Network equipment double-flow throughput testing method

Abstract

The invention discloses a network equipment double-flow throughput testing method, and relates to the technical field of network equipment testing, and the method comprises the steps: carrying out the multi-group double-flow testing of target network equipment under the condition that the CPU utilization rate of the target network equipment is within a first set range, and obtaining the throughput of the target network equipment through the multi-group double-flow testing; determining a linear fitting relational expression between the tunnel traffic and the Internet traffic; dividing the tunnel traffic into a plurality of traffic levels, performing a double-traffic test on the target network equipment for the traffic level of each tunnel traffic, and collecting effective double-traffic data pairs of which the change rate of the CPU utilization rate is within a second set range; for the traffic level of each tunnel traffic, substituting the normalized tunnel traffic and Internet traffic into a linear fitting relational expression to obtain a fitting error, if the fitting error does not exceed a set value, determining that the test is valid, otherwise, re-fitting the linear fitting relational expression; and outputting the related data obtained by the test corresponding to the traffic level of each tunnel traffic. According to the invention, cross-device multiplexing can be realized.

Description

Technical Field

[0001] This application relates to the field of network device testing technology, and in particular to a method for testing the throughput of network devices with dual traffic. Background Technology

[0002] Many enterprise network devices (such as Secure Access Service Edge (SASE) and multi-line converged devices) now need to handle two types of traffic simultaneously: internet traffic and tunnel traffic. To determine the performance of a device, it's necessary to test how much internet traffic decreases when both types of traffic are running together—this is called dual-traffic throughput testing, and the results can help enterprises select equipment and plan network resources.

[0003] However, current testing methods have a major problem: when testing different devices, the test procedures and parameters must be redesigned each time, which is inefficient; moreover, after the device is upgraded, the previous test results become inaccurate, with errors exceeding 20%. The root cause is that traffic is not correlated with the device's core hardware (such as the CPU), so the test results cannot reflect the true resource contention situation.

[0004] Existing technology 1: Fixed slope formula testing method. This method is for a single device, using a fixed data packet type for testing (e.g., mixed packets), and then manually setting 5-8 tunnel traffic rates (e.g., 0, 50, 100...300Mbps). For each tunnel traffic rate, internet traffic is measured 3 times, and the average value is taken. Finally, an Excel line is plotted to fit a formula: "Internet traffic = -k × tunnel traffic + C" (where k is the fixed slope and C is the intercept). If a different device is used, 5-8 sets of data must be measured again, and a new formula must be fitted.

[0005] The disadvantages of the existing technology are as follows.

[0006] 1) Testing different devices is too troublesome: 5-8 sets of data need to be measured again for each device. The testing time for 10 devices is 3 times longer than that for 3 devices, which is too inefficient.

[0007] 2) Test results become unusable after device upgrade: For example, if the device could originally run a maximum of 295Mbps of internet traffic, but after the upgrade it can run 350Mbps, the error of the results calculated by the previous formula will increase from 5% to 23%, making it completely unusable.

[0008] 3) The results are meaningless: Only looking at traffic data, regardless of CPU activity level, means that when the CPU is idle, there is more internet traffic, and when the CPU is busy, there is less internet traffic. The test results cannot explain this phenomenon, nor can they guide the optimization of the equipment.

[0009] The second relevant existing technology: a resource-unconstrained normalization testing method. To allow for comparison of test results from different devices, a method has been devised: first, measure the device's maximum internet throughput (NetMax) and maximum tunnel throughput (TunMax), then divide the real-time traffic by the maximum throughput to obtain "normalized traffic" (e.g., NetNorm = NetReal / NetMax), and then fit the formula "NetNorm = -m × TunNorm + n" (where m and n are parameters). However, the load on hardware resources such as CPU and memory is ignored during the test.

[0010] The disadvantages of the existing technology 2 are as follows.

[0011] 1) Parameter fluctuates: Because the CPU load was not controlled, when the CPU ran from 60% to 90%, the calculated parameter m could differ by 12%, and the final test error exceeded 18%; 2) Retesting is required when changing scenarios: If the data packets used for testing are changed from mixed packets to large packets, the previous parameters m and n are no longer useful and must be recalculated, increasing the cost by half. Summary of the Invention

[0012] The purpose of this application is to provide a method for testing the throughput of dual traffic in network devices, which can achieve cross-device reuse.

