Storage performance analysis method and device, computer device and storage medium

Abstract

The application provides a storage performance analysis method and device, computer equipment and a storage medium, and relates to the technical field of computers. The application method realizes accurate perception from device performance to application performance by collecting I / O performance data generated by business applications. By calculating data conversion ratios, the performance differences between different I / O operation types are quantified, and evaluation deviations caused by the diversity of I / O operation types are avoided. By calculating the benchmark I / O total amount of the business application under standard I / O conditions, performance evaluation is standardized, and the comparability and consistency of the evaluation results are improved. By calculating the performance density, the efficiency of the storage system in meeting actual business requirements is more accurately evaluated. According to the performance density, the performance level is determined, accurate grading evaluation of the performance state of the storage system is realized, and the performance of the storage system in application scenarios such as financial risk control applications and medical image storage can be accurately reflected, thereby providing strong support for resource planning and performance optimization.

Description

Technical Field

[0001] This application relates to the field of computer technology, and in particular to a storage performance analysis method, apparatus, computer equipment, and storage medium. Background Technology

[0002] With the rapid development of cloud computing, big data, and artificial intelligence technologies, storage systems, as the core of data infrastructure, face severe challenges in performance evaluation and billing mechanisms. Currently, the performance evaluation of mainstream storage systems mainly relies on three performance indicators: IOPS (Input / Output Operations Per Second), throughput, and latency. However, these indicators only reflect the physical performance of storage devices and cannot be directly correlated with the actual needs of business applications.

[0003] As digital transformation accelerates across industries, the demands on storage system performance are constantly increasing. The financial industry needs to support high-concurrency scenarios such as AI training and real-time risk control, while the healthcare industry needs to handle petabyte-scale image data storage and the concurrent access needs of thousands of users.

[0004] Current technologies for I / O performance analysis often focus on a single performance metric, failing to correlate the performance of underlying storage devices with the actual I / O behavior of upper-layer business applications. This results in biased performance evaluations that fail to fully reflect the true performance status of the storage system, hindering refined management and cost optimization of storage resources. For example, a database application might generate numerous small-block random reads and writes, while a video transcoding system primarily uses large-block sequential reads and writes. Traditional evaluation models cannot differentiate between these differences, leading to inaccurate performance matching. Users struggle to intuitively determine "which storage is suitable for my business," causing a surge in communication costs and even resource mismatches such as "over-configuration for under-utilization" or "under-configuration exceeding limits."

[0005] Existing storage performance evaluation technologies can no longer meet the needs of these emerging business scenarios. How to deeply integrate the underlying storage performance with the upper-layer business requirements, achieve refined evaluation of storage performance, and improve the accuracy of storage performance evaluation has become an urgent technical problem to be solved. Summary of the Invention

[0006] This application provides a storage performance analysis method, apparatus, computer equipment, and storage medium, aiming to solve the technical problem that the existing technology cannot dynamically reflect the actual business load performance requirements, resulting in low accuracy of storage performance evaluation.

[0007] In a first aspect, this application provides a storage performance analysis method, which includes the following steps: Collect I / O performance data generated by at least one current I / O operation corresponding to the business application in the storage system; Calculate the data conversion ratio corresponding to the current I / O operation based on the baseline performance data and the I / O performance data corresponding to the current I / O operation; Based on the data conversion ratio corresponding to the current I / O operation and the I / O operation quantity (IOPS) of the current I / O operation, calculate the baseline total I / O of the service application under standard I / O conditions; Calculate the performance density of the storage system based on the number of beneficiaries and the total baseline I / O; Based on the performance density, the performance level of the storage system is determined to evaluate the performance of the storage system.

[0008] Secondly, this application also provides a storage performance analysis device, the storage performance analysis device comprising: The data acquisition module is used to collect I / O performance data generated by at least one current I / O operation corresponding to the business application in the storage system; The conversion ratio calculation module is used to calculate the data conversion ratio corresponding to the current I / O operation based on the baseline performance data and the I / O performance data corresponding to the current I / O operation; The baseline I / O total calculation module is used to calculate the baseline I / O total of the service application under standard I / O conditions based on the data conversion ratio corresponding to the current I / O operation and the I / O operation quantity (IOPS) of the current I / O operation. The performance density calculation module is used to calculate the performance density of the storage system based on the number of benefits and the total baseline I / O. A storage performance evaluation module is used to determine the performance level of the storage system based on the performance density, so as to evaluate the performance of the storage system.

[0009] Thirdly, this application also provides a computer device, the computer device including a processor, a memory, and a computer program stored in the memory and executable by the processor, wherein when the computer program is executed by the processor, it implements the steps of the storage performance analysis method described above.

[0010] Fourthly, this application also provides a computer-readable storage medium storing a computer program, wherein when the computer program is executed by a processor, it implements the steps of the storage performance analysis method described above.

