Storage device performance evaluation method and device based on dynamic pressure regulation

Abstract

The present application discloses a storage device performance evaluation method and apparatus based on dynamic pressure regulation, which relates to the technical field of storage device performance testing, including: generating a pressure thread and pressurizing a storage device; monitoring the performance indicators of the storage device and recording performance data; analyzing the performance data, calculating the variance value of the performance data, judging whether the performance of the storage device has reached a steady state, and obtaining corresponding analysis results; dynamically adjusting the pressure value according to the performance data and the analysis results, thereby solving the technical problem in related technologies of relying too much on manual experience to set pressure parameters, resulting in low efficiency of storage performance testing and poor accuracy of test results. The method achieves the technical effect of accurately identifying the performance bottleneck of the storage device by monitoring the performance indicators of the storage device in real time and dynamically adjusting the test pressure in combination with an adaptive algorithm, so as to further test the operating status of the storage device under different performance conditions, thereby improving the design level and quality level of the product.

Description

Technical Field

[0001] The present application relates to the technical field of storage device performance testing, and in particular to a storage device performance evaluation method and apparatus based on dynamic pressure regulation. Background Art

[0002] During storage device performance testing, as the test pressure gradually increases, the performance of the storage device typically shows an upward trend until it reaches a performance bottleneck and then stabilizes. However, current testing methods lack a precise control mechanism to apply test pressure.

[0003] Currently, although existing benchmarking tools have multi-mode load simulation capabilities, the setting of their test parameters is still highly dependent on manual configuration; although existing stress testing frameworks can simulate high-concurrency access scenarios, they lack dynamic stress adjustment functions; in addition, although existing performance monitoring systems can collect key indicators such as CPU (Central Processing Unit) utilization, memory usage and disk I / O (Input / Output), they have not yet achieved closed-loop linkage with stress generation modules.

[0004] In addition, existing storage performance testing methods have the following defects:

[0005] 1. Subjectivity of pressure parameters:

[0006] The number of concurrent threads or I / O load intensity required during testing is usually manually preset based on experience, which makes it difficult to accurately match the actual load capacity of the device. This can lead to resource waste and insufficient test pressure, making it impossible to accurately identify performance critical points.

[0007] 2. Static nature of the test pattern:

[0008] The load-raising method with a fixed number of threads lacks a dynamic adjustment mechanism, cannot respond to changes in device performance in real time, and is difficult to fully evaluate the performance characteristics of the device under different load conditions;

[0009] 3. Lack of steady-state judgment:

[0010] The lack of an automated mechanism for detecting stable performance during the testing process directly affects the accuracy and reliability of the final evaluation results.

[0011] 4. Insufficient testing efficiency:

[0012] Due to the reliance on trial and error to adjust pressure parameters, a single test cycle often takes several hours or even longer, significantly reducing test efficiency and making it difficult to meet the needs of rapid iterative development.

[0013] In summary, existing storage performance testing technology relies too much on manual experience to set stress parameters, resulting in low efficiency of storage performance testing and poor accuracy of test results, which urgently needs to be solved. Summary of the Invention

[0014] The present application provides a storage device performance evaluation method and apparatus based on dynamic pressure regulation, in order to at least solve the technical problem in the related art that the pressure parameters are set too much by manual experience, resulting in low efficiency of storage performance testing and poor accuracy of test results.

[0015] The present application provides a storage device performance evaluation method based on dynamic pressure regulation, comprising the following steps: based on a preset initial pressure value, pressurizing a target storage device, and after the target storage device is pressurized for a preset period of time, collecting the performance index value of the target storage device at preset intervals to generate multiple groups of performance index values ​​of the target storage device; calculating the variance value of each group of performance index values ​​in the multiple groups of performance index values, and judging whether the target storage device meets the preset performance steady-state requirements based on the variance value, wherein, if the target storage device meets the performance steady-state requirements, obtaining the current pressure value and the current performance index value of the target storage device to judge the current performance steady-state requirements. Whether the indicator value is within a preset expected range; if the current performance indicator value is not within the expected range, then based on the preset step value, increase or decrease the current pressure value, and use the increased or decreased current pressure value to re-pressurize the target storage device; if the current performance indicator value is within the expected range, then iteratively reduce the current pressure value of the target storage device, and match the performance indicator value corresponding to the reduced current pressure value in each iteration process, so as to determine the performance bottleneck point of the target storage device based on the matched performance indicator value, and generate the performance evaluation test result of the target storage device according to the target pressure value and target performance indicator value corresponding to the performance bottleneck point.

[0016] The present application also provides a storage device performance evaluation device based on dynamic pressure regulation, comprising: a storage device pressurization module, for pressurizing a target storage device based on a preset initial pressure value, and after the target storage device is pressurized for a preset period of time, collecting the performance index value of the target storage device at preset intervals to generate multiple groups of performance index values ​​of the target storage device; a performance steady-state judgment module, for calculating the variance value of each group of performance index values ​​in the multiple groups of performance index values, and judging whether the target storage device meets the preset performance steady-state requirements based on the variance value, wherein, if the target storage device meets the performance steady-state requirements, the current pressure value and the current performance index value of the target storage device are obtained to judge the Whether the current performance index value is within a preset expected range; a pressure-performance matching module, which is used to increase or decrease the current pressure value based on a preset step value if the current performance index value is not within the expected range, and re-pressurize the target storage device using the increased or decreased current pressure value; if the current performance index value is within the expected range, iteratively reduce the current pressure value of the target storage device, and match the performance index value corresponding to the reduced current pressure value during each iteration, so as to determine the performance bottleneck point of the target storage device based on the matched performance index value, and generate the performance evaluation test result of the target storage device according to the target pressure value and target performance index value corresponding to the performance bottleneck point.

[0017] The present application also provides an electronic device, comprising: a memory for storing a computer program; and a processor for implementing the steps of any of the above-mentioned storage device performance evaluation methods based on dynamic pressure regulation when executing the computer program.

[0018] The present application also provides a non-volatile computer-readable storage medium, in which a computer program is stored. When the computer program is executed by a processor, the steps of any of the above-mentioned storage device performance evaluation methods based on dynamic pressure regulation are implemented.

[0019] The present application also provides a computer program product, including a computer program, which, when executed by a processor, implements the steps of any of the above-mentioned storage device performance evaluation methods based on dynamic pressure regulation.

[0020] Through the present application, the target storage device can be pressurized based on a preset initial pressure value, and after the target storage device is pressurized for a preset period of time, the performance index value of the target storage device is collected at preset intervals to generate multiple groups of performance index values ​​of the target storage device; the variance value of each group of performance index values ​​in the multiple groups of performance index values ​​is calculated, and whether the target storage device meets the preset performance steady-state requirements is judged based on the variance value, wherein, if the target storage device meets the performance steady-state requirements, the current pressure value and the current performance index value of the target storage device are obtained to judge whether the current performance index value is within a preset expected range; if the current performance index value is not within the expected range, the current pressure value is increased or decreased based on a preset step value, and the target storage device is re-pressurized using the increased or decreased current pressure value; if the current performance index If the value is within the expected range, the current pressure value of the target storage device is iteratively reduced, and the performance index value corresponding to the current pressure value after reduction in each iteration is matched, so as to determine the performance bottleneck point of the target storage device based on the matched performance index value, and generate the performance evaluation test result of the target storage device according to the target pressure value and the target performance index value corresponding to the performance bottleneck point. Therefore, it can solve the technical problem in the related technology that too much reliance on manual experience to set pressure parameters makes the efficiency of storage performance testing low and the accuracy of test results poor, and achieves the technical effect of accurately identifying the performance bottleneck of the storage device by monitoring the performance indicators of the storage device in real time and dynamically adjusting the test pressure with an adaptive algorithm, so as to further test the operating status of the storage device under different performance conditions, thereby improving the design level and quality level of the product. BRIEF DESCRIPTION OF THE DRAWINGS

[0021] In order to more clearly illustrate the embodiments of the present application, 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 application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without any creative work.