[0013] To achieve the above objectives, this application provides the following solution: This application provides a method for testing the throughput of a network device under dual traffic conditions, the method comprising: Under the condition that the CPU utilization of the target network device is within a first set range, the linear fitting relationship between tunnel traffic and Internet traffic is determined by performing multiple sets of dual-traffic tests on the target network device. The tunnel traffic is divided into multiple traffic levels. For each traffic level of the tunnel traffic, a dual traffic test is performed on the target network device. During the dual traffic test, real-time tunnel traffic and Internet traffic are collected. If the rate of change of CPU utilization exceeds the second set range during the collection process, the dual traffic test is performed again on the target network device for that traffic level until the rate of change of CPU utilization during the collection process is within the second set range, and a valid dual traffic data pair for that traffic level is obtained. For each tunnel traffic volume, the corresponding valid dual traffic data is normalized for both tunnel traffic and internet traffic. The normalized tunnel traffic and internet traffic are then substituted into the linear fitting formula to obtain the fitting error. If the fitting error does not exceed a set value, the test is considered valid. Otherwise, the process returns to the step of determining the linear fitting formula for tunnel traffic and internet traffic by performing multiple sets of dual traffic tests on the target network device under the condition that the CPU utilization of the target network device is within a first set range, until the fitting error corresponding to each traffic volume of the tunnel traffic does not exceed the set value. Output the tunnel traffic, internet traffic, normalized tunnel traffic, normalized internet traffic, and fitting error corresponding to the traffic level of each tunnel traffic.

[0014] Optionally, under the condition that the CPU utilization of the target network device is within a first set range, a linear fitting relationship between tunnel traffic and Internet traffic is determined by performing multiple sets of dual-traffic tests on the target network device, specifically including: Three tunnel flow rates are set according to 30%, 60%, and 90% of the maximum tunnel flow rate; Under the condition that the CPU utilization of the target network device is within a first set range, three sets of dual-traffic tests are performed on the target network device based on three levels of tunnel traffic, and the corresponding Internet traffic is recorded during each set of dual-traffic tests; each level of tunnel traffic and its corresponding Internet traffic constitute three sets of dual-traffic test data; Based on three sets of dual-flow test data, a linear fitting formula for tunnel flow and Internet flow was determined.

[0015] Optionally, under the condition that the CPU utilization of the target network device is within a first set range, three sets of dual-traffic tests are performed on the target network device based on three levels of tunnel traffic, and the corresponding Internet traffic during each set of dual-traffic tests is recorded, specifically including: Take 30% of the maximum tunnel flow as the first level of tunnel flow, measure the Internet flow for 5 minutes, record the Internet flow once every 1 minute, and take the average of 5 times to obtain the Internet flow of the first group. Use 60% of the maximum tunnel flow as the second level of tunnel flow, measure the Internet flow for 5 minutes, record the Internet flow once every 1 minute, and take the average of 5 times to obtain the Internet flow of the second group. The third level of tunnel flow was determined by using 90% of the maximum tunnel flow. Internet traffic was measured for 5 minutes, with one record of internet traffic every minute. The average of the five records was taken to obtain the third group of internet traffic.

[0016] Optionally, based on three sets of dual-flow test data, a linear fitting formula for tunnel traffic and Internet traffic is determined, specifically including: Normalize the tunnel traffic and Internet traffic in the three sets of dual-flow test data; Based on the normalized three sets of dual-flow test data, the linear fitting relationship between tunnel flow and Internet flow is determined by the least squares method, or by using Excel's Solver tool to determine the linear fitting relationship between tunnel flow and Internet flow. The linear fitting relationship is expressed as follows: ; in, a and b All are fitting coefficients. For normalized internet traffic, This represents the normalized tunnel flow rate.

[0017] Optionally, the normalized tunnel traffic and Internet traffic are substituted into the linear fitting equation to obtain the fitting error, which specifically includes: Substituting the normalized tunnel traffic and internet traffic into the linear fitting equation, the difference between the calculated result and 1 is the fitting error; the calculated result is... .

[0018] Optionally, before determining the linear fitting relationship between tunnel traffic and Internet traffic by performing multiple sets of dual-traffic tests on the target network device, under the condition that the CPU utilization of the target network device is within a first set range, the dual-traffic throughput test method of the network device further includes: The test measures the target network device's maximum internet traffic when tunnel traffic is zero, and the maximum tunnel traffic when internet traffic is zero. The maximum tunnel traffic is used for tunnel traffic normalization, and the maximum internet traffic is used for internet traffic normalization.

[0019] Optionally, before determining the linear fitting relationship between tunnel traffic and Internet traffic by performing multiple sets of dual-traffic tests on the target network device, under the condition that the CPU utilization of the target network device is within a first set range, the dual-traffic throughput test method of the network device further includes: A network testing tool is used to send preset background data to the target network device, so that the CPU utilization of the target network device is within a first set range.

