A method, system, device and medium for detecting quality of a network card link

Abstract

The application discloses a network card link quality detection method, system and device and a medium, and relates to the field of signal transmission. The application solves the problem of lacking a unified standard for evaluating network card link quality. The number of connectors between a central processor and a network card, the number and size of cables are used to score the network card link structure to obtain a first score; a signal transmission test is performed on the central processor and the network card to obtain a second score; and the quality score of each network card link is obtained according to the first score and the second score. The number of connectors, the number and length of cables are related to signal attenuation and interference, and the number of connectors, the number and length of cables are included in the scoring system, so that the advantages and disadvantages of different links in structure can be quantitatively evaluated; in combination with the signal transmission test result, the actual performance of the network card link can be more comprehensively reflected, the signal quality of multiple network card links can be uniformly quantitatively evaluated, and then the quality of multiple links can be sorted according to the quality score, so that the link with the optimal quality can be selected.

Description

Technical Field

[0001] The present invention relates to the field of signal transmission, and in particular to a method, system, device and medium for detecting the quality of a network card link. Background Art

[0002] As a high-performance network interface card designed specifically for data centers, the OCP (Open Compute Project) network interface card (NIC) promotes standardization and interoperability of data center hardware through open standards and modular design, thereby reducing costs and improving efficiency. OCP NICs, with their standardized design, high performance, flexibility, manageability, and energy efficiency, have become a vital component of modern data center network infrastructure. They not only improve network performance and reliability, but also simplify deployment and maintenance, reducing total cost of ownership. OCP NICs communicate with central processing units (CPUs) using PCIe (Peripheral Component Interconnect Express), a high-performance, low-latency computer expansion bus standard widely used in various computing devices. Its high-speed transmission, low latency, hot-swappable and hot-removable, flexibility, and power management make it an indispensable component of modern computing systems. As PCIe speeds continue to increase, the signal quality requirements for OCP NICs are also becoming increasingly stringent.

[0003] However, there is a problem with the current design of OCP network card links in servers: although more and more OCP network card links have been designed, each link is currently measured separately to determine whether its quality is qualified. There is a lack of unified standards to evaluate the signal quality of OCP network card links, resulting in the inability to rank the quality of multiple OCP network card links and thus unable to select the link with the best quality from multiple OCP network card links. Summary of the Invention

[0004] The purpose of the present invention is to provide a network card link quality detection method, system, device and medium, which can uniformly quantify the signal quality of multiple network card links, and then rank the quality of multiple links according to the quality score, so as to select the link with the best quality.

[0005] In a first aspect, the present application provides a method for detecting the quality of a network card link, comprising:

[0006] For each of the network card links, obtaining structural parameters of the network card link and scoring the structural parameters to obtain a first score; the structural parameters include the number of connectors, the number of cables, and the length of the cables in the network card link; the network card link is a connection link structure between the network card and a central processing unit;

[0007] Performing a signal transmission test between the central processing unit and the network card to obtain a test result, and scoring the test result to obtain a second score;

[0008] A quality score of each network card link is obtained by performing calculation according to the first score and the second score in a preset manner.

[0009] Optionally, before scoring the structural parameters to obtain the first score, the method further includes:

[0010] According to the structural parameters of the network card link and a first preset level classification method, obtaining a structural level corresponding to the structural parameters of each network card link;

[0011] Scoring the test result to obtain the second score further includes:

[0012] Obtaining a test level corresponding to each test result of the network card link according to the test result of the network card link and a second preset level classification method; the test level includes one or more of a test value not reaching a standard test value, a test value reaching a standard test value, and a standard test value reaching a preset multiple, where the preset multiple is a value greater than 1 and less than 1.5;

[0013] Scoring the structural parameters to obtain a first score includes:

[0014] determining, according to the structural level corresponding to the structural parameter of each of the network card links, a first score corresponding to the structural level;

[0015] Scoring the test result to obtain a second score includes:

[0016] According to the test level corresponding to the test result of each network card link, a second score corresponding to the test level is determined.

[0017] Optionally, before determining the first score corresponding to the structural level according to the structural parameter of each network card link, the method further includes:

[0018] Eliminating network card links of a preset structural level from all the network card links; the number of connectors in the network card links of the preset structural level is greater than a first preset number and / or the number of cables is greater than a second preset number and / or the length of the cables is greater than a preset length;

[0019] Before determining the second score corresponding to the test level according to the test result of each network card link, the method further comprises:

[0020] Eliminating the network card link of a preset test level from all the network card links; the preset test level is a level not reaching the standard test value.

[0021] Optionally, the signal transmission test between the central processor and the network card is performed to obtain a test result, comprising:

[0022] The network card is sequentially executed the test of shutdown and startup to obtain a first test result;

[0023] The network card is sequentially executed the second test of shutdown, power-off, power-on and startup to obtain a second test result;

[0024] The test result comprises at least the first test result and the second test result.

[0025] Optionally, the method further comprises:

[0026] The first test result is assigned a first weight, and the second test result is assigned a second weight;

[0027] The restart test score of the network card link is obtained according to the first weight and the first test result, and the second weight and the second test result.