[0022] Figure 1 A flowchart of a storage device performance evaluation method based on dynamic pressure regulation according to an embodiment of the present application;

[0023] Figure 2 A schematic diagram of the logical architecture of a storage device performance evaluation method based on dynamic pressure regulation provided in one embodiment of the present application;

[0024] Figure 3 This is an example diagram of a storage device performance evaluation device based on dynamic pressure regulation according to an embodiment of the present application.

[0025] Among them, 10-storage device performance evaluation device based on dynamic pressure regulation, 100-storage device pressurization module, 200-performance steady-state judgment module, 300-pressure performance matching module. DETAILED DESCRIPTION

[0026] The following will be combined with the accompanying drawings in the embodiments of this application to clearly and completely describe the technical solutions in the embodiments of this application. Obviously, the embodiments described are only part of the embodiments of this application, not all of them. Based on the embodiments in this application, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of this application.

[0027] It should be noted that, in the description of this application, the terms "comprises," "includes," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or device 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 device. The terms "first," "second," etc., in this application are used to distinguish similar objects, and are not used to describe a particular order or sequence.

[0028] In order to enable those skilled in the art to better understand the present application, the present application is further described in detail below with reference to the accompanying drawings and specific implementation methods.

[0029] In conjunction with the specific application environment architecture or specific hardware architecture on which the execution of the storage device performance evaluation method based on dynamic pressure regulation depends, the specific application environment architecture or specific hardware architecture is described herein.

[0030] An embodiment of the present application provides a storage device performance evaluation method based on dynamic pressure regulation.

[0031] like Figure 1 FIG. 1 is a flow chart of a storage device performance evaluation method based on dynamic pressure regulation according to an embodiment of the present application, wherein the storage device performance evaluation method based on dynamic pressure regulation includes the following steps:

[0032] In step S101, the target storage device is pressurized based on a preset initial pressure value, and after the target storage device is pressurized for a preset time, performance index values ​​of the target storage device are collected at preset intervals to generate multiple groups of performance index values ​​of the target storage device.

[0033] Those skilled in the art should understand that with the rapid development of technologies such as cloud computing, big data, and AI (artificial intelligence), storage devices, as the core carriers of data storage, have a significant impact on the responsiveness, throughput, and stability of business systems. For example, financial trading systems must support high-concurrency transactions of tens of thousands per second and are sensitive to storage latency and IOPS (Input / Output Operations Per Second). Video streaming platforms must handle high-throughput read and write of petabytes of data, placing extremely high demands on storage bandwidth and concurrency. Distributed storage systems must ensure data consistency and scalability, and stress testing is a key method for verifying their reliability.

[0034] However, current performance testing methods for storage devices lack precise control mechanisms to apply test stress, as described below:

[0035] Subjective load control: Test stress relies on manually pre-setting the number of concurrent threads, but testers often struggle to accurately predict the actual stress intensity generated by a specific number of threads. For example, when 256 stress threads are preset, the device's performance may have already reached its limit at 128 threads, yet the test results may falsely indicate that the device can handle 256 threads.

[0036] Distortion of the performance curve: This extensive load control method will cause non-smooth transitions in the performance curve obtained from the test, which cannot accurately reflect the actual performance boundaries of the device, thereby affecting the accuracy and reliability of the test results.

[0037] In addition, although existing benchmarking tools such as Iometer support simulating multiple load modes, they require manual parameter configuration; existing stress testing frameworks such as JMeter and LoadRunner can generate high-concurrency requests, but they lack intelligent adjustment capabilities; currently, although performance monitoring systems can analyze performance by collecting indicators such as CPU, memory, and disk I / O, they are not linked to stress regulation.

[0038] Therefore, based on the technical problems and defects of the current performance testing method, the embodiment of the present application can propose a storage device performance evaluation method based on dynamic pressure regulation, so as to dynamically adjust the test pressure in combination with an adaptive algorithm, accurately identify the performance bottleneck of the storage device, and further test the operating status of the storage device under different performance conditions, thereby improving the design level and quality level of the product.

[0039] Specifically, the embodiments of the present application may first set initial parameters such as the expected value of the storage performance indicator (P\_target), the initial pressure value (T\_initial), and the increase / decrease step size (S\_step); for example, in the actual execution process, the embodiments of the present application may set the expected value of the storage performance indicator to 1000 IOPS (i.e., the number of input / output operations per second), the initial pressure value to 100 threads, and the increase / decrease step size to 10 threads.

[0040] Secondly, the embodiment of the present application can pressurize the storage device according to the initial pressure value, for example, start 100 threads to perform a pressurization operation on the storage device.

[0041] Afterwards, after a preset pressurization time, such as 10 minutes, the embodiment of the present application can monitor the performance indicators of the storage device and record the performance indicator values ​​of the storage device every few minutes to continuously record multiple groups of performance indicator values.

[0042] For example, an embodiment of the present application can record the performance index value of the storage device after pressurization for 10 minutes, record once every 1 minute, record continuously for 30 times, and divide it into 3 groups, each group of 10 times, thereby obtaining 3 groups of performance index values.

[0043] Therefore, the embodiment of the present application generates a pressure thread to pressurize the storage device, monitors the performance indicators of the storage device, and records the corresponding performance data, thereby providing reliable data guidance and basis for subsequent determination of performance steady state, etc.

[0044] In step S102, the variance value of each group of performance indicator values ​​in the multiple groups of performance indicator values ​​is calculated, and whether the target storage device meets the preset performance steady-state requirements is determined based on the variance value. If the target storage device meets the performance steady-state requirements, the current pressure value and the current performance indicator value of the target storage device are obtained to determine whether the current performance indicator value is within the preset expected range.

[0045] Furthermore, the embodiments of the present application respectively calculate the variance values ​​of multiple groups of performance indicator values ​​to determine whether the storage device meets the corresponding performance steady-state requirements based on the corresponding variance values. If the storage device meets the performance steady-state requirements, the current pressure value and the current performance indicator value of the storage device are recorded; otherwise, the initial pressure value is maintained and the storage device is continued to be pressurized for a preset period of time.

[0046] Therefore, the embodiments of the present application perform steady-state performance judgment on the storage device, thereby providing reliable data support for subsequent pressure-performance matching.

[0047] Optionally, in one embodiment of the present application, whether the target storage device meets the preset performance steady-state requirements is judged based on the variance value, wherein, when the target storage device meets the performance steady-state requirements, the current pressure value and the current performance index value of the target storage device are obtained, including: judging whether the target storage device meets the performance steady-state requirements based on multiple groups of performance index values; if the target storage device does not meet the performance steady-state requirements, pressurizing the target storage device for a preset time period according to the initial pressure value to re-collect new multiple groups of performance index values ​​corresponding to the target storage device after the preset time period; if the target storage device does not meet the performance steady-state requirements, determining the current pressure value and the current performance index value of the target storage device.

[0048] During the actual implementation process, the embodiments of the present application can calculate the variance values ​​of the three groups of data respectively based on the steady-state judgment algorithm, denoted as S1, S2, and S3. If S1≥S2≥S3 and 0≤S2-S3≤1 (that is, the performance steady-state requirement), it is determined that the performance of the storage device has reached a steady state, that is, the performance steady-state requirement is met; for example, when it is calculated that S1=5, S2=3 and S3=2, it satisfies S1≥S2≥S3 and 0≤S2-S3≤1, so it can be determined that the performance of the storage device has reached a steady state.

[0049] In an embodiment of the present application, if the performance of the storage device does not meet the steady-state requirements, the embodiment of the present application can maintain the initial pressure value and continue to pressurize the storage device for time T (such as 10 minutes), and re-monitor the performance indicators of the storage device to obtain new multiple sets of performance indicators.