[0020] Optionally, the tunnel traffic is divided into multiple traffic tiers. For each traffic tier, a dual-traffic test is performed on the target network device. During the dual-traffic test, real-time tunnel traffic and internet traffic are collected. If the rate of change in CPU utilization exceeds a second preset range during the collection process, the dual-traffic test is repeated on the target network device for that traffic tier until the rate of change in CPU utilization during the collection process is within the second preset range, thus obtaining a valid dual-traffic data pair for that traffic tier. Specifically, this includes: The tunnel flow rate is divided into six flow rate levels based on 0%, 20%, 40%, 60%, 80%, and 100% of the maximum tunnel flow rate. For each tunnel traffic level, the target network device is subjected to three dual-traffic tests, and the rate of change in CPU utilization during the three dual-traffic tests does not exceed the second set range. The average tunnel traffic is calculated based on the three tunnel traffic values ​​obtained from three dual-traffic tests for each traffic level, and the average internet traffic is calculated based on the three internet traffic values. The average tunnel traffic value and the average internet traffic value constitute a valid dual-traffic data pair.

[0021] Optionally, the first setting range is greater than 90% and less than 95%, and the second setting range is greater than -5% and less than 5%.

[0022] Optionally, the setting value is 3%.

[0023] According to the specific embodiments provided in this application, the following technical effects are disclosed: This application provides a method for testing the throughput of dual traffic on network devices. By setting a constraint on the CPU utilization of the target network device within a first set range, the method reflects the real resource competition of dual traffic. The method sets the change rate of CPU utilization within a second set range as the data collected as valid data, thus ensuring the validity of the collected traffic data. After normalizing the tunnel traffic and Internet traffic in the corresponding valid dual traffic data, the fitting error is calculated, which can eliminate the differences between network devices and realize cross-device reuse. Attached Figure Description

[0024] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0025] Figure 1 This is a flowchart illustrating a method for testing the throughput of a network device with dual traffic, provided in an embodiment of this application.

[0026] Figure 2 This is a detailed flowchart illustrating a method for testing the throughput of a network device with dual traffic, provided as an embodiment of this application.

[0027] Figure 3 This is a schematic flowchart of a constant factor calibration process provided in an embodiment of this application. Detailed Implementation

[0028] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0029] This application provides a method for testing the throughput of a network device using dual traffic, such as... Figure 1 and Figure 2 As shown, the throughput test method for dual traffic of the network device includes steps 101-104.

[0030] Step 101: Under the condition that the CPU utilization of the target network device is within a first set range, determine the linear fitting relationship between tunnel traffic and Internet traffic by performing multiple sets of dual-traffic tests on the target network device.

[0031] Dual traffic refers to two types of traffic processed simultaneously by a target network device: tunnel traffic and internet traffic. Internet traffic is ordinary network data directly sent and received by the network device, denoted as "NetReal Internet Traffic" (Mbps). Tunnel traffic is data transmitted through an encrypted tunnel, denoted as "TunReal Tunnel Traffic" (Mbps). Tunnel traffic and internet traffic share the network device's core hardware resources, such as the CPU.

[0032] Step 102: Divide the tunnel traffic into multiple traffic levels. For each traffic level of the tunnel traffic, perform a dual traffic test on the target network device. During the dual traffic test, collect real-time tunnel traffic and Internet traffic. If the rate of change of CPU utilization exceeds the second set range during the collection process, then perform the dual traffic test again on the target network device for that traffic level until the rate of change of CPU utilization during the collection process is within the second set range, thus obtaining a valid dual traffic data pair for that traffic level.

[0033] Step 103: For each tunnel traffic volume, normalize the tunnel traffic and internet traffic in the corresponding valid dual traffic data. Substitute the normalized tunnel traffic and internet traffic into the linear fitting formula. If the fitting error does not exceed the set value, the test is considered valid. Otherwise, return to the step of determining the linear fitting formula of tunnel traffic and internet traffic by performing multiple sets of dual traffic tests on the target network device under the condition that the CPU utilization of the target network device is within the first set range, until the fitting error corresponding to each traffic volume of the tunnel traffic does not exceed the set value.

[0034] Step 104: Output the tunnel traffic, internet traffic, normalized tunnel traffic, normalized internet traffic, and fitting error corresponding to the traffic volume range for each tunnel traffic.

[0035] This application solves the problem of "testing different network devices is too cumbersome": it allows test methods to be reused on different network devices without having to redesign the steps each time; it also solves the problem of "inaccurate results after network device upgrades": after device upgrades, the test error can still be controlled within 5%. Furthermore, it allows test results to reflect CPU resource usage, helping enterprises optimize device performance.

[0036] Key terms defined in this application include maximum throughput, normalized flow, and constant factor.