[0028] Optionally, the signal transmission test between the central processor and the network card is performed to obtain a test result, comprising:

[0029] The original signal and the original data packet are sent from the central processor, and the test signal and the test data packet received by the network card are obtained; the original signal corresponds to the test signal, and the original data packet corresponds to the test data packet;

[0030] The test result is obtained according to the signal parameter of the test signal and the data packet parameter of the test data packet;

[0031] The signal parameter is a parameter for characterizing signal quality, and the data packet parameter is a parameter for characterizing data transmission performance.

[0032] Optionally, the quality score of each network card link is obtained by calculating according to the first score and the second score in a preset manner, comprising:

[0033] A first influence factor corresponding to the first score and a second influence factor corresponding to the second score are determined;

[0034] A quality score of each of the network card links is determined according to the first impact factor, the first score, the second impact factor, and the second score.

[0035] In a second aspect, the present application provides a network card link quality detection system, comprising:

[0036] a first scoring unit configured to obtain, for each network card link, structural parameters of the network card link and score the structural parameters to obtain a first score; the structural parameters include the number of connectors, the number of cables, and the length of the cables in the network card link; the network card link is a connection link structure between the network card and a central processing unit;

[0037] a second scoring unit, configured to perform a signal transmission test between the central processing unit and the network card, obtain a test result, and score the test result to obtain a second score;

[0038] The quality evaluation unit is configured to calculate according to a preset method based on the first score and the second score to obtain a quality score of each network card link.

[0039] In a third aspect, the present application provides a network card link quality detection device, comprising:

[0040] Memory for storing computer programs;

[0041] The processor is configured to implement the steps of the above-mentioned method for detecting the quality of the network card link when executing the computer program.

[0042] In a fourth aspect, the present application provides a computer-readable storage medium having a computer program stored thereon, and when the computer program is executed by a processor, the steps of the network card link quality detection method as described above are implemented.

[0043] The present application provides a quality detection method, system, device and medium for a network card link, which relates to the field of signal transmission and solves the problem of no unified standard for evaluating the quality of a network card link. The network card link structure is scored according to the number of connectors, the number and size of cables between the central processing unit and the network card to obtain a first score; a signal transmission test is performed on the central processing unit and the network card to obtain a second score; and a quality score for each network card link is obtained based on the first score and the second score. The number of connectors, the number and length of cables are related to signal attenuation and interference. Incorporating the number of connectors, the number and length of cables into the scoring system can quantitatively evaluate the structural advantages and disadvantages of different links; combined with the signal transmission test results, the actual performance of the network card link can be more comprehensively reflected, and the signal quality of multiple network card links can be uniformly quantitatively evaluated. Then, the quality of multiple links can be ranked according to the quality score, thereby selecting the link with the best quality. BRIEF DESCRIPTION OF THE DRAWINGS

[0044] In order to more clearly illustrate the embodiments of the present invention, the following is a brief introduction to the drawings required for use in the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without any creative work.

[0045] Figure 1 This is a flow chart of a quality detection method for a network card link provided by the present invention;

[0046] Figure 2 A schematic diagram of impedance change of a connector provided by the present invention;

[0047] Figure 3 A schematic diagram of impedance change of a cable provided by the present invention;

[0048] Figure 4 A schematic diagram of the first network card link provided by the present invention;

[0049] Figure 5 This is a schematic diagram of the second network card link provided by the present invention;

[0050] Figure 6 This is a schematic diagram of the third network card link provided by the present invention;

[0051] Figure 7 A flowchart of a network card link screening provided by the present invention;

[0052] Figure 8 This is a schematic diagram of a network card link quality detection system provided by the present invention. DETAILED DESCRIPTION

[0053] The core of the present invention is to provide a quality detection method, system, device and medium for network card links, which can uniformly quantify the signal quality of multiple network card links, and then rank the quality of multiple links according to the quality score, so as to select the link with the best quality.

[0054] To make the objectives, technical solutions, and advantages of the embodiments of the present invention more clear, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the accompanying drawings 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 making creative efforts shall fall within the scope of protection of the present invention.

[0055] First, as Figure 1As shown, the present application provides a method for detecting the quality of a network card link, comprising:

[0056] S11: For each network card link, obtain structural parameters of the network card link and score the structural parameters to obtain a first score; the structural parameters include the number of connectors, the number of cables, and the length of the cables in the network card link; the network card link is the connection link structure between the network card and the central processing unit;

[0057] In this step, we first obtain key structural parameters for each NIC link structure. The core of this step is to quantitatively evaluate the physical configuration of the NIC link. The NIC link refers to the connection architecture between the NIC and the central processing unit (CPU), typically involving connectors, cables, and their corresponding configurations.

[0058] like Figure 2 As can be seen, the impedance at the connector fluctuates significantly. The connector count refers to the total number of connectors in a network card link, and the number of connectors directly affects the stability and reliability of signal transmission. Excessive connectors can introduce additional signal loss, interference, or poor physical contact. Therefore, the number of connectors significantly affects the quality of the network card link.

[0059] like Figure 3 It shows the impedance fluctuation of the cable. Impedance discontinuity will cause reflection, and reflection will cause signal fluctuation. Vr represents the reflected voltage, Vi represents the incident voltage, Z1 and Z2 represent the instantaneous impedance of the two transmission lines respectively. Represents the reflection coefficient. According to the formula Vr / Vi=(Z2-Z1) / (Z2+Z1)= It can be inferred that the more drastic the impedance change, the worse the signal quality, and thus it can be inferred that the topology containing connectors should be worse than the topology without connectors, and the topology containing cables should be worse than the topology without cables.