[0050] If the performance of the storage device meets the steady-state requirement, the current pressure value and the current performance index value of the storage device are recorded for the next step of judgment.

[0051] Therefore, the embodiment of the present application determines whether the storage device has reached a steady state through variance analysis of multiple sets of performance indicator data, dynamically monitors the performance indicators of the storage device based on the judgment results, and adjusts the pressure value in real time to ensure the matching degree of pressure and performance during the test process.

[0052] In addition, during actual implementation, the embodiment of the present application may also perform stress testing and performance monitoring in the following manner to determine whether the storage device meets the steady-state requirements. The specific process is as follows:

[0053] 1. Intelligent pressure loading strategy

[0054] The embodiment of the present application may adopt a dual-mode pressure regulation system to automatically select the loading mode based on the device type:

[0055] (1) For SSD (Solid State Drive) devices, use "step-pulse hybrid loading": first increase the pressure stepwise by 10%, with each step maintained for 30 seconds; when the delay exceeds 200% of the baseline, switch to 5-second pulse pressurization;

[0056] (2) For HDD (Hard Disk Drive) devices, “temperature compensation loading” is used: the pressure increase is dynamically adjusted according to the head temperature sensor data. For every 5°C increase in temperature, the pressure increase is reduced by 2%.

[0057] 2. Construction of holographic performance portrait

[0058] Establish 3D monitoring model collection:

[0059] (1) Time dimension: histogram of read and write latency distribution with millisecond sampling;

[0060] (2) Spatial dimension: local performance heat map divided by storage block;

[0061] (3) Operational dimension: association matrix between random / sequential read / write ratio and error type.

[0062] 3. Steady-state determination based on federated learning

[0063] Deploy distributed edge computing nodes:

[0064] (1) Each node uses local data to train a lightweight GRU (Gated Recurrent Unit) model to extract the second-order derivative features of the performance curve;

[0065] (2) Exchange model parameters (not raw data) through a secure aggregation protocol;

[0066] (3) The central coordinator integrates the prediction results of the 10 nodes, and when the following conditions are met for 5 consecutive cycles:

[0067] 1) The variance of the predicted steady-state probability is less than 0.1;

[0068] 2) The difference in confidence between nodes is less than 15%;

[0069] It is determined that the global steady state is reached.

[0070] 4. Pressure release verification mechanism

[0071] Design reverse verification process:

[0072] (1) Phase 1: Maintain 90% of peak pressure for 1 minute, with throughput fluctuation required to be less than 3%;

[0073] (2) Second stage: After rapidly dropping to 50% pressure, verify that the delay recovers to within ±10% of the initial value within 15 seconds;

[0074] (3) Phase 3: After the pressure is completely unloaded, all error counters must stop increasing.

[0075] Therefore, the embodiments of the present application greatly improve the pressure loading efficiency by combining the intelligent loading strategy with the distributed learning framework, and can identify the "pseudo-steady state" phenomenon that traditional methods cannot detect.

[0076] In step S103, if the current performance index value is not within the expected range, the current pressure value is increased or decreased based on the preset step value, and the target storage device is re-pressurized using the increased or decreased current pressure value. If the current performance index value is within the expected range, the current pressure value of the target storage device is iteratively reduced, and the performance index value corresponding to the reduced current pressure value in each iteration process is matched, so as to determine the performance bottleneck point of the target storage device based on the matched performance index value, and generate the performance evaluation test result of the target storage device according to the target pressure value and target performance index value corresponding to the performance bottleneck point.

[0077] Afterwards, the embodiment of the present application can determine an expected interval based on the expected performance index value of the storage device set above. For example, the expected interval can be set from 95% of the expected performance index value to 105% of the expected performance index value, that is, the minimum value of the expected interval is 95% of the expected performance index value, and the maximum value of the expected interval is 105% of the expected performance index value; furthermore, the embodiment of the present application can determine whether the current performance index value is within the expected interval. If the current performance index value of the storage device is less than 95% of the expected performance index value, the current pressure value is increased according to the increase or decrease step size in the initial parameters determined above; if the current performance index value of the storage device is greater than 105% of the expected performance index value, the pressure value is decreased according to the increase or decrease step size.

[0078] In addition, if the current performance index value of the storage device is within the range of 95% - 105% of the set expected value (i.e., the expected interval), a pressure and performance matching algorithm is performed; specifically, the embodiment of the present application can reduce the current pressure value by half, and use the halved pressure value to pressurize the storage device for a preset period of time to determine whether the performance index value of the storage device decreases after the pressurization for the preset period of time, and iteratively perform the above operations until the performance index value decreases, then determine the performance bottleneck point of the storage device, and record the target pressure value and target performance index value corresponding to the performance bottleneck point.

[0079] Therefore, the embodiment of the present application dynamically adjusts the pressure value through the preset performance index value, and at the same time performs a matching judgment on the storage device performance index value and the pressure value, thereby ensuring the matching degree between the pressure value and the storage performance and improving the accuracy.

[0080] Optionally, in one embodiment of the present application, if the current performance index value is not within the expected range, the current pressure value is adjusted, and the target storage device is re-pressurized using the adjusted current pressure value, including: if the current performance index value of the target storage device is less than the minimum value of the expected range, then based on the preset step value and the dynamic pressure adjustment mechanism, the current pressure value is increased, so as to use the increased current pressure value to pressurize the target storage device for a preset period of time; if the current performance index value of the target storage device is greater than the maximum value of the expected range, then based on the step value and the dynamic pressure adjustment mechanism, the current pressure value is reduced, so as to use the reduced current pressure value to pressurize the target storage device for a preset period of time.

[0081] During the specific implementation process, the embodiments of the present application can determine whether the performance of the storage device meets the expected value based on the performance index value of the storage device; if the current performance index value of the storage device is less than 95% of the set expected value (that is, the minimum value of the expected range), it indicates that the current pressure does not meet the performance expectations of the storage device. The embodiments of the present application can increase the pressure value according to the set step value based on the dynamic pressure adjustment mechanism, so as to use the increased pressure value to pressurize the storage device for a preset period of time.

[0082] For example, if the current performance indicator value is 900 IOPS, it is less than 95% (950 IOPS) of the expected value of 1000 IOPS. Therefore, the embodiment of the present application can increase the pressure value by 10 threads to re-pressurize and monitor the storage device.

[0083] If the current performance index value of the storage device is greater than 105% of the set expected value (i.e., the maximum value of the expected range), it indicates that the current pressure exceeds the performance expectations of the storage device. The embodiment of the present application can reduce the pressure value according to the set step size based on the dynamic pressure adjustment mechanism to use the reduced pressure value to pressurize the storage device for a preset time period.

[0084] For example, if the current performance indicator value is 1050IOPS, it is greater than 105% (1050IOPS) of the expected value of 1000IOPS. Therefore, the embodiment of the present application can reduce the pressure value by 10 threads to re-pressurize and monitor the storage device by reducing the pressure value by 10 threads.

[0085] In addition, during the actual execution process, if the performance indicator value of the storage device is within the range of 95%-105% of the set expected value, the embodiment of the present application can perform a pressure and performance matching algorithm determination.

[0086] It should be noted that the pressure-performance matching algorithm of the embodiment of the present application can be divided into two stages: preliminary matching and precise matching to verify the correlation between pressure and performance, as described below:

[0087] 1. Preliminary matching: In the embodiment of the present application, the current pressure value can be halved, and the storage device can be pressurized for a preset period of time using the halved pressure value to obtain the performance index value after pressurization, and the performance of the storage device can be judged based on the performance index value after pressurization. If the performance drops by more than 5%, the pressure is too high; if it remains unchanged, it indicates that there is redundancy.