[0037] Maximum throughput refers to the maximum amount of data that a device can process when a single type of traffic is running alone—maximum Internet throughput (NetMax, unit: Mbps, i.e., the maximum Internet traffic when tunnel traffic is 0); maximum tunnel throughput (TunMax, unit: Mbps, i.e., the maximum tunnel traffic when Internet traffic is 0).

[0038] Normalized traffic: The ratio of real-time traffic to maximum throughput (eliminating performance differences between different network devices), specifically including: normalized internet traffic (NetNorm = NetReal / NetMax) and normalized tunnel traffic (TunNorm = TunReal / TunMax). NetReal is the actual internet traffic, and TunReal is the actual tunnel traffic.

[0039] A constant factor is a coefficient used to reflect the competition for CPU resources between the two traffic streams. It is fixed after testing and calibration and is used to calculate the throughput relationship when the two traffic streams coexist.

[0040] The core of this application is a dual-traffic normalization test bound to CPU load, which consists of three steps: first, a "basic test" of the network device is performed; then, dual-traffic is measured using general procedures; and finally, the accuracy of the results is verified.

[0041] Basic tests include testing maximum throughput, stable CPU load, and calibrating constant factors. a , b .

[0042] In an exemplary embodiment, step 101 specifically includes steps 201-203.

[0043] Step 201: Set three tunnel flow rates according to 30%, 60%, and 90% of the maximum tunnel flow rate.

[0044] Step 202: Under the condition that the CPU utilization of the target network device is within the first set range, perform three sets of dual-traffic tests on the target network device based on three levels of tunnel traffic, and record the corresponding Internet traffic for each set of dual-traffic tests; each level of tunnel traffic and its corresponding Internet traffic constitute three sets of dual-traffic test data.

[0045] Step 203: Based on the three sets of dual-flow test data, determine the linear fitting relationship between tunnel flow and Internet flow.

[0046] Step 101 is to calibrate the constant factor. a and b The process involves setting the tunnel traffic at 30%, 60%, and 90% of TunMax (e.g., if TunMax = 100Mbps, set it to 30%, 60%, and 90Mbps), and recording the corresponding Internet traffic (NetReal). Then, the normalized traffic is calculated (NetNorm = NetReal / NetMax, TunNorm = tunnel traffic / TunMax), and the normalized traffic is substituted into the formula "a × NetNorm + b × TunNorm = 1" to calculate the normalized traffic using a mathematical method (least squares). a and b The error of each set of data (the difference between the formula result and 1) is required to be no more than 3%.

[0047] TunMax's 30%, 60%, and 90% settings can cover low, medium, and high loads.

[0048] Among them, such as Figure 3 As shown, step 201 specifically includes: using 30% of the maximum tunnel traffic as the first level of tunnel traffic, measuring internet traffic for 5 minutes, recording internet traffic every minute, and taking the average of 5 measurements to obtain the first group of internet traffic; using 60% of the maximum tunnel traffic as the second level of tunnel traffic, measuring internet traffic for 5 minutes, recording internet traffic every minute, and taking the average of 5 measurements to obtain the second group of internet traffic; using 90% of the maximum tunnel traffic as the third level of tunnel traffic, measuring internet traffic for 5 minutes, recording internet traffic every minute, and taking the average of 5 measurements to obtain the third group of internet traffic. If the CPU fluctuation exceeds ±2% during any test group, the test is retested to ensure data reliability.

[0049] The above-mentioned 5-minute duration can also be 3 minutes, with each group measured for 3 minutes (recorded once every 30 seconds), with an error of ≤3.5%, suitable for scenarios where time is of the essence.

[0050] Specifically, step 203 includes: normalizing the tunnel flow and Internet flow in the three sets of dual-flow test data; and determining the linear fitting relationship between tunnel flow and Internet flow using the least squares method based on the normalized three sets of dual-flow test data, or using Excel's Solver to determine the linear fitting relationship between tunnel flow and Internet flow.

[0051] The linear fitting relationship is expressed as follows: .

[0052] in, a and b All are fitting coefficients. For normalized internet traffic, This represents the normalized tunnel flow rate.

[0053] a and b Defined as a constant factor, it is a coefficient used to reflect the competition relationship between dual traffic and CPU resources. It is fixed after being calibrated through testing and is used to calculate the throughput relationship when dual traffic coexists.

[0054] In step 103, the normalized tunnel traffic and internet traffic are substituted into the linear fitting equation to obtain the fitting error. Specifically, this includes substituting the normalized tunnel traffic and internet traffic into the linear fitting equation, and the difference between the calculated result and step 1 is the fitting error; the calculated result is... The difference between the calculated result and 1 is... -1.

