard deviation of several abnormal coefficients in the data set as tgf, and compare the abnormal coefficient standard deviation tgf with the preset abnormal coefficient standard deviation threshold tyu;

[0108] If the abnormal coefficient standard deviation tgf is not less than the preset abnormal coefficient standard deviation threshold tyu, a special alarm signal is issued, and the FPGA-based message delay optimization technology and network equipment are tested for a long time;

[0109] If the abnormal coefficient standard deviation tgf is not less than the preset abnormal coefficient standard deviation threshold tyu, no special alarm signal will be issued. At this time, there is no need to conduct long-term testing of FPGA-based message delay optimization technology and network equipment. The messages of the network equipment can be directly processed through the FPGA message delay optimization technology.

[0110] In one implementation method, when a special alarm signal is issued, it indicates that the compatibility between the FPGA-based message delay optimization technology and the network equipment is not high. Although the network equipment needs to process a large number of messages quickly in a short period of time, due to the low compatibility, it is still necessary to wait for the complete FPGA and network equipment test to be completed and remind the staff at the same time before processing the message, thereby reducing the impact on the performance of the overall network equipment; when no special alarm signal is issued, it indicates that the compatibility between the FPGA-based message delay optimization technology and the network equipment is high. When the network equipment needs to process a large number of messages quickly in a short period of time, it is not necessary to wait for the complete FPGA and network equipment test to be completed before processing the message. The message can be processed by the FPGA-based message delay optimization technology. In this way, the delay of FPGA-based message processing is reduced, so that the message can be processed faster and more quickly.

[0111] The above is a detailed description of an embodiment of the present invention, but the content is only a preferred embodiment of the present invention and cannot be considered to limit the scope of implementation of the present invention. All equivalent changes and improvements made within the scope of the present invention should still fall within the scope of the patent coverage of the present invention.

Claims

1. An intelligent optimization method for message delay based on FPGA, characterized in that: The following steps are involved: Connect the FPGA board to a network environment of a network device, obtain the compatibility between the FPGA and the network device within a preset time interval, and divide the preset time interval into several sub-intervals according to a preset time window; Acquire data throughput information of network devices at different times in each sub-interval, and process the data throughput information to obtain a throughput increase coefficient of the sub-interval; Acquire data identification information of network devices at different times in each sub-interval, and process the data identification information to obtain a data identification error coefficient of the sub-interval; Acquire data processing information of network devices at different times in each sub-interval, and process the data processing information to obtain a data tampering coefficient of the sub-interval; The compatibility coefficient of each sub-interval is obtained according to the throughput increase coefficient, data recognition error coefficient and data tampering coefficient in the sub-interval; A compatible coefficient set is established based on the compatibility coefficients in all sub-intervals, and all abnormal coefficients in the compatible coefficient set are screened out, and an abnormal set is established for all the abnormal coefficients. The standard deviation in the abnormal set is calculated, and the standard deviation is compared with the preset standard deviation value. The compatibility of the FPGA-based message delay optimization technology and the network equipment is determined based on the comparison results.

2. The intelligent optimization method for message delay based on FPGA according to claim 1, characterized in that: The data throughput information of the network device at different times in each sub-interval is obtained, and the data throughput information is processed to obtain the throughput increase coefficient of the sub-interval, including: Obtain the actual data throughput number before the FPGA board is connected and the actual data throughput number after the FPGA board is connected of the network device at different times in each sub-interval, and mark the actual data throughput number before the FPGA board is connected and the actual data throughput number after the FPGA board is connected as Q respectively. 前 y and Q 后 y , y represents the number of actual data throughputs before the FPGA board is connected and the number of actual data throughputs after the FPGA board is connected of the network device at different times, y=1, 2, 3, 4, ..., m, m is the number of numbers and is a positive integer; Q 前 y and Q 后 y After processing, the throughput increase coefficient is obtained. The specific calculation expression is: Where RF is the throughput increase factor.

3. The intelligent optimization method for message delay based on FPGA according to claim 1, characterized in that: The data identification information of the network device at different times in each sub-interval is obtained, and the data identification information is processed to obtain the data identification error coefficient of the sub-interval, including: The data recognition error coefficient includes a data recognition missing coefficient and a data recognition delay coefficient; The data identification missing coefficients include: Get the number of data actually recognized by the network device at different times in each sub-interval and mark it as S 实 g , g represents the number of data actually recognized by the network device at different times, g = 1, 2, 3, 4, ..., n, n is the number of numbers and is a positive integer; Get the number of preset recognized data of the network device at different times in each sub-interval and mark it as S 预 f , f represents the number of preset identified data of the network device at different times, f=1, 2, 3, 4, ..., l, l is the number of numbers and is a positive integer; Calculate the data recognition missing coefficient, the calculation formula is: UY = ln (l / n + 1), where UY is the data recognition missing coefficient.