[0088] 2. Precise matching: Gradually approach the optimal pressure value through the dichotomy method until the performance fluctuation range is ≤5%.

[0089] Therefore, the embodiments of the present application monitor the performance indicators of storage devices in real time and dynamically adjust the test pressure in combination with an adaptive algorithm, thereby accurately identifying the performance bottlenecks of storage devices, effectively reducing manual intervention, shortening the test cycle by more than 50%, and solving the problems of low test efficiency and inaccurate results caused by traditional manual setting of pressure parameters.

[0090] Optionally, in one embodiment of the present application, the current pressure value is increased based on a preset step value and a dynamic pressure regulation mechanism, including: calculating the performance indicator difference between the current performance indicator value and the preset performance indicator expected value; constructing a dynamic pressure regulation expression based on the performance indicator difference, the current pressure value and the preset regulation coefficient to determine the dynamic pressure regulation mechanism according to the dynamic pressure regulation expression.

[0091] It should be noted that the embodiment of the present application can calculate the performance index difference between the current performance index value and the expected performance index value; furthermore, the embodiment of the present application can adopt a step size trial strategy combined with a feedback control model to dynamically adjust the pressure value according to the performance index difference between the performance index and the expected value to determine the dynamic pressure regulation mechanism. The dynamic pressure regulation expression corresponding to the dynamic pressure regulation mechanism is as follows:

[0092] P new =P current + α ·(T target -T current )·ln(1+ β ΔT)

[0093] Among them, P current is the current pressure value; T current is the current performance index value; T target is the expected value of the performance indicator; T is the performance indicator (such as IOPS); α 、 β are adjustment coefficients; ΔT is the performance index difference.

[0094] Therefore, the embodiments of the present application can dynamically adjust pressure parameters through an adaptive dynamic pressure regulation mechanism, thereby accurately identifying performance bottlenecks and obtaining corresponding pressure values.

[0095] Optionally, in one embodiment of the present application, if the current performance indicator value is within the expected range, a pressure-performance matching operation is performed to determine the performance bottleneck point of the target storage device, including: lowering the current pressure value to obtain a corresponding pressure verification value, and pressurizing the target storage device for a preset period of time using the pressure verification value to collect the performance indicator verification value after pressurization for the preset period of time; judging whether the performance indicator verification value is less than the minimum value of the expected range; when the performance indicator verification value is less than the minimum value of the expected range, increasing the pressure verification value based on a preset dynamic pressure adjustment mechanism and step value to pressurize the target storage device for a preset period of time using the increased pressure verification value; when the performance indicator verification value is within the expected range, continuing to lower the pressure verification value to obtain a new pressure verification value, and re-pressurizing the target storage device for a preset period of time using the new pressure verification value to obtain a new performance indicator verification value after pressurization for the preset period of time; based on the new pressure verification value and the new performance indicator verification value, iteratively performing pressurization and performance indicator value collection operations on the target storage device until the collected performance indicator value is less than the minimum value of the expected range to determine the performance bottleneck point of the target storage device.

[0096] It should be noted that the process of determining the pressure and performance matching algorithm in the embodiment of the present application is as follows:

[0097] In the specific implementation process, the embodiment of the present application can halve the current pressure value based on the pressure halving strategy, and after pressurizing the storage device for a preset period of time (such as 10 minutes) using the halved current pressure value (i.e., the pressure verification value), monitor the new performance indicator value (i.e., the performance indicator verification value) corresponding to the storage device, and judge whether the performance of the storage device has declined based on the performance indicator verification value. If the performance indicator verification value drops below 95% of the set expected value, that is, less than the minimum value of the expected interval, the pressure verification value is increased according to the step value, so as to pressurize the storage device for a preset period of time using the increased pressure verification value, and re-execute the corresponding performance monitoring operation.

[0098] For example, in an embodiment of the present application, if the current pressure value is 100 threads, the embodiment of the present application can be adjusted to 50 threads, and the storage device is pressurized by 50 threads for 10 minutes to determine whether the performance of the storage device has decreased after ten minutes of pressurization; if the performance indicator value (i.e., the performance indicator verification value) drops to 850 IOPS, which is less than the expected value of 950 IOPS, the pressure value is increased by 10 threads, and the corresponding pressurization and performance monitoring operations are re-executed.

[0099] In addition, if the performance indicator verification value is within the expected range, that is, the performance of the storage device remains unchanged, it indicates that the current performance of the storage device has reached a bottleneck, and the pressure exceeds at least half. At this time, the embodiment of the present application can continue to halve the current pressure value to obtain a new pressure verification value, and monitor the new performance indicator verification value after the new pressure verification value is applied to the storage device for a preset period of time, and judge whether the performance of the storage device has decreased based on the new pressure verification value; thereafter, the embodiment of the present application can iteratively perform the above operations according to the bottleneck detection mechanism until the performance of the storage device decreases, so as to obtain the performance bottleneck point of the storage device, and record the corresponding current pressure value and performance value at this time. For example, after adjusting the pressure value, the performance indicator value remains unchanged at 1000 IOPS, and the pressure value continues to be reduced to 25 threads, and the performance value is checked until the performance drops below 950 IOPS, and the corresponding current pressure value and performance value (i.e., the target pressure value and the target performance indicator value) are recorded.

[0100] Therefore, the embodiments of the present application are based on a pressure and performance matching algorithm that includes a pressure halving strategy and a bottleneck detection mechanism. It detects performance bottlenecks by gradually reducing pressure and records the current pressure value and performance value, thereby accurately determining the maximum pressure that the storage device can withstand under given performance indicators, thereby improving the accuracy and reliability of the test.

[0101] As an achievable method, the embodiment of the present application can also achieve dynamic location of storage device performance bottlenecks in the following manner. The specific process is as follows:

[0102] 1. Intelligent pressure waveform generation

[0103] Adaptive pressure waveform synthesis technology is used to dynamically generate test loads that meet the characteristics of storage devices:

[0104] (1) Basic waveform: Generate standard pressure waveforms such as step type, pulse type, and ramp type according to equipment specifications;

[0105] (2) Modulation waveform: Superimpose high-frequency jitter (±3% pressure fluctuation) to simulate real environmental interference;

[0106] (3) Intelligent parameter adjustment: Automatically adjust the waveform amplitude / frequency based on real-time feedback to ensure that the pressure value is always within the device response sensitive range.

[0107] 2. Multi-granularity performance sampling

[0108] Establish a three-level performance observation system:

[0109] (1) Microscopic level (μs level): Capturing bus-level signal integrity (eye opening, clock jitter);

[0110] (2) Meso-level (ms level): Statistics on command queue scheduling efficiency (percentage of valid instructions, arbitration delay);

[0111] (3) Macro layer (s-level): Monitor media-level health (bad block growth trend, write amplification factor).

[0112] 3. Pressure-performance characteristic alignment

[0113] Constructing a time domain-frequency domain joint analysis model:

[0114] (1) Time domain alignment: Matching pressure changes and performance fluctuation curves through dynamic time warping algorithm;

[0115] (2) Frequency domain decomposition: Use wavelet transform to extract the frequency domain characteristics of performance indicators and identify the resonant frequency points with pressure harmonics;

[0116] (3) Mutation detection: Capturing abnormal deviation points of performance indicators.

[0117] 4. Decoupling of bottleneck features

[0118] Use a three-step analysis approach to isolate bottleneck causes:

[0119] (1) Hardware layer decoupling: identifying the contribution of controller / medium / interface through pressure-delay transfer function;

[0120] (2) Protocol layer decoupling: Analyze the correlation between command retry rate and pressure value to locate the bottleneck of protocol stack;

[0121] (3) System-level decoupling: Establish a ternary relationship model of energy consumption, temperature, and pressure to discover the heat dissipation limitation point.