[0055] In an exemplary embodiment, prior to step 101, the network device dual-traffic throughput testing method further includes measuring the maximum throughput: testing the target network device's maximum Internet traffic when tunnel traffic is 0, and the maximum tunnel traffic when Internet traffic is 0; the maximum tunnel traffic is used for tunnel traffic normalization, and the maximum Internet traffic is used for Internet traffic normalization.

[0056] More specifically, measuring maximum throughput includes: setting the tunnel traffic to 0 and slowly increasing the internet traffic until the traffic no longer increases, denoted as NetMax; conversely, setting the internet traffic to 0 and measuring the maximum value of the tunnel traffic, denoted as TunMax.

[0057] Before measuring maximum throughput, the process also includes: initializing the device: restoring the device to factory settings, turning off services such as log reporting and unnecessary calculations, leaving only the function of handling dual traffic, and avoiding unrelated programs occupying the CPU.

[0058] In an exemplary embodiment, prior to step 101, the network device dual-traffic throughput testing method further includes stabilizing CPU load: using a network testing tool to send preset background data to the target network device, so that the CPU utilization of the target network device is within a first set range.

[0059] More specifically, stabilizing the CPU load involves sending preset background data to the target network device using Hping3 or Iperf3, maintaining the CPU utilization of the target network device at 90%-95% (with fluctuations not exceeding ±5%). This is because the CPU is the primary competitive resource for dual-traffic applications, and a fixed load is necessary to ensure test stability.

[0060] In an exemplary embodiment, steps 102-103 are a dual-flow test process that can be repeated, including setting test parameters, collecting valid data, and calculating normalized flow.

[0061] Step 102 specifically includes steps 301-303.

[0062] Step 301: Select the data packet type to be tested (e.g., mixed packets), and divide the tunnel traffic into six traffic levels according to 0%, 20%, 40%, 60%, 80%, and 100% of the maximum tunnel traffic. These six traffic levels cover the full load, providing a more comprehensive test.

[0063] Step 302: For each tunnel traffic level, perform three dual-traffic tests on the target network device, and during the three dual-traffic tests, the rate of change of CPU utilization does not exceed the second set range.

[0064] Step 303: Calculate the average tunnel traffic based on the three tunnel traffic values ​​obtained from the three dual traffic tests for each traffic level, and calculate the average internet traffic value based on the three internet traffic values. The average tunnel traffic value and the average internet traffic value constitute a valid dual traffic data pair.

[0065] The first set range is greater than 90% and less than 95%, and the second set range is greater than -5% and less than 5%.

[0066] The set value is 3%.

[0067] In step 104, the tunnel traffic, internet traffic, normalized tunnel traffic, normalized internet traffic, and fitting error output are used to generate a report in tabular form. The report also includes the device model and CPU stability range.

[0068] The specific data collected includes: recording real-time internet traffic (NetReal), real-time tunnel traffic (TunReal), and CPU utilization during each test. If CPU usage fluctuates by more than ±5%, the data set is discarded and the test is repeated.

[0069] The verification process for this application includes substituting the normalized flow into the formula " a ×NetNorm + b If the difference between the result and 1 is no more than 5%, the test is valid; otherwise, recalibrate. a , b .

[0070] The final output includes a table summarizing tunnel traffic, internet traffic, normalized values ​​(normalized tunnel traffic and normalized internet traffic), and fitting error, and a graph showing the relationship between tunnel traffic and internet traffic to visually demonstrate the impact of the two types of traffic.

[0071] In one exemplary embodiment, the device is calibrated as "F1003-S". a , b "For example, this application illustrates the constant factor in a dual-traffic throughput testing method for network devices." a and b The calibration process, specifically the steps, is as follows.

[0072] Step 1: Prepare the equipment.

[0073] 1. Initialization: Restore factory settings, turn off log service, and leave only dual traffic processing function enabled.

[0074] 2. Measure maximum throughput: NetMax=2000Mbps (when tunnel traffic=0), TunMax=1000Mbps (when internet traffic=0).

[0075] 3. Stable CPU: When sending background data using Hping3, the CPU usage remained stable at 93%-95%, with a fluctuation of ±1.2%.

[0076] Step 2: Measure 3 sets of data.

[0077] 1. Set the gradients as follows: 300Mbps (30%×1000), 600Mbps (60%×1000), 900Mbps (90%×1000).

[0078] 2. Measurement data: At 300Mbps, 5 internet traffic spikes: 1700, 1710, 1695, 1705, 1700 → average 1702Mbps (Net1=1702). At 600Mbps, the average speed is 1402Mbps (Net2=1402). At 900Mbps, the average speed is 1102Mbps (Net3=1102). The CPU fluctuations for all three groups were ≤±1.2%, indicating that the data is valid.