[0060] The number of cables refers to the number of cables used to connect various devices (such as a network card and CPU). Excessive cables can complicate the layout and increase the difficulty of line management. Excessive cables can also cause signal interference or crosstalk, affecting signal transmission quality. Cable length determines the distance a signal can travel. Long cables can cause signal attenuation or delay, especially at high-speed data transmission. Excessive cable length can negatively impact signal quality. Therefore, cable length must be appropriately controlled to ensure that connectivity requirements are met while avoiding transmission delays and signal loss caused by excessive cables.

[0061] These structural parameters reflect the design quality of the network card link and its performance potential in real-world environments. After obtaining these parameters, each parameter is scored. Scoring criteria may be set based on actual needs. For example, a reasonable range may be set based on the number of connectors; exceeding this range may result in a lower score. There are also corresponding scoring criteria for the number and length of cables, with exceeding a certain threshold considered a factor affecting link quality.

[0062] S12: performing a signal transmission test between the central processing unit and the network card, obtaining a test result, and scoring the test result to obtain a second score;

[0063] The core task of this step is to conduct actual testing of signal transmission between the network card and the central processing unit (CPU) to evaluate the performance of the network link in a real-world operating environment. Specifically, the quality and efficiency of data transmission between the network card and the CPU must be evaluated through actual signal transmission testing. This process involves sending a signal from the CPU to the network card and monitoring key parameters such as signal integrity, transmission speed, latency, and packet loss rate in real time during transmission to ensure that the signal can operate stably within the predetermined transmission protocol and rate. This testing can identify potential issues such as signal attenuation, interference, transmission delay, or packet loss, which may negatively impact the overall performance of the system.

[0064] Specific testing procedures may include, but are not limited to: transmission speed testing, which assesses whether the data transmission rate from the CPU to the network card meets design requirements; latency testing, which detects delays in data transmission, particularly during high-speed data exchange, where excessive latency can impact system response time and real-time performance; packet loss rate testing, which assesses whether data packets are lost during transmission. Severe packet loss can lead to data integrity issues, impacting system reliability and efficiency; and signal integrity testing, which checks for signal distortion or errors during transmission to ensure data reaches the target device without error. During testing, the transmission process is typically monitored using standardized testing tools and equipment, and compared against predetermined standards and requirements to produce a specific test result.

[0065] The test results are ultimately converted into a second score, which reflects the quality of actual signal transmission. If the test results indicate that signal transmission meets the preset standards, the score is high; otherwise, the score is low. This second score accurately reflects the performance of the NIC link in real-world applications. As a key component of the NIC link quality score, it, along with the structural parameter score (the first score), constitutes a comprehensive assessment of NIC link quality.

[0066] S13: Calculate the quality score of each network card link according to the first score and the second score in a preset manner.

[0067] In this step, the quality of the network card link is scored by comprehensively evaluating the structural parameter score (the first score) and the signal transmission test score (the second score). The core principle of this process is to combine the quality of the structural design with the actual operating performance to form a comprehensive scoring system.

[0068] First, the first score represents the quality of the physical design of the NIC link. This score measures the rationality and optimization of the link design based on parameters such as the number of connectors, the number of cables, and the length of the cables. A good design can reduce signal transmission loss and interference, ensuring network stability and efficiency. The second score is obtained through an evaluation of actual signal transmission tests. It reflects the signal transmission quality between the NIC and the CPU in actual operation, including factors such as transmission speed, latency, and packet loss rate. Therefore, the second score mainly measures the performance of the link in actual working conditions.

[0069] In this step, the two scores need to be combined according to a preset method (which may be weighted, summed, or other mathematical models). For example, the structural parameter score (first score) may account for a certain proportion of the total score, while the signal transmission test score (second score) may account for another proportion. Combining the results of these two can produce a more comprehensive NIC link quality score. This quality score provides a quantitative basis for evaluating the overall quality of the NIC link. By comprehensively considering design optimization (structural score) and actual performance (test score), this quality score can accurately reflect the reliability, stability, and performance level of the NIC link in real-world environments, helping engineers or users determine whether the NIC link meets design requirements and where there may be room for improvement in actual applications.

[0070] As an optional embodiment: Based on the quality scores of all network card links, the quality scores of all network card links can be compared to select the link with the highest score as the optimal network card link. The specific method is to sort the comprehensive quality score of each network card link (derived from the structural parameter score and the signal transmission test score), and the network card link with the highest score is the optimal link. Generally, the scoring will take into account the comprehensive performance of each link in multiple dimensions such as design rationality and transmission performance. Therefore, the link with the highest score represents the design solution with the best balance and the best performance in these dimensions. In addition, according to specific needs or application scenarios, scoring thresholds or weights can be set to further refine the selection criteria to ensure that the optimal network card link can meet the performance requirements and stability requirements of the system to the greatest extent possible.