[0122] 5. Dynamic bottleneck verification

[0123] Execute the closed-loop verification process:

[0124] (1) Forward verification: Maintain pressure at the suspected bottleneck point and observe whether the performance degradation rate meets the prediction;

[0125] (2) Reverse verification: suddenly release the pressure and detect the hysteresis characteristics of the performance recovery trajectory;

[0126] (3) Cross-validation: Alternately shield different subsystems and observe the migration patterns of bottleneck features.

[0127] 6. Bottleneck heat map generation

[0128] Build a 3D visualization model:

[0129] (1) X-axis: pressure value (from threshold 10% to 120%);

[0130] (2) Y-axis: performance indicators (latency / throughput / error rate);

[0131] (3) Z-axis: bottleneck intensity index (comprehensive hardware / protocol / system impact factors);

[0132] (4) Color coding: red indicates the absolute bottleneck area, and blue indicates the safe operating area.

[0133] Therefore, the embodiments of the present application can achieve dynamic positioning of storage device performance bottlenecks through strategies such as intelligent waveform synthesis, multi-granularity observation, time-frequency joint analysis, and dynamic decoupling verification, thereby identifying a variety of implicit performance limiting factors and guiding storage devices to achieve optimal performance output within a safe boundary.

[0134] Optionally, in one embodiment of the present application, a performance evaluation test result of the target storage device is generated based on the target pressure value and the target performance index value corresponding to the performance bottleneck point, including: determining the performance fluctuation range and steady-state achievement time corresponding to the target storage device based on the target pressure value and the target performance index value; drawing the performance curve, pressure-performance scatter plot and bottleneck analysis heat map corresponding to the target storage device based on the target pressure value, the target performance index value, the performance fluctuation range and the steady-state achievement time; generating the performance evaluation test result based on the performance fluctuation range, steady-state achievement time, the performance curve, the pressure-performance scatter plot and the bottleneck analysis heat map.

[0135] Specifically, the embodiments of the present application can determine key indicators such as the performance fluctuation range and steady-state achievement time corresponding to the storage device based on the above-mentioned target pressure value (i.e., the optimal pressure value) and the target performance index value; thereafter, the embodiments of the present application can use the above-mentioned key indicators to draw performance curves, pressure-performance scatter plots, bottleneck analysis heat maps and other visual charts to generate the final performance evaluation test report, i.e., the performance evaluation test results, and send the report to the corresponding R&D test personnel.

[0136] It can be understood that the pressure values ​​evaluated in the embodiments of the present application and the corresponding performance evaluation test reports are of vital importance to R&D testers. They can enable R&D testers to effectively reduce the number and time of pressure value tests each time they attempt, accurately control the performance of storage devices, and thus further test the operating status of storage devices under different performance conditions, thereby improving the design level and quality level of the product.

[0137] Optionally, in one embodiment of the present application, based on the target pressure value and the target performance index value, the performance fluctuation range and steady-state achievement time corresponding to the target storage device are determined, including: starting from the target pressure value, based on a preset step value, gradually adjusting the pressure value in at least one target direction, and obtaining the performance data corresponding to each adjusted pressure value, and eliminating the instantaneous fluctuation interference of the performance data to obtain the performance attenuation value corresponding to each adjusted pressure value; judging whether the performance attenuation value is greater than a preset threshold, wherein when the performance attenuation value is greater than the preset threshold, stopping the gradual adjustment of the pressure value in both high and low directions to obtain the performance fluctuation range.

[0138] It should be noted that, in the embodiment of the present application, the process of determining the performance fluctuation range of the storage device is as follows:

[0139] 1. Pressure Adjustment Strategy

[0140] 1. Reference point setting:

[0141] The target pressure value corresponding to the performance bottleneck point is used as the starting benchmark (denoted as P0). This value is determined through preliminary testing and represents the critical pressure level at which the storage device reaches maximum stable performance.

[0142] 2. Bidirectional step adjustment:

[0143] Positive pressure: Starting from P0, gradually increase the pressure according to the preset step size (such as 5%), and record the performance data of pressure points such as P0+ΔP, P0+2ΔP, etc.;

[0144] Reverse pressure reduction: Start from P0 synchronously, gradually reduce the pressure with the same step length, and record the performance data of pressure points such as P0-ΔP, P0-2ΔP, etc.

[0145] Step size selection principles: The step size is typically set to 1%-5% of the baseline pressure and needs to be adjusted based on the device type (e.g., a smaller step size can be used for an all-flash array). Alternatively, the step size set during the pressure-performance matching algorithm can be used.

[0146] 2. Data Collection and Processing

[0147] 1. Performance data collection:

[0148] At each pressure adjustment point:

[0149] (1) Continuously collect performance indicators (such as IOPS and latency) for at least three complete load cycles;

[0150] (2) Ensure data coverage with a mix of read and write operations (typical ratio is 70% read / 30% write).

[0151] 2. Sliding average processing:

[0152] A sliding average algorithm with a window width of 5 sampling points is used

[0153] (1) Calculation method: Take the arithmetic mean of 5 consecutive sampling values ​​as the representative value of the pressure point;

[0154] (2) Special processing: When a data mutation is detected (difference between adjacent points > 15%), the window is automatically expanded to 10 sampling points.

[0155] 3. Performance degradation analysis

[0156] 1. Attenuation value calculation:

[0157] Taking the performance value of the reference pressure P0 as the benchmark (denoted as Q0), the attenuation value of each pressure point = (Q0-performance value of the current point) / Q0×100%, and the attenuation curve is calculated for the forward and reverse pressure sequences respectively.

[0158] 2. Threshold determination mechanism:

[0159] The default threshold is usually set to 15% (can be adjusted according to device specifications)

[0160] (1) Stop condition: the attenuation value of two consecutive pressure points exceeds the threshold;

[0161] (2) Boundary determination: The pressure value before the first threshold value is taken as the boundary of the fluctuation range.

[0162] 4. Determination of Fluctuation Range

[0163] 1. Positive boundary:

[0164] Record the last high pressure point with a decay value ≤ 15% (e.g. P0+3ΔP);

[0165] 2. Reverse boundary:

[0166] Record the last low pressure point with a decay value ≤ 15% (e.g. P0-2ΔP);

[0167] 3. Final scope:

[0168] The performance fluctuation range is determined to be [P0-2ΔP, P0+3ΔP], and the performance degradation within this interval is controlled within the allowable range.

[0169] Therefore, the embodiment of the present application can simultaneously detect performance changes under overload and underload conditions through bidirectional pressure scanning, which is more comprehensive than unidirectional testing; in addition, the sliding average processing operation of the embodiment of the present application can effectively eliminate instantaneous fluctuations caused by system background tasks, etc. (such as performance jitter caused by garbage collection), and has good anti-interference ability; then, the threshold judgment mechanism of the embodiment of the present application effectively takes into account the test efficiency and accuracy requirements, avoiding excessive testing.

[0170] Optionally, in one embodiment of the present application, based on the target pressure value, target performance indicator value, performance fluctuation range and steady-state achievement time, a performance curve, a pressure-performance scatter plot and a bottleneck analysis heat map corresponding to the target storage device are drawn, including: constructing a two-dimensional matrix of pressure-performance indicators based on the target pressure value and the target performance indicator value, and dividing the two-dimensional matrix into multiple pressure intervals, and calculating the performance fluctuation characteristics of different pressure intervals, wherein the fluctuation characteristics include the degree of discreteness and skewness characteristics of the performance indicators; calculating the statistical distribution parameters of the performance indicators in each pressure interval in the multiple pressure zones, wherein the statistical distribution parameters include standard deviation, skewness and kurtosis; performing a visual color coding operation based on the degree of discreteness, skewness characteristics and statistical distribution parameter values ​​to generate a bottleneck analysis heat map.