[0079] Step 3: Calculate a , b And verify.

[0080] 1. Normalization: NetNorm1=1702 / 2000≈0.851, TunNorm1=300 / 1000=0.3; NetNorm2=1402 / 2000≈0.701, TunNorm2=600 / 1000=0.6; NetNorm3=1102 / 2000≈0.551, TunNorm3=900 / 1000=0.9; 2. Calculation a , b Substituting into the formula yields the system of equations → a ≈0.75, b ≈0.8, with an error of ≤0.6% for each group.

[0081] 3. Reuse verification: Test with a larger data packet, 300Mbps tunnel traffic corresponds to Internet traffic.

[0082] 1705Mbps → NetNorm = 1705 / 2000 ≈ 0.852 → Error = |0.75×0.852 + 0.8×0.3-1| ≈ 0.1% ≤ 3%, can be reused.

[0083] Step 4: Generate the test report.

[0084] Test report highlights: F1003-S device, NetMax=2000Mbps, TunMax=1000Mbps, CPU stability 93%-95% (±1.2%). a =0.75, b =0.8, error ≤0.6%, can be used across mixed packages / large packages.

[0085] The constant factor calibration method of this application is faster, reducing the calibration time from 8 hours to 2.5 hours per calibration, which is 70% faster. a , b More stable: How long can the equipment be used? a , b Deviations are all ≤2%, and formula errors are ≤5%. Reusable: No need to recalculate when changing data packet types, reducing costs by 50%.

[0086] Multi-resource control: When the bottleneck of a device is memory, add a memory factor. c The fitting relationship was replaced with " a ×NetNorm+ b ×TunNorm+ c ×MemNorm=1.

[0087] In an exemplary embodiment, taking the "SASE200 device mixed packet test scenario" as an example, this application describes a method for testing the throughput of dual traffic of a network device. The specific steps are as follows.

[0088] Step 1: Basic test.

[0089] 1. Measure the maximum throughput: When the tunnel traffic is 0, the maximum Internet traffic is 295.6Mbps, so NetMax=295.6Mbps; when the Internet traffic is 0, the maximum tunnel traffic is 100Mbps, so TunMax=100Mbps.

[0090] 2. Stable CPU load: When sending background data using Hping3, the CPU load remained stable at 92%, fluctuating by ±1.5%.

[0091] 3. Calibration a , b .

[0092] Tunnel traffic 30Mbps (30% × 100) → Internet traffic 220Mbps → NetNorm = 220 / 295.6 ≈ 0.744, TunNorm = 30 / 100 = 0.3; Tunnel traffic 60Mbps (60% × 100) → Internet traffic 145Mbps → NetNorm = 145 / 295.6 ≈ 0.490, TunNorm = 60 / 100 = 0.6; Tunnel traffic 90Mbps (90% × 100) → Internet traffic 70Mbps → NetNorm = 70 / 295.6 ≈ 0.237, TunNorm = 90 / 100 = 0.9; Substituting into the formula: 0.744 a +0.3 b =1; 0.490 a +0.6 b =1; 0.237 a +0.9 b =1→Calculate a ≈0.98, b ≈1.05, and the error (fitting error) of each group is ≤2.8%.

[0093] Step 2: Dual-flow test.

[0094] 1. Setting parameters: Mixed package, tunnel traffic is divided into 6 levels: 0, 20, 40, 60, 80, 100Mbps (corresponding to 0, 20%×100...100%×100).

[0095] 2. Collect valid data.

[0096] When the tunnel traffic is 20Mbps, the CPU fluctuation is ±2%. Internet traffic was measured 3 times: 240, 238, 242 → average 240Mbps → NetNorm=240 / 295.6≈0.812, TunNorm=20 / 100=0.2; When the tunnel traffic is 40Mbps, the CPU fluctuation is ±1.8%, and the average Internet traffic is 185Mbps → NetNorm=185 / 295.6≈0.626, TunNorm=40 / 100=0.4.

[0097] Step 3: Verification and Output 1. Check accuracy: 20Mbps range: 0.98×0.812+1.05×0.2≈0.996 → the difference between 0.4% and 1 is ≤5%; 40Mbps range: 0.98×0.626+1.05×0.4≈0.999→difference 0.1%≤5%.

[0098] 2. Report generation: Tables 1-4 contain data for each range, including "For every 20Mbps increase in tunnel traffic, internet traffic decreases by approximately 55-60Mbps".