[0071] As an optional embodiment, before scoring the structural parameters and obtaining the first score, the method further includes:

[0072] According to the structural parameters of the network card link and the first preset level classification method, a structural level corresponding to the structural parameters of each network card link is obtained;

[0073] Scoring of test results, before obtaining a second score, also includes:

[0074] According to the test results of the network card link and the second preset level classification method, a test level corresponding to the test results of each network card link is obtained; the test level includes several types of test values: not reaching the standard test value, reaching the standard test value, and reaching the standard test value by a preset multiple, where the preset multiple is a value greater than 1 and less than 1.5;

[0075] Structural parameters are scored to obtain a first score, including:

[0076] Determining a first score corresponding to the structural level according to the structural parameter of each network card link;

[0077] The test results are scored to obtain a second score, including:

[0078] According to the test level corresponding to the test result of each network card link, a second score corresponding to the test level is determined.

[0079] In this embodiment, the scoring process for the network card link quality is further refined, and the scoring is made more systematic and objective through the division into structural levels and test levels.

[0080] Regarding the structural parameters of the NIC link, a first preset grading method is used to assign a structural grade to each NIC link based on its structural characteristics (such as the number of connectors, number of cables, and cable length). This structural grade reflects the rationality and degree of optimization of the NIC link design and is typically set based on actual application requirements. For example, the acceptable range for the number of connectors, number of cables, and lengths is considered. If certain parameters fall outside the preset range, they are assigned a lower grade; otherwise, they may receive a higher grade. This process quantifies and grades the structural characteristics of each NIC link, allowing for a scoring based on the quality of the structural design.

[0081] Then, before scoring the test results, the test results of each network card link are assigned to a test level using a second preset level division method based on the test results of the network card link (such as signal transmission speed, delay, packet loss rate, etc.). Test levels usually include multiple levels, such as: "failed to reach the standard test value", "reached the standard test value" and "reached the preset multiple of the standard test value". Among them, the preset multiple is a value greater than 1 and less than 1.5, indicating that the test result can reach a certain multiple of the standard test value. For example, if the standard test value is a certain transmission rate, reaching 1.2 times or 1.3 times the standard test value can be considered a better level. In this way, the network card link can be assigned to different test levels according to the quality of the test results, so as to accurately evaluate the actual performance.

[0082] Then, when scoring the structural parameters, the corresponding first score is determined based on the structural grade of each NIC link. The structural grade directly determines the score. A higher structural grade indicates a more rational NIC link design, resulting in a higher first score. Similarly, when scoring the test results, the second score is determined based on the test grade of each NIC link. A higher test grade indicates better actual signal transmission performance, resulting in a higher second score.

[0083] This tiered scoring approach quantifies both structural design and actual performance testing and links them to corresponding grades and scores, ensuring a more objective and comprehensive NIC link quality score and helping to optimize NIC link selection and application. Ultimately, the combined assessment of the first and second scores provides an accurate and actionable basis for each NIC link's quality score.

[0084] As an optional embodiment, before determining the first score corresponding to the structural level according to the structural parameter of each network card link, the method further includes:

[0085] Eliminating network card links of a preset structural level from all network card links; the number of connectors in the network card links of the preset structural level is greater than a first preset number and / or the number of cables is greater than a second preset number and / or the length of the cables is greater than a preset length;

[0086] Before determining the second score corresponding to the test level according to the test result of each network card link, the method further includes:

[0087] Eliminate network card links with a preset test level from all network card links; the preset test level is a level corresponding to a test value that does not reach the standard.

[0088] In this embodiment, the quality scoring process of the network card link introduces a screening mechanism for structural parameters and test results, making the scoring process more accurate and having the ability to exclude links that do not meet the standards.

[0089] Specifically, before determining the structural level of each network card link and calculating the first score, those link structures that do not meet the requirements are eliminated from all network card links according to preset standards. Specifically, the preset structural level refers to the excess range of certain link structure parameters, such as the number of connectors, the number of cables, and the length of the cables. If these parameters of a network card link exceed the preset threshold (such as the number of connectors is greater than a first preset number, the number of cables is greater than a second preset number, or the cable length is greater than the preset length), the network card link will be eliminated and cannot participate in the subsequent scoring calculation. The purpose of this is to filter out those network card links that are significantly unreasonable in structural design, to prevent them from affecting the quality assessment of the entire system, and to ensure that the scoring results can better reflect the high-quality network card links.

[0090] Next, before determining the second score based on the test results of each network card link, a similar screening mechanism is used to eliminate from all network card links those links whose test results do not meet the standards. The preset test levels are mainly aimed at the performance of the network card link during signal transmission, such as transmission speed, latency, packet loss rate, etc. If the test result of a network card link fails to meet the standard test value (for example, the transmission speed is lower than the predetermined standard), the network card link will be classified as "unqualified" and will be eliminated from the range of further scoring. This means that only those network card links whose signal transmission performance meets the standard or exceeds a certain preset multiple (such as the test result is greater than 1.2 times or 1.3 times the standard value) will be considered for the final quality score calculation.

[0091] This two-stage screening mechanism effectively eliminates links that clearly fail to meet design or performance standards from a large number of network interface cards (NICs), preventing these substandard links from influencing the final quality assessment results. This not only ensures fairness and accuracy in scoring, but also ensures that the scoring system focuses on and highlights NIC links that excel in both structural design and actual performance, thereby improving overall system stability and performance.

[0092] As an optional embodiment, a signal transmission test is performed between the central processing unit and the network card to obtain a test result, including:

[0093] Performing a power-off and power-on test on the network card in sequence to obtain a first test result;

[0094] Perform a second test on the network card by shutting down, powering off, powering on, and booting up in sequence to obtain a second test result;

[0095] The test results include at least a first test result and a second test result.