[0171] Specifically, the process of generating a storage device bottleneck analysis heat map in an embodiment of the present application is as follows:

[0172] 1. Basic data preparation stage

[0173] 1. Key parameter acquisition:

[0174] (1) Obtain the target pressure value (denoted as P*) corresponding to the performance bottleneck point. This value represents the pressure level when the storage device reaches the optimal performance state;

[0175] (2) Record the corresponding target performance indicator value (denoted as Q*), such as IOPS, latency and other core indicators;

[0176] (3) Obtain the performance fluctuation range and steady-state reaching time data of the storage device.

[0177] 2. Two-dimensional matrix construction:

[0178] (1) Establish a two-dimensional data matrix of pressure and performance index, with the horizontal axis representing the pressure value (from 0 to 120% P*) and the vertical axis representing the performance index value;

[0179] (2) Divide the pressure axis into 24 pressure intervals at intervals of 5%P (e.g., 0-5%P, 5-10%P*).

[0180] 2. Pressure range analysis stage

[0181] 1. Calculation of performance fluctuation characteristics:

[0182] (1) In each pressure range:

[0183] Degree of dispersion: Calculate the standard deviation of the performance index value to reflect the degree of data dispersion;

[0184] Skewness characteristics: The skewness coefficient is calculated using the moment method to determine the data distribution shape (left skewness / right skewness);

[0185] Kurtosis: Quantifies the sharpness of the data distribution.

[0186] For example, when the load is in the 80-85% P* range, the standard deviation is 850 IOPS (reflecting the performance fluctuation range), the skewness coefficient is 1.2 (indicating a right-skewed distribution with abnormally high values), and the kurtosis is 4.8 (the distribution shape is sharper than a normal distribution).

[0187] 3. Visual Coding Stage

[0188] 1. Heat map coloring rules:

[0189] (1) Hue selection: Gradual change from blue to red according to the pressure range (low pressure blue changes to high pressure red);

[0190] (2) Brightness control: Use standard deviation value mapping, the greater the fluctuation, the darker the color;

[0191] (3) Saturation adjustment: According to the skewness coefficient, positive skewness increases the red saturation, and negative skewness increases the blue saturation.

[0192] 2. Auxiliary marking method:

[0193] (1) Use contour lines to mark key resource utilization thresholds (e.g., CPU > 75%).

[0194] (2) Add special marks at the mutation points (e.g., add a lightning icon where the performance drops sharply when the pressure is > 90% P*).

[0195] 4. Chart Generation Stage

[0196] 1. Performance curve drawing:

[0197] (1) Draw the main curve with the pressure value as the horizontal axis and the median of the performance index as the vertical axis;

[0198] (2) The upper and lower quartiles are superimposed to form a band-like area to intuitively display the fluctuation range.

[0199] 2. Scatter plot generation:

[0200] (1) Each sampling point is plotted according to the actual pressure-performance value;

[0201] (2) The size of the point corresponds to the comprehensive resource utilization rate at the sampling moment;

[0202] (3) The shape of the dot distinguishes the type of read and write operation (if it is a circle, it is a read, if it is a square, it is a write).

[0203] 3. Heatmap synthesis:

[0204] (1) Superimpose the colored pressure interval matrix with the auxiliary markers;

[0205] (2) Add a legend to illustrate the correspondence between colors and statistical parameters;

[0206] (3) Add text labels to key areas (such as "high volatility risk area").

[0207] Therefore, the embodiments of the present application can simultaneously display the pressure-performance relationship, data distribution characteristics and resource correlation through multi-dimensional analysis, and can encode multiple types of information through the three elements of color (hue, brightness, and saturation), so as to intuitively identify performance inflection points (such as sudden changes in color to dark red areas) and quickly locate resource competition hotspots.

[0208] To sum up, the embodiments of the present application can effectively reduce manual intervention, shorten the test cycle by more than 50%, reduce labor costs by 60%, and the performance curve volatility is less than or equal to 5%, which is better than the 15%-20% of the traditional method and has higher accuracy; in addition, the embodiments of the present application can also support distributed storage and multi-dimensional evaluation in cloud-native environments, and can reduce the risk of hardware loss through abnormal protection mechanisms, greatly improving the scalability and reliability of storage device performance evaluation strategies.

[0209] The following describes the execution logic of the storage device performance evaluation method based on dynamic pressure regulation of the present application in conjunction with the accompanying drawings.

[0210] Figure 2 This is a schematic diagram of the execution logic of the storage device performance evaluation method based on dynamic pressure regulation of this application. Figure 2 As shown, the execution process of the storage device performance evaluation method based on dynamic pressure regulation of the present application is as follows:

[0211] S201: initializing the storage device;

[0212] S202: Pressurizing the storage device;

[0213] S203: The pressure of the storage device is maintained for a preset time period;

[0214] S204: Determine whether the storage device has reached a steady-state state after the preset pressurization time. If so, proceed to S205; otherwise, proceed to S203.

[0215] S205: Record the current pressure value and the current performance index value;

[0216] S206: Determine whether the performance of the storage device meets the standard. If so, proceed to S207. If the current performance index value is less than 95% of the expected storage performance index value or greater than 105% of the expected storage performance index value, it does not meet the standard, and proceed to S208.

[0217] S207: Perform pressure and performance matching algorithm determination on the storage device, and proceed to S209 and S2010 respectively;

[0218] S208: Dynamic pressure adjustment;

[0219] S209: Output pressure value and performance index value;

[0220] S2010: Pressure value adjustment, pressurization is performed using the adjusted pressure value, and performance changes of the storage device are determined. If the performance decreases, the process goes to S208; if the performance remains the same, S2010 is executed again.

[0221] In addition, the present application can also construct a corresponding storage device performance evaluation system based on dynamic pressure regulation based on the execution logic of the above-mentioned storage device performance evaluation method based on dynamic pressure regulation. The storage device performance evaluation system based on dynamic pressure regulation mainly includes a pressure generation module, a performance monitoring module, a data analysis module, a pressure regulation module, and a control module.

[0222] The pressure generation module is used to generate a pressure thread to pressurize the storage device;

[0223] Performance monitoring module, used to monitor the performance indicators of storage devices and record performance data;

[0224] The data analysis module is used to analyze the performance data, calculate the variance value of the performance data, determine whether the performance of the storage device has reached a steady state, and obtain the corresponding analysis results;

[0225] Pressure regulation module, used to dynamically adjust the pressure value based on performance data and analysis results;

[0226] The control module is used to coordinate the work of each module and control the entire evaluation process.

[0227] Therefore, the storage device performance evaluation system based on dynamic pressure regulation of the present application improves the flexibility and scalability of the system through modular design and coordinated control mechanism, facilitating subsequent functional upgrades and optimizations.

[0228] Through the description of the above implementation methods, those skilled in the art can clearly understand that the method according to the above embodiment can be implemented by means of software plus the necessary general hardware platform, and of course it can also be implemented by hardware, but in many cases the former is a better implementation method.

[0229] An embodiment of the present application also provides a storage device performance evaluation apparatus based on dynamic pressure regulation.

[0230] like Figure 3As shown, the storage device performance evaluation apparatus 10 based on dynamic pressure regulation includes: a storage device pressurization module 100 , a performance steady-state judgment module 200 , a pressure-performance matching module 300 and a test report generation module 400 .

[0231] Among them, the storage device pressurization module 100 is used to pressurize the target storage device based on a preset initial pressure value, and after the target storage device is pressurized for a preset time, collect the performance indicator value of the target storage device every preset period to generate multiple groups of performance indicator values ​​of the target storage device.

[0232] The performance steady-state judgment module 200 is used to calculate the variance value of each group of performance indicator values ​​in multiple groups of performance indicator values, and judge whether the target storage device meets the preset performance steady-state requirements based on the variance value. If the target storage device meets the performance steady-state requirements, the current pressure value and current performance indicator value of the target storage device are obtained to judge whether the current performance indicator value is within the preset expected range.