4. The intelligent optimization method for message delay based on FPGA according to claim 3, characterized in that: The data identification delay coefficient includes: Obtain the highest threshold of the time interval of the effective identification data of the network device at different times in each sub-interval, and mark the highest threshold as F G , and 0 <F G ; Obtain the time interval of the data actually recognized by the network device at different times in each sub-interval, and mark the time interval of the data actually recognized as F 实 s , s represents the number of the time interval of the data actually recognized by the network device at different times, s = 1, 2, 3, 4, ..., k, k is the number of numbers and is a positive integer; will be greater than F G The time interval of the actual recognition data of the network device is recorded as F 大 j ,j is greater than F G The number of the time interval of the data actually recognized by the network device, j = 1, 2, 3, 4, ..., b, b is the number of numbers and is a positive integer; Calculate the data recognition delay coefficient, the calculation expression is: DH is the data recognition delay factor; The data recognition omission coefficient and the data recognition delay coefficient are processed to obtain the data recognition error coefficient, and the calculation expression is: KU = α×RF+β×GM; where KU is the data recognition error coefficient, α and β are the preset proportional coefficients of the data recognition omission coefficient and the data recognition delay coefficient, respectively, and α and β are both greater than 0.

5. The intelligent optimization method for message delay based on FPGA according to claim 1, characterized in that: The data processing information of the network device at different times in each sub-interval is obtained, and the data processing information is processed to obtain the data tampering coefficient of the sub-interval, including: Convert the data not processed by the FPGA-based message delay optimization technology into binary, and add a certain number of zero bits at the end of each data packet, so that the data not processed by the FPGA-based message delay optimization technology and the data processed by the FPGA-based message delay optimization technology are received in the same format; Performing a modulo-2 division operation on the binary value of the data that has not been processed by the FPGA-based message delay optimization technology to obtain a remainder; The remainder is added to the end of the data that is not processed by the message delay optimization technology based on the FPGA to obtain a check code of the data that is not processed by the message delay optimization technology based on the FPGA; The data processed by the FPGA-based message delay optimization technology is also converted into the same binary as above, and the steps as described above are performed to obtain the check code of the data processed by the FPGA-based message delay optimization technology; The check codes of the data not processed by the FPGA-based message delay optimization technology and the check codes of the data processed by the FPGA-based message delay optimization technology are compared, the number of consistent comparisons is recorded as GH, the number of inconsistent comparisons is recorded as GS, and the data tampering coefficient is calculated. The calculation expression is: KH=GS / GH+GS, where KH is the data tampering coefficient.

6. The intelligent optimization method for message delay based on FPGA according to claim 1, characterized in that: The compatibility coefficient of each sub-interval is obtained based on the throughput increase coefficient, data recognition error coefficient and data tampering coefficient in each sub-interval, including: The throughput increase coefficient, data recognition error coefficient and data tampering coefficient of each sub-interval are weighted and summed to obtain the compatibility coefficient of the sub-interval. The calculation formula is: Wherein, Lgt is the compatibility coefficient, RF, KU, and KH are the throughput increase coefficient, data recognition error coefficient, and data tampering coefficient, respectively; a1, a2, and a3 are the preset proportional coefficients of the throughput increase coefficient, data recognition error coefficient, and data tampering coefficient, respectively; and a1, a2, and a3 are all greater than 0.

7. The intelligent optimization method for message delay based on FPGA according to claim 1, characterized in that: A compatible coefficient set is established based on the compatible coefficients in all sub-intervals, and all abnormal coefficients in the compatible coefficient set are screened out, including: A compatible coefficient set is established according to the compatible coefficients in all sub-intervals, and all abnormal coefficients in the compatible coefficient set are screened out; When the compatibility coefficient is not less than the preset compatibility coefficient reference threshold, the compatibility coefficient is marked as an abnormal coefficient; When the compatibility coefficient is less than a preset compatibility coefficient reference threshold, the compatibility coefficient is marked as a normal coefficient.

8. The intelligent optimization method for message delay based on FPGA according to claim 1, characterized in that: All abnormal coefficients are used to establish an abnormal set, and the standard deviation in the abnormal set is calculated. The standard deviation is compared with the preset standard deviation value. According to the comparison results, the compatibility of the FPGA-based message delay optimization technology and the network equipment is judged, including: All abnormal coefficients are used to establish an abnormal set, and the standard deviation in the abnormal set is calculated. The standard deviation is compared with the preset standard deviation value, and the accuracy of the analysis results output by the system is determined based on the comparison results; A data set is established by using several abnormal coefficients and calibrated as W, then W = {TG d }, where h = {1, 2, 3…o}, o is a positive integer; And calculate the standard deviation of several abnormal coefficients in the data set, and calibrate the standard deviation of several abnormal coefficients in the data set as tgf, and compare the abnormal coefficient standard deviation tgf with the preset abnormal coefficient standard deviation threshold tyu; If the abnormal coefficient standard deviation tgf is not less than the preset abnormal coefficient standard deviation threshold tyu, a special alarm signal is issued, and the FPGA-based message delay optimization technology and network equipment are tested for a long time; If the abnormal coefficient standard deviation tgf is not less than the preset abnormal coefficient standard deviation threshold tyu, no special alarm signal will be issued. At this time, there is no need to conduct long-term testing of FPGA-based message delay optimization technology and network equipment. The messages of the network equipment can be directly processed through the FPGA message delay optimization technology.

Citation Information

Patent Citations

  • Feiteng platform FPGA accelerator card introduction test method and system

    CN113157515A

  • Network equipment compatibility test method, terminal equipment and storage medium

    CN113824612A

  • Interconnection interface test method based on chip FPGA prototype verification equipment

    CN115269293A

  • Hardware software communication channel to support direct programming interface methods on FPGA-based prototype platforms

    US20230401132A1

Cited By

  • Network security model comprehensive evaluation system and evaluation method thereof

    CN119210859A