[0096] In this embodiment, the signal transmission test is designed to comprehensively evaluate the signal transmission performance and stability of the network card under different operating conditions through a multi-stage testing process. Specifically, the testing process may include, but is not limited to, two main stages: a first test and a second test. Each test stage verifies the network card's signal transmission capabilities and its response to power supply changes by simulating different power management and startup processes.

[0097] First, the first testing phase involves a simple power cycle (also known as a DC reboot) of the network card. This test verifies the stability of signal transmission during normal power-on and power-off cycles. After powering off and back on, the card needs to be reinitialized and restored to normal operation. By measuring parameters such as signal transmission latency, packet loss rate, and data rate during this process, we can determine whether the card can maintain stable signal transmission quality despite power state changes.

[0098] Next, the second testing phase involves four steps: "power off," "power off," "power on," and "power on" (also known as the AC reboot test). This series of operations simulates more severe power state changes, testing the network card's performance during a complete power outage and power restoration. The "power off" and "power on" operations, in particular, simulate scenarios of unstable power supply or system power restoration after a power outage, which is critical to the resilience of the network card and system, signal stability, and network connection reliability. By testing this process, we can more comprehensively evaluate whether the network card can successfully recover after a complete power outage and ensure that signal transmission quality is not affected after power is restored.

[0099] Ultimately, the test results include at least the first and second test results. These two combined results provide a comprehensive view of the NIC's performance under both normal power-on and power-off operations and extreme power-state fluctuations. This multi-stage test effectively identifies potential issues such as power instability or signal recovery, providing a more accurate basis for assessing NIC link quality.

[0100] As an optional embodiment, the method further includes:

[0101] assigning a first weight to the first test result and a second weight to the second test result;

[0102] A restart test score of the network card link is obtained according to the first weight and the first test result, the second weight and the second test result.

[0103] In this embodiment, the calculation process of the restart test score is further refined by assigning different weights to different test results to ensure that the score can reflect the importance of each test stage and its actual impact on the network card link quality.

[0104] Specifically, the weights of the first and second test results in the scoring can be 60% and 40% respectively. This means that when evaluating the restart performance of the network card link, the first test phase (power off, power on) is relatively considered more critical because it reflects the stability and reliability of signal transmission during the regular startup process of the network card. The second test phase (power off, power off, power on, power on) mainly tests the network card's recovery ability after unstable power supply or complete power failure.

[0105] The results of the first test directly affect the reliability of the network card link in the actual environment. If the network card performs well in this test phase, that is, it can quickly recover after being shut down and transmit signals stably, then it can get a higher score, indicating that it has higher stability. The results of the second test can more comprehensively simulate the power fluctuation or power outage recovery scenario. Although the weight of this test phase is lower than the first test, it is still very important because it reflects the performance of the network card in the face of power failure or system restart, and determines whether the network card can recover normally and maintain signal quality in an unstable power environment.

[0106] Ultimately, the reboot test score is calculated by combining the first and second test results with their respective weights. Specifically, the formula is: reboot test score = (first test result × 60%) + (second test result × 40%). This scoring method weights the impact of the two test results, ensuring that the test results from both operating environments have a combined impact on the final score. This weighted approach more accurately reflects the performance of the NIC link in actual use, especially its stability and reliability in environments with frequent power cycles or power fluctuations.

[0107] As an optional embodiment, a signal transmission test is performed between the central processing unit and the network card to obtain a test result, including:

[0108] Sending original signals and original data packets from the central processing unit, and acquiring test signals and test data packets received by the network card; the original signals correspond to the test signals, and the original data packets correspond to the test data packets;

[0109] Obtaining a test result based on the signal parameters of the test signal and the data packet parameters of the test data packet;

[0110] Signal parameters are parameters used to characterize signal quality, and data packet parameters are parameters used to characterize data transmission performance.

[0111] In this embodiment, the principle of signal transmission testing is to comprehensively evaluate signal quality and data transmission performance. Two testing methods, margin testing and iperf testing, can be used, but are not limited to. Signal parameters and data packet parameters are combined to obtain accurate test results.

[0112] Specifically, in the margin test, the central processing unit (CPU) sends the original signal to the network card. This signal is usually used to test the signal quality of the network link. By comparing the original signal sent with the signal received by the network card, the attenuation, distortion or interference of the signal during transmission can be evaluated. Signal parameters are key indicators for describing signal quality, which usually include signal strength, noise level, signal-to-noise ratio (SNR), bit error rate (BER), etc. These parameters reflect the reliability and clarity of the signal during transmission, thus providing effective information about the quality of the signal received by the network card.

[0113] Secondly, the iperf test is a common method for evaluating network performance through data packet transmission. In this test, the central processing unit sends original data packets, and the network card receives the data packets and returns them to the central processing unit. Data packet parameters are used to represent data transmission performance, usually including throughput (bandwidth), delay, packet loss rate, etc. Throughput measures the amount of data transmitted per unit of time, delay reflects the time required for data packets to be transmitted and received, and packet loss rate shows the proportion of data packets lost during transmission. These parameters help to judge the performance of the network card in actual data transmission, especially under high load and network congestion conditions.