[0233] The pressure-performance matching module 300 is used to increase or decrease the current pressure value based on a preset step value if the current performance index value is not within the expected range, and re-pressurize the target storage device using the increased or decreased current pressure value; if the current performance index value is within the expected range, iteratively reduce the current pressure value of the target storage device, and match the performance index value corresponding to the reduced current pressure value in each iteration process, so as to determine the performance bottleneck point of the target storage device based on the matched performance index value, and generate a performance evaluation test result of the target storage device according to the target pressure value and target performance index value corresponding to the performance bottleneck point.

[0234] Optionally, in one embodiment of the present application, the performance steady-state judgment module 200 includes: a first judgment unit, a collection unit, and a first determination unit.

[0235] The first judgment unit is configured to judge whether the target storage device meets the performance steady-state requirement based on multiple groups of performance indicator values.

[0236] The collection unit is used to pressurize the target storage device for a preset time period according to the initial pressure value if the target storage device does not meet the performance steady-state requirement, so as to re-collect multiple sets of new performance indicator values ​​corresponding to the target storage device after the preset time period.

[0237] The first determining unit is configured to determine a current pressure value and a current performance index value of the target storage device if the target storage device does not meet the performance steady-state requirement.

[0238] Optionally, in one embodiment of the present application, the pressure performance matching module 300 includes: a first pressurization unit and a pressure reduction unit.

[0239] Among them, the first boosting unit is used to increase the current pressure value based on a preset step value and a dynamic pressure adjustment mechanism if the current performance indicator value of the target storage device is less than the minimum value of the expected interval, so as to use the increased current pressure value to pressurize the target storage device for a preset time period.

[0240] The decompression unit is used to reduce the current pressure value based on the step value and the dynamic pressure adjustment mechanism if the current performance indicator value of the target storage device is greater than the maximum value of the expected interval, so as to use the reduced current pressure value to pressurize the target storage device for a preset time period.

[0241] Optionally, in one embodiment of the present application, the boost unit includes: a calculation subunit and a construction subunit.

[0242] The calculation subunit is used to calculate the performance indicator difference between the current performance indicator value and the preset performance indicator expected value.

[0243] The subunit is constructed to construct a dynamic pressure regulation expression based on the performance index difference, the current pressure value and the preset regulation coefficient, so as to determine the dynamic pressure regulation mechanism according to the dynamic pressure regulation expression.

[0244] Optionally, in one embodiment of the present application, the pressure performance matching module 300 further includes: a first halving unit, a second judging unit, a second pressurizing unit, a second halving unit, and an iterating unit.

[0245] The first halving unit is configured to reduce the current pressure value to obtain a corresponding pressure verification value, and pressurize the target storage device for a preset time period using the pressure verification value to collect a performance indicator verification value after pressurization for the preset time period.

[0246] The second judgment unit is used to judge whether the performance indicator verification value is less than the minimum value of the expected interval.

[0247] The second boosting unit is used to increase the pressure verification value based on a preset dynamic pressure adjustment mechanism and step value when the performance indicator verification value is less than the minimum value of the expected interval, so as to pressurize the target storage device for a preset time period through the increased pressure verification value.

[0248] The second halving unit is used to continue to reduce the pressure verification value when the performance indicator verification value is within the expected range to obtain a new pressure verification value, and re-pressurize the target storage device for a preset time period using the new pressure verification value to obtain a new performance indicator verification value after pressurization for the preset time period.

[0249] The iterative unit is used to iteratively perform pressure application and performance indicator value collection operations on the target storage device based on the new pressure verification value and the new performance indicator verification value until the collected performance indicator value is less than the minimum value of the expected interval, so as to determine the performance bottleneck point of the target storage device.

[0250] Optionally, in one embodiment of the present application, the pressure performance matching module 300 further includes: a second determining unit, a drawing unit, and a generating unit.

[0251] The second determining unit is configured to determine the performance fluctuation range and steady-state reaching time corresponding to the target storage device based on the target pressure value and the target performance indicator value.

[0252] The drawing unit is used to draw the performance curve, pressure-performance scatter plot and bottleneck analysis heat map corresponding to the target storage device based on the target pressure value, target performance indicator value, performance fluctuation range and steady-state achievement time.

[0253] The generation unit is used to generate performance evaluation test results based on the performance fluctuation range, steady-state achievement time, performance curve, pressure-performance scatter plot and bottleneck analysis heat map.

[0254] Optionally, in one embodiment of the present application, the second determining unit includes: a bidirectional scanning subunit and a stopping subunit.

[0255] Among them, the bidirectional scanning subunit is used to start from the target pressure value, and based on the preset step value, gradually adjust the pressure value in at least one target direction, and obtain the performance data corresponding to each adjusted pressure value, and eliminate the instantaneous fluctuation interference of the performance data to obtain the performance attenuation value corresponding to each adjusted pressure value.

[0256] The stop subunit is used to determine whether the performance degradation value is greater than a preset threshold value. When the performance degradation value is greater than the preset threshold value, the pressure value is stopped from being gradually adjusted in both high and low directions to obtain a performance fluctuation range.

[0257] Optionally, in one embodiment of the present application, the rendering unit includes: a matrix construction subunit, a parameter calculation subunit and a color coding subunit.

[0258] Among them, the matrix construction subunit is used to construct a two-dimensional matrix of pressure-performance indicators based on the target pressure value and the target performance indicator value, divide the two-dimensional matrix into multiple pressure intervals, and calculate the performance fluctuation characteristics of different pressure intervals, wherein the fluctuation characteristics include the discrete degree and skewness characteristics of the performance indicators.

[0259] The parameter calculation subunit is used to calculate the statistical distribution parameters of the performance index in each pressure interval of the multiple pressure zones, wherein the statistical distribution parameters include standard deviation, skewness and kurtosis.

[0260] The color coding subunit is used to perform visual color coding operations based on the degree of dispersion, skewness characteristics, and statistical distribution parameter values ​​to generate bottleneck analysis heat maps.

[0261] For descriptions of features in the embodiments corresponding to the storage device performance evaluation apparatus based on dynamic pressure regulation, reference may be made to the relevant descriptions of the embodiments corresponding to the storage device performance evaluation method based on dynamic pressure regulation, which will not be detailed here.

[0262] An embodiment of the present application also provides an electronic device, including a memory and a processor, wherein the memory stores a computer program, and the processor is configured to run the computer program to execute the steps in any of the above-mentioned embodiments of the storage device performance evaluation method based on dynamic pressure regulation.

[0263] An embodiment of the present application also provides a non-volatile computer-readable storage medium, which stores a computer program, wherein the computer program is configured to execute the steps of any of the above-mentioned storage device performance evaluation method embodiments based on dynamic pressure regulation when running.

[0264] In an exemplary embodiment, the non-volatile computer-readable storage medium may include, but is not limited to, various media that can store computer programs, such as a USB flash drive, a read-only memory (ROM), a random access memory (RAM), a mobile hard disk, a magnetic disk, or an optical disk.

[0265] An embodiment of the present application further provides a computer program product, which includes a computer program. When the computer program is executed by a processor, the steps in any of the above-mentioned storage device performance evaluation method embodiments based on dynamic pressure regulation are implemented.

[0266] An embodiment of the present application also provides another computer program product, including a non-volatile computer-readable storage medium, the non-volatile computer-readable storage medium storing a computer program, and when the computer program is executed by a processor, implementing the steps in any of the above-mentioned embodiments of the storage device performance evaluation method based on dynamic pressure regulation.