[0099] Table 1 Internet Bandwidth Report

[0100] Here, g_LttMaxNum, packet_pool_size, and ltt_packet_pool_size represent the global maximum number of Light TCP Tunnel (Ltt) entries, the general packet memory pool size, and the Ltt-specific packet memory pool size, respectively. In LightWAN, Ltt refers to "Light TCP Tunnel," a tunneling technology. Ltt is used for SD-WAN Overlay tunnels to improve network transmission efficiency and security.

[0101] Table 2 LightWAN Network Bandwidth Report

[0102] Table 3 Report on LightWAN Networking + Internet Hybrid Scenarios

[0103] Table 4 Report on LightWAN Networking + Internet Hybrid Scenarios 2

[0104] This application has calibrated the constant factor. a , b Afterwards, a different data packet type was tested (e.g., from mixed packets to large packets), and the verification error remained ≤3%, indicating that... a , b It can be used in different scenarios. That is, when switching data packet types, there is no need to relabel. a , b .

[0105] In an exemplary embodiment, gradient adjustment: if the maximum tunnel flow of the device is only 80%×TunMax (e.g., low-end devices), the calibration gradient can be changed to 20%, 50%, or 80%×TunMax, and the error is still ≤4%.

[0106] Multi-resource association: If the bottleneck of the network device is memory, the memory utilization rate can be fixed at 85%-90% simultaneously by adding a memory factor to the formula. c ,become" a ×NetNorm+ b ×TunNorm+ c ×MemNorm=1” (MemNorm is the memory normalization value).

[0107] Algorithm optimization: For high-load data (such as 90%×TunMax), the weights of multiple points can be calculated, and the error can be reduced to ≤3%, which is suitable for high-precision scenarios.

[0108] The beneficial effects of the technical solution in this application are as follows.

[0109] Faster equipment testing: The testing time for 10 types of equipment has been reduced from 45 days to 18 days, a 60% reduction.

[0110] The results remain accurate after the upgrade: After the equipment is upgraded, only the NetMax test needs to be repeated, and the error will still be ≤4.5%.

[0111] It can optimize the device: the error can indicate whether the CPU is being used well. For example, when the error exceeds 8%, adjusting the CPU scheduling strategy can increase the device's maximum Internet traffic by 8%.

[0112] Easy to use: You can calculate it with Excel. a , b Even factory testing personnel can easily learn to use it.

[0113] The technical features of the above embodiments can be combined in any way. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0114] This document uses specific examples to illustrate the principles and implementation methods of this application. The descriptions of the above embodiments are only for the purpose of helping to understand the methods and core ideas of this application. Furthermore, those skilled in the art will recognize that, based on the ideas of this application, there will be changes in the specific implementation methods and application scope. Therefore, the content of this specification should not be construed as a limitation of this application.

Claims

1. A method for testing the throughput of a network device using dual traffic, characterized in that, The throughput testing method for dual-traffic networks includes: Under the condition that the CPU utilization of the target network device is within a first set range, the linear fitting relationship between tunnel traffic and Internet traffic is determined by performing multiple sets of dual-traffic tests on the target network device. The tunnel traffic is divided into multiple traffic levels. For each traffic level of the tunnel traffic, a dual traffic test is performed on the target network device. During the dual traffic test, real-time tunnel traffic and Internet traffic are collected. If the rate of change of CPU utilization exceeds the second set range during the collection process, the dual traffic test is performed again on the target network device for that traffic level until the rate of change of CPU utilization during the collection process is within the second set range, and a valid dual traffic data pair for that traffic level is obtained. For each tunnel traffic volume, the corresponding valid dual traffic data is normalized for both tunnel traffic and internet traffic. The normalized tunnel traffic and internet traffic are then substituted into the linear fitting formula to obtain the fitting error. If the fitting error does not exceed a set value, the test is considered valid. Otherwise, the process returns to the step of determining the linear fitting formula for tunnel traffic and internet traffic by performing multiple sets of dual traffic tests on the target network device under the condition that the CPU utilization of the target network device is within a first set range, until the fitting error corresponding to each traffic volume of the tunnel traffic does not exceed the set value. Output the tunnel traffic, internet traffic, normalized tunnel traffic, normalized internet traffic, and fitting error corresponding to the traffic level of each tunnel traffic.

2. The method for testing the throughput of network devices with dual traffic according to claim 1, characterized in that, Under the condition that the CPU utilization of the target network device is within a first set range, the linear fitting relationship between tunnel traffic and Internet traffic is determined by performing multiple sets of dual-traffic tests on the target network device, specifically including: Three tunnel flow rates are set according to 30%, 60%, and 90% of the maximum tunnel flow rate; Under the condition that the CPU utilization of the target network device is within a first set range, three sets of dual-traffic tests are performed on the target network device based on three levels of tunnel traffic, and the corresponding Internet traffic is recorded during each set of dual-traffic tests; each level of tunnel traffic and its corresponding Internet traffic constitute three sets of dual-traffic test data; Based on three sets of dual-flow test data, a linear fitting formula for tunnel flow and Internet flow was determined.