[0114] By combining these two testing methods, the margin test is mainly used to evaluate signal quality, while the iperf test is used to evaluate data transmission performance. Finally, based on the signal parameters of the test signal and the data packet parameters of the test data packet, the test results can be comprehensively obtained. These parameters together constitute a comprehensive evaluation of the performance of the network card, ensuring that the network card can stably and efficiently transmit data in different signal environments.

[0115] In summary, through this dual testing method, the system can not only detect the signal quality of the network card, but also comprehensively evaluate its data transmission capability, thus providing accurate and multi-dimensional test basis for the quality evaluation of the network card link.

[0116] As an optional embodiment, the quality score of each network card link is calculated according to the first score and the second score in a preset manner, including:

[0117] determining a first influence factor corresponding to the first score and a second influence factor corresponding to the second score;

[0118] A quality score of each network card link is determined according to the first impact factor, the first score, the second impact factor, and the second score.

[0119] In this embodiment, the quality score of the network card link not only depends on the first score and the second score, but also introduces the concept of influencing factors to further refine the scoring process and ensure that the scoring result is more reasonable and operational.

[0120] Specifically, the scoring process first uses influencing factors to assign different weights or degrees of influence to different test dimensions (such as structural parameters and signal transmission test results). This allows the focus of the scoring to be adjusted according to different actual needs, ultimately resulting in a comprehensive quality score.

[0121] The first and second scores reflect the performance of the NIC link in terms of structural parameter scoring and signal transmission test scoring, respectively. However, simply the first and second scores may not fully reflect the actual quality of the NIC link, as different NIC links may have different levels of influence in these two dimensions. In this case, an impact factor is assigned to each score to adjust its importance.

[0122] The First Impact Factor corresponds to the First Score and is typically used to reflect the degree to which the structural parameters of the NIC link (such as the number of connectors, number of cables, and cable length) influence the overall quality of the NIC. Structural design optimizations can significantly improve the stability and reliability of the NIC, so the First Impact Factor is typically set based on the importance of these design factors. If a NIC link has superior structural parameters, its First Impact Factor may be higher, thus having a greater positive impact on the final quality score.

[0123] Similarly, the second impact factor corresponds to the second score and is primarily related to the NIC's signal transmission test results (such as margin and iperf tests). The magnitude of this impact factor reflects the impact of these test results on the NIC's link quality. Signal transmission quality directly impacts the NIC's performance in real-world network environments. Therefore, the second impact factor may be more important in certain scenarios, especially when signal transmission quality is critical to system stability and performance.

[0124] Finally, the quality score of the network card link is obtained by comprehensively calculating the first impact factor, the first score, the second impact factor, and the second score. The calculation formula can be expressed as: quality score = (first impact factor × first score) + (second impact factor × second score). This weighted calculation method ensures that the impact of different dimensions can be adjusted according to their actual importance, thereby obtaining a more accurate quality score. By setting different impact factors, the scoring criteria can be flexibly adjusted according to different application scenarios or needs. For example, in scenarios with higher requirements for structural design, the first impact factor can be higher, while in scenarios with higher requirements for signal stability, the second impact factor can be relatively large.

[0125] Overall, this impact factor-based scoring method makes the quality score of the network card link not just a simple sum of two scores, but an accurate score that comprehensively considers multiple factors, which helps to select the best network card link solution for different application scenarios.

[0126] In one embodiment of the present invention, a method for evaluating the signal quality of a network interface card link employs a multi-dimensional comprehensive evaluation strategy. This strategy uses multiple influencing factors to score the signal quality of the network interface card link and uses a weighted scoring method to determine the final quality rating. These influencing factors include link type (A, B, C), loss level, margin test results, iperf test results, and restart test results. Each factor has a different impact on the final score.

[0127] First, network card links (denoted as T1) are divided into three types: A link (such as Figure 4 As shown), B link (as shown Figure 5 as shown) and C-link (as Figure 6 Each link type has a different impact on signal quality, so different weights are assigned to it, recorded as A, B, and C. Based on experiments or system experience, these link types may have different signal quality impact factors, for example, link A is assigned a value of 60, link B is assigned a value of 25, and link C is assigned a value of 15. These weights are used to quantify the relative contributions of different link types to signal quality.

[0128] Next, loss (denoted by the codename T2), a key factor affecting signal quality, is divided into three levels: high (H), medium (M), and low (L), each assigned a different weight. The higher the loss level, the more severe the signal attenuation and the greater the impact on signal quality. Loss levels are set based on system experience. For example, the M level is assigned a value of 40, the L level is assigned a value of 60, and the H level has a greater impact, potentially resulting in poor network card signal quality, thus having a significant negative impact on the final score.

[0129] Next, the margin test (T3) and the iperf test (T4) are performed to evaluate the network card's signal quality and data transmission performance. Margin test results are categorized into three levels (D, E, and F), corresponding to "does not meet spec (standard test value)", "110%*spec", and "120%*spec", respectively. Similarly, iperf test results are categorized into three levels (G, H, and I), corresponding to "does not meet spec", "110%*spec", and "120%*spec", respectively. These test results further determine the performance of the network card, providing additional dimensions for signal quality assessment.

[0130] In addition, the reboot test (codenamed T5) includes DC reboots and AC reboots, with weights of 60% and 40%, respectively. This ensures a comprehensive assessment of the network card's stability and signal quality under different reboot conditions. Reboot test results are also categorized by deviation from the standard test value, with three levels: X, Y, and Z, corresponding to "does not meet spec," "110%*spec," and "120%*spec," respectively.