[0267] Professionals may further appreciate that the units and algorithm steps of each example described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, computer software, or a combination of the two. In order to clearly illustrate the interchangeability of hardware and software, the above description has generally described the components and steps of each example according to their functions. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professionals and technicians may use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.

[0268] The above is a detailed introduction to the storage device performance evaluation method, apparatus, equipment and medium based on dynamic pressure regulation provided by the present application. This article uses specific examples to illustrate the principles and implementation methods of the present application. The description of the above embodiments is only used to help understand the method of the present application and its core ideas. It should be pointed out that for ordinary technicians in this technical field, without departing from the principles of the present application, several improvements and modifications can be made to the present application, and these improvements and modifications also fall within the scope of protection of the claims of the present application.

Claims

1. A storage device performance evaluation method based on dynamic pressure regulation, characterized in that: The following steps are involved: Pressurizing the target storage device based on a preset initial pressure value, and collecting performance indicator values ​​of the target storage device at preset intervals after the target storage device has been pressurized for a preset period of time to generate multiple sets of performance indicator values ​​of the target storage device; Calculating a variance value of each group of performance indicator values ​​in the multiple groups of performance indicator values, and determining whether the target storage device meets a preset steady-state performance requirement based on the variance value, wherein if the target storage device meets the steady-state performance requirement, obtaining a current pressure value and a current performance indicator value of the target storage device, and determining whether the current performance indicator value is within a preset expected range; If the current performance index value is not within the expected range, the current pressure value is increased or decreased based on a preset step value, and the target storage device is re-pressurized using the increased or decreased current pressure value; if the current performance index value is within the expected range, the current pressure value of the target storage device is iteratively decreased, and the performance index value corresponding to the reduced current pressure value in each iteration is matched, so as to determine the performance bottleneck point of the target storage device based on the matched performance index value, and generate a performance evaluation test result of the target storage device according to the target pressure value and target performance index value corresponding to the performance bottleneck point; If the current performance indicator value is within the expected range, iteratively reducing the current pressure value of the target storage device and matching the performance indicator value corresponding to the reduced current pressure value in each iteration to determine the performance bottleneck of the target storage device based on the matched performance indicator value includes: Lowering the current pressure value to obtain a corresponding pressure verification value, and applying pressure to the target storage device for the preset duration using the pressure verification value to collect a performance indicator verification value after applying pressure for the preset duration; Determining whether the performance indicator verification value is less than the minimum value of the expected interval; When the performance indicator verification value is less than the minimum value of the expected interval, based on a preset dynamic pressure adjustment mechanism and the step value, increasing the pressure verification value, so as to pressurize the target storage device for the preset time period using the increased pressure verification value; When the performance indicator verification value is within the expected range, continue to reduce the pressure verification value to obtain a new pressure verification value, and re-pressurize the target storage device for the preset time period using the new pressure verification value to obtain a new performance indicator verification value after pressurization for the preset time period; Based on the new pressure verification value and the new performance indicator verification value, the target storage device is iteratively pressurized and the performance indicator value is collected until the collected performance indicator value is less than the minimum value of the expected interval, so as to determine the performance bottleneck point of the target storage device.

2. The storage device performance evaluation method based on dynamic pressure regulation according to claim 1, characterized in that: The determining, based on the variance value, whether the target storage device meets a preset steady-state performance requirement, wherein if the target storage device meets the steady-state performance requirement, obtaining a current pressure value and a current performance index value of the target storage device includes: Based on the multiple groups of performance indicator values, determining whether the target storage device meets the performance steady-state requirement; If the target storage device does not meet the steady-state performance requirement, pressurizing the target storage device for the preset time period according to the initial pressure value to re-collect multiple sets of new performance indicator values ​​corresponding to the target storage device after the preset time period; If the target storage device meets the performance steady-state requirement, the current pressure value and the current performance indicator value of the target storage device are determined.

3. The storage device performance evaluation method based on dynamic pressure regulation according to claim 2, characterized in that: If the current performance indicator value is not within the expected range, increasing or decreasing the current pressure value based on a preset step value, and re-pressurizing the target storage device using the increased or decreased current pressure value, including: If the current performance indicator value of the target storage device is less than the minimum value of the expected interval, increasing the current pressure value based on a preset step value and a dynamic pressure adjustment mechanism, so as to pressurize the target storage device for the preset time period using the increased current pressure value; If the current performance index value of the target storage device is greater than the maximum value of the expected interval, the current pressure value is reduced based on the step value and the dynamic pressure adjustment mechanism to pressurize the target storage device for the preset time period using the reduced current pressure value.

4. The storage device performance evaluation method based on dynamic pressure regulation according to claim 3, characterized in that: The increasing the current pressure value based on a preset step value and a dynamic pressure adjustment mechanism includes: Calculating a performance indicator difference between the current performance indicator value and a preset performance indicator expected value; A dynamic pressure regulation expression is constructed based on the performance indicator difference, the current pressure value and a preset regulation coefficient, so as to determine the dynamic pressure regulation mechanism according to the dynamic pressure regulation expression.

5. The storage device performance evaluation method based on dynamic pressure regulation according to claim 1, characterized in that: Generating the performance evaluation test result of the target storage device according to the target pressure value and the target performance index value corresponding to the performance bottleneck point includes: Determining a performance fluctuation range and a steady-state reaching time corresponding to the target storage device based on the target pressure value and the target performance indicator value; Based on the target pressure value, the target performance indicator value, the performance fluctuation range, and the steady-state reaching time, draw a performance curve, a pressure-performance scatter plot, and a bottleneck analysis heat map corresponding to the target storage device; The performance evaluation test result is generated according to the performance fluctuation range, the steady-state reaching time, the performance curve, the pressure-performance scatter plot, and the bottleneck analysis heat map.

6. The storage device performance evaluation method based on dynamic pressure regulation according to claim 5, characterized in that: The determining, based on the target pressure value and the target performance indicator value, a performance fluctuation range and a steady-state reaching time corresponding to the target storage device includes: Starting from the target pressure value, the pressure value is gradually adjusted in at least one target direction based on a preset step value, and performance data corresponding to each adjusted pressure value is obtained, and instantaneous fluctuation interference of the performance data is eliminated to obtain a performance degradation value corresponding to each adjusted pressure value; Determine whether the performance degradation value is greater than a preset threshold, wherein when the performance degradation value is greater than the preset threshold, stop gradually adjusting the pressure value in both high and low directions to obtain the performance fluctuation range.

7. The storage device performance evaluation method based on dynamic pressure regulation according to claim 5, characterized in that: Drawing a performance curve, a pressure-performance scatter plot, and a bottleneck analysis heat map corresponding to the target storage device based on the target pressure value, the target performance indicator value, the performance fluctuation range, and the steady-state achievement time includes: Based on the target pressure value and the target performance index value, constructing a two-dimensional matrix of pressure and performance index, dividing the two-dimensional matrix into a plurality of pressure intervals, and calculating performance fluctuation characteristics of different pressure intervals, wherein the fluctuation characteristics include the degree of dispersion and skewness characteristics of the performance index; Calculating statistical distribution parameters of the performance indicator in each pressure interval of the plurality of pressure zones, wherein the statistical distribution parameters include standard deviation, skewness, and kurtosis; A visual color coding operation is performed based on the degree of dispersion, the skewness characteristics, and the statistical distribution parameter value to generate the bottleneck analysis heat map.

8. An electronic device, characterized in that: include: memory for storing computer programs; A processor is configured to implement the steps of the storage device performance evaluation method based on dynamic pressure regulation as claimed in any one of claims 1 to 7 when executing the computer program.

9. A non-volatile computer-readable storage medium, characterized in that: The non-volatile computer-readable storage medium stores a computer program, wherein when the computer program is executed by a processor, the steps of the storage device performance evaluation method based on dynamic pressure regulation according to any one of claims 1 to 7 are implemented.

Images