3. The method for testing the throughput of network devices with dual traffic according to claim 2, characterized in that, Under the condition that the CPU utilization of the target network device is within a first set range, three sets of dual-traffic tests are performed on the target network device based on three levels of tunnel traffic. The Internet traffic corresponding to each set of dual-traffic tests is recorded, specifically including: Take 30% of the maximum tunnel flow as the first level of tunnel flow, measure the Internet flow for 5 minutes, record the Internet flow once every 1 minute, and take the average of 5 times to obtain the Internet flow of the first group. Use 60% of the maximum tunnel flow as the second level of tunnel flow, measure the Internet flow for 5 minutes, record the Internet flow once every 1 minute, and take the average of 5 times to obtain the Internet flow of the second group. The third level of tunnel flow was determined by using 90% of the maximum tunnel flow. Internet traffic was measured for 5 minutes, with one record of internet traffic every minute. The average of the five records was taken to obtain the third group of internet traffic.

4. The method for testing the throughput of network devices with dual traffic according to claim 2, characterized in that, Based on three sets of dual-flow test data, a linear fitting formula for tunnel traffic and Internet traffic was determined, specifically including: Normalize the tunnel traffic and Internet traffic in the three sets of dual-flow test data; Based on the normalized three sets of dual-flow test data, the linear fitting relationship between tunnel flow and Internet flow is determined by the least squares method, or by using Excel's Solver tool to determine the linear fitting relationship between tunnel flow and Internet flow. The linear fitting relationship is expressed as follows: ; in, a and b All are fitting coefficients. For normalized internet traffic, This represents the normalized tunnel flow rate.

5. The method for testing the throughput of a network device with dual traffic according to claim 4, characterized in that, Substituting the normalized tunnel traffic and internet traffic into the linear fitting equation, we obtain the fitting error, which specifically includes: Substituting the normalized tunnel traffic and internet traffic into the linear fitting equation, the difference between the calculated result and 1 is the fitting error; the calculated result is... .

6. The method for testing the throughput of a network device with dual traffic according to claim 1, characterized in that, Before determining the linear fitting relationship between tunnel traffic and Internet traffic by conducting multiple sets of dual-traffic tests on the target network device, under the condition that the CPU utilization of the target network device is within a first set range, the dual-traffic throughput test method of the network device further includes: The test measures the target network device's maximum internet traffic when tunnel traffic is zero, and the maximum tunnel traffic when internet traffic is zero. The maximum tunnel traffic is used for tunnel traffic normalization, and the maximum internet traffic is used for internet traffic normalization.

7. The method for testing the throughput of a network device with dual traffic according to claim 1, characterized in that, Before determining the linear fitting relationship between tunnel traffic and Internet traffic by conducting multiple sets of dual-traffic tests on the target network device, under the condition that the CPU utilization of the target network device is within a first set range, the dual-traffic throughput test method of the network device further includes: A network testing tool is used to send preset background data to the target network device, so that the CPU utilization of the target network device is within a first set range.

8. The method for testing the throughput of a network device with dual traffic according to claim 1, characterized in that, The tunnel traffic is divided into multiple traffic tiers. For each traffic tier, a dual-traffic test is performed on the target network device. During the dual-traffic test, real-time tunnel traffic and internet traffic are collected. If the rate of change in CPU utilization exceeds a second preset range during the collection process, the dual-traffic test is repeated on the target network device for that traffic tier until the rate of change in CPU utilization during the collection process falls within the second preset range, thus obtaining a valid dual-traffic data pair for that traffic tier. Specifically, this includes: The tunnel flow rate is divided into six flow rate levels based on 0%, 20%, 40%, 60%, 80%, and 100% of the maximum tunnel flow rate. For each tunnel traffic level, the target network device is subjected to three dual-traffic tests, and the rate of change in CPU utilization during the three dual-traffic tests does not exceed the second set range. The average tunnel traffic is calculated based on the three tunnel traffic values ​​obtained from three dual-traffic tests for each traffic level, and the average internet traffic is calculated based on the three internet traffic values. The average tunnel traffic value and the average internet traffic value constitute a valid dual-traffic data pair.

9. The method for testing the throughput of a network device with dual traffic according to claim 1, characterized in that, The first set range is greater than 90% and less than 95%, and the second set range is greater than -5% and less than 5%.

10. The method for testing the throughput of a network device with dual traffic according to claim 1, characterized in that, The set value is 3%.

Images