[0131] All these factors are combined to produce 243 possible combinations through complex permutations and combinations. On this basis, some combinations with poor performance (such as those containing H and D level links) are eliminated, and those network card links with extremely poor signal quality (such as Figure 7 The remaining 216 sets of data enter the nested subroutine module (the calculation method of 216 is: [C(3,3)×C(3,3)×C(3,3)×C(3,3)×C(3,3)]- [C(3,3)×C(1,1)×C(1,1)×C(3,3)×C(3,3)]=3 5 -3 3 =216), and continue with more detailed classification to ensure that the signal quality of each network card link can be accurately classified.

[0132] Nested subroutine modules further weight the contribution of defined influencing factors (such as P1, P2, P3, P4, and P5) to the network interface card (NIC) quality. For example, P1 represents the impact of link type, P2 represents the impact of loss, P3 and P4 represent the impact of margin testing and iperf testing, respectively, and P5 represents the impact of reboot testing. The value of each factor is determined based on system experience or statistical analysis. In practice, the weights of P1, P2, P3, P4, and P5 (these influencing factors) can be set based on system experience or dynamically adjusted based on extensive data and statistical probabilities (for correction methods, refer to the Naive Bayes algorithm) to adapt to different application scenarios and data environments.

[0133] Finally, the system uses a weighted formula to combine the weights of each factor with the scoring results to calculate the final quality score. This weighted calculation method ensures that the impact of each dimension can be flexibly adjusted according to actual needs, making the quality assessment of the network interface card link not only accurate but also dynamically optimized for different scenarios. The final quality score Grade = P1 × T1 × (3 / 3) + P2 × T2 × (2 / 3) + P3 × T3 × (2 / 3) + P4 × T4 × (3 / 3) + P5 × T5 × (3 / 3). T1-T5 are the corresponding values ​​assigned above. P1-P5 are the corresponding impact factors. 3 / 3 indicates that no grades that do not meet the requirements are eliminated from the corresponding item, and 2 / 3 indicates that one grade that does not meet the requirements is eliminated from the corresponding item.

[0134] In summary, this invention uses a multi-dimensional, comprehensive analysis of the various factors involved in network interface card link quality assessment, employing a weighted scoring approach to accurately assess each network interface card link using a scientific method. This approach can accurately determine the signal quality of different link structures in complex network environments, providing an effective basis for subsequent link optimization and performance improvement.

[0135] In a specific embodiment, P1, P2, P3, P4, and P5 are assigned values ​​of 1, 3, 3, 2, and 1, respectively.

[0136] For example, let's evaluate the signal quality of three OCP network adapter links. The first link is assumed to be X1 (A, H, F, I, Z), the second link is assumed to be X2 (C, L, F, I, Z), and the third link is assumed to be X3 (A, M, F, H, X). After screening, the X1 link has the worst signal quality.

[0137] Because H exists in X1, it is directly excluded from the poor quality category and does not participate in the calculation of the quality score.

[0138] According to the steps described above, enter each impact factor and calculate the score.

[0139] The quality score of X2 is: Grade = 1 × 15 × (3 / 3) + 3 × 60 × (2 / 3) + 3 × 55 × (2 / 3) + 2 × 50 × (3 / 3) + 1 × 50 × (3 / 3) = 395;

[0140] The quality score of X3 is: Grade = 1 × 60 × (3 / 3) + 3 × 40 × (2 / 3) + 3 × 55 × (2 / 3) + 2 × 40 × (3 / 3) + 1 × 10 × (3 / 3) = 340;

[0141] Comparing the signal quality of the three links, we conclude that the signal quality of the X2 link is the best, the signal quality of the X3 link is second, and the signal quality of the X1 link is the worst.

[0142] In summary, the present invention determines the OCP network card link with the best signal quality based on the results of the margin test, iperf test, DC restart test, and AC restart test combined with an algorithm formula.

[0143] Through the above method, the signal quality of different OCP links can be determined, and the link with the best signal quality can be determined.

[0144] Second, as Figure 8 As shown, the present application provides a network card link quality detection system, including:

[0145] A first scoring unit 21 is configured to obtain, for each network interface card link, structural parameters of the network interface card link and score the structural parameters to obtain a first score; the structural parameters include the number of connectors, the number of cables, and the length of the cables in the network interface card link; the network interface card link is the connection link structure between the network interface card and the central processing unit;

[0146] The second scoring unit 22 is configured to perform a signal transmission test between the central processing unit and the network card, obtain a test result, and score the test result to obtain a second score;

[0147] The quality evaluation unit 23 is configured to perform calculation according to a preset method based on the first score and the second score to obtain a quality score of each network card link.

[0148] For further introduction to the quality detection system for the network card link, please refer to the above embodiments, and this application will not go into details here.

[0149] In a third aspect, the present application provides a network card link quality detection device, comprising:

[0150] Memory for storing computer programs;

[0151] The processor is configured to implement the steps of the above-mentioned network card link quality detection method when executing the computer program. For other introductions to the network card link quality detection device, please refer to the above-mentioned embodiments, which will not be described in detail in this application.

[0152] Fourthly, the present application provides a computer-readable storage medium having a computer program stored thereon. When executed by a processor, the computer program implements the steps of the above-described network card link quality detection method. For further information on the computer-readable storage medium, please refer to the above-described embodiments, and this application will not elaborate further here.

[0153] In a fifth aspect, the present application provides a computer program product, including a computer program / instruction, which, when executed by a processor, implements the steps of the aforementioned network card link quality detection method. For further description of the computer program product, please refer to the aforementioned embodiments, and this application will not elaborate further here.

[0154] It should also be noted that, in this specification, relational terms such as first and second, etc., are used only to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any actual relationship or order between these entities or operations. Moreover, the terms "comprises," "comprising," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus comprising a series of elements includes not only those elements, but also other elements not explicitly listed, or elements inherent to such process, method, article, or apparatus. In the absence of further limitations, an element defined by the phrase "comprising a ..." does not exclude the presence of additional identical elements in the process, method, article, or apparatus comprising the element.

[0155] The above description of the disclosed embodiments is intended to enable one skilled in the art to implement or use the present invention. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the present invention. Therefore, the present invention is not limited to the embodiments shown herein but is intended to conform to the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A method for detecting the quality of a network card link, characterized in that: include: For each of the network card links, obtaining structural parameters of the network card link, and scoring the structural parameters to obtain a first score; The structural parameters include the number of connectors in the network card link, the number of cables, and the length of the cables; The network card link is a connection link structure between the network card and the central processing unit; Performing a signal transmission test between the central processing unit and the network card to obtain a test result, and scoring the test result to obtain a second score; A quality score of each network card link is obtained by performing calculation according to the first score and the second score in a preset manner.

2. The quality detection method of the network card link according to claim 1, characterized in that: Scoring the structural parameters before obtaining the first score further includes: According to the structural parameters of the network card link and a first preset level classification method, obtaining a structural level corresponding to the structural parameters of each network card link; Scoring the test result to obtain the second score further includes: Obtaining a test level corresponding to each test result of the network card link according to the test result of the network card link and a second preset level classification method; the test level includes one or more of a test value not reaching a standard test value, a test value reaching a standard test value, and a standard test value reaching a preset multiple, where the preset multiple is a value greater than 1 and less than 1.5; Scoring the structural parameters to obtain a first score includes: determining, according to the structural level corresponding to the structural parameter of each of the network card links, a first score corresponding to the structural level; Scoring the test result to obtain a second score includes: According to the test level corresponding to the test result of each network card link, a second score corresponding to the test level is determined.

3. The quality detection method of the network card link according to claim 2, characterized in that: Before determining the first score corresponding to the structural level according to the structural parameter of each network card link, the method further includes: Eliminating network card links of a preset structural level from all the network card links; the number of connectors in the network card links of the preset structural level is greater than a first preset number and / or the number of cables is greater than a second preset number and / or the length of the cables is greater than a preset length; Before determining a second score corresponding to the test level corresponding to the test result of each network card link, the method further includes: The network card links with a preset test level are eliminated from all the network card links; the preset test level is a level that does not reach the standard test value.

4. The quality detection method of the network card link according to claim 1, characterized in that: Performing a signal transmission test between the central processing unit and the network card to obtain a test result includes: Performing power-off and power-on tests on the network card in sequence to obtain a first test result; Performing a second test of shutting down, powering off, powering on, and booting up the network card in sequence to obtain a second test result; The test results include at least the first test result and the second test result.

5. The quality detection method of the network card link according to claim 4, characterized in that: Also includes: assigning a first weight to the first test result, and assigning a second weight to the second test result; A restart test score of the network card link is obtained according to the first weight and the first test result, the second weight and the second test result.

6. The quality detection method of the network card link according to claim 1, characterized in that: Performing a signal transmission test between the central processing unit and the network card to obtain a test result includes: Sending an original signal and an original data packet from the central processor, and acquiring a test signal and a test data packet received by the network card; the original signal corresponds to the test signal, and the original data packet corresponds to the test data packet; Obtaining the test result according to the signal parameters of the test signal and the data packet parameters of the test data packet; The signal parameters are parameters used to characterize signal quality, and the data packet parameters are parameters used to characterize data transmission performance.

7. The method for detecting the quality of a network card link according to any one of claims 1 to 6, wherein: Calculating according to a preset method based on the first score and the second score to obtain a quality score of each network card link includes: Determining a first impact factor corresponding to the first score and a second impact factor corresponding to the second score; A quality score of each of the network card links is determined according to the first impact factor, the first score, the second impact factor, and the second score.

8. A quality detection system for a network card link, characterized in that: include: a first scoring unit, configured to obtain, for each of the network card links, structural parameters of the network card link, and score the structural parameters to obtain a first score; The structural parameters include the number of connectors in the network card link, the number of cables, and the length of the cables; The network card link is a connection link structure between the network card and the central processing unit; a second scoring unit, configured to perform a signal transmission test between the central processing unit and the network card, obtain a test result, and score the test result to obtain a second score; The quality evaluation unit is configured to calculate according to a preset method based on the first score and the second score to obtain a quality score of each network card link.

9. A quality detection device for a network card link, characterized in that: include: memory for storing computer programs; The processor is configured to implement the steps of the method for detecting the quality of a network card link according to any one of claims 1 to 7 when executing a computer program.

10. A computer-readable storage medium, characterized in that The computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the steps of the method for detecting the quality of a network card link according to any one of claims 1 to 7 are implemented.

Images