[0011] This application provides a storage performance analysis method, apparatus, computer device, and storage medium. The method collects I / O performance data generated by business applications, achieving accurate perception from device performance to application performance. This accurately reflects the storage performance requirements of business applications in the actual operating environment, ensuring the real-time nature and relevance of the data. By calculating the data conversion ratio, the collected actual performance data is compared with predefined benchmark performance data. Complex and diverse non-benchmark I / O operation types are normalized into standard benchmark I / O units, quantifying the performance differences between different I / O operation types and avoiding evaluation bias caused by the diversity of I / O operation types. By combining the data conversion ratio with the actual I / O operation volume (IOPS), the benchmark total I / O of business applications under standard I / O conditions is calculated, achieving standardization of performance evaluation. This provides a unified quantitative indicator for the overall performance evaluation of the storage system, allowing the storage performance of different business applications to be compared under the same benchmark conditions, improving the comparability and consistency of evaluation results. By calculating performance density, a correlation is established between the benchmark total I / O and the benefit number, considering not only the performance output of the storage system but also the actual resource usage. By calculating performance density, the efficiency of a storage system in meeting actual business needs can be more accurately assessed, thus avoiding resource waste or inadequacy that may result from solely relying on performance metrics. Determining performance levels based on performance density enables precise grading and assessment of the storage system's performance status. This not only allows users to quickly understand the storage system's performance level but also enables the development of corresponding optimization strategies based on different performance levels. Performance density-based tiered assessment more accurately reflects the storage system's performance in real-world business scenarios, thereby providing strong support for resource planning and performance optimization. Attached Figure Description

[0012] To more clearly illustrate the technical solutions of the embodiments of this application, the drawings used in the description of the embodiments will be briefly introduced below. Obviously, the drawings described below are some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0013] Figure 1 This is a schematic diagram of an application environment for a storage performance analysis method according to an embodiment of the present invention; Figure 2 A flowchart illustrating an embodiment of a storage performance analysis method provided in this application; Figure 3 This is a schematic diagram of the structure of an embodiment of a storage performance analysis device provided in this application. Figure 4 This is a schematic block diagram of the structure of a computer device provided in an embodiment of this application.

[0014] The realization of the purpose, functional features and advantages of this application will be further explained in conjunction with the embodiments and with reference to the accompanying drawings. Detailed Implementation

[0015] 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, 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.

[0016] The flowchart shown in the attached diagram is for illustrative purposes only and does not necessarily include all content and operations / steps, nor does it necessarily have to be performed in the order described. For example, some operations / steps can be broken down, combined, or partially merged, so the actual execution order may change depending on the actual situation.

[0017] The following detailed description of some embodiments of this application is provided in conjunction with the accompanying drawings. Unless otherwise specified, the following embodiments and features can be combined with each other.

[0018] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of the present invention. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0019] The storage performance analysis method provided in this embodiment of the invention can be applied to, for example... Figure 1 In this application environment, the client communicates with the server via a network. When the server receives a storage performance evaluation request from the client, it can collect I / O performance data generated by at least one current I / O operation corresponding to the business application in the storage system; calculate the data conversion ratio corresponding to the current I / O operation based on the baseline performance data and the I / O performance data; calculate the baseline total I / O of the business application under standard I / O conditions based on the data conversion ratio and the I / O operation quantity (IOPS) of the current I / O operation; calculate the performance density of the storage system based on the benefit number and the baseline total I / O; and determine the performance level of the storage system based on the performance density to evaluate the performance of the storage system.

[0020] This invention addresses the technical problem in existing technologies where, in application scenarios such as financial risk control and medical image data storage, the inability to dynamically reflect actual business load performance requirements leads to low accuracy in storage performance evaluation. By collecting I / O performance data generated by business applications, it achieves precise perception from device performance to application performance, accurately reflecting the storage performance requirements of business applications in the actual operating environment and ensuring data real-time performance and relevance. By calculating the data conversion ratio, the collected actual performance data is compared with predefined benchmark performance data. This normalizes the complex and diverse non-benchmark I / O operation types into standard benchmark I / O units, quantifying the performance differences between different I / O operation types and avoiding evaluation bias caused by the diversity of I / O operation types. By combining the data conversion ratio with the actual I / O operation volume (IOPS), the benchmark total I / O of the business application under standard I / O conditions is calculated, achieving standardization of performance evaluation. This provides a unified quantitative indicator for the overall performance evaluation of the storage system, allowing the storage performance of different business applications to be compared under the same benchmark conditions, improving the comparability and consistency of evaluation results. By calculating performance density and establishing a correlation between baseline I / O volume and benefit, this approach considers not only the storage system's performance output but also actual resource usage. Performance density calculations allow for a more accurate assessment of the storage system's efficiency in meeting real-world business needs, avoiding resource waste or inadequacy that might result from solely relying on performance metrics. Determining performance levels based on performance density enables precise grading and evaluation of the storage system's performance status. This not only allows users to quickly understand the storage system's performance level but also enables the development of corresponding optimization strategies based on different performance levels. Performance density-based tiered evaluation more accurately reflects the storage system's performance in real-world business scenarios, thus providing strong support for resource planning and performance optimization.

[0021] The client can be, but is not limited to, various personal computers, laptops, smartphones, tablets, and portable wearable devices. The server can be implemented using a standalone server or a server cluster consisting of multiple servers. The invention will now be described in detail through specific embodiments.

[0022] Please refer to Figure 2 , Figure 2 This is a flowchart illustrating an embodiment of a storage performance analysis method provided in this application.

[0023] like Figure 2 As shown, the storage performance analysis method includes steps S101 to S104.

[0024] S101. Collect I / O performance data generated by at least one current I / O operation corresponding to the business application in the storage system.

[0025] The collection of I / O performance data is the foundational data collection stage of the SPPS (Storage Performance Profiling System) algorithm model. It aims to capture the real I / O behavior characteristics generated by business applications from the storage system, and then transform complex low-level I / O operations into quantifiable performance parameters, providing data support for subsequent I / O behavior modeling and performance evaluation.

[0026] The business applications running in the storage system can be database services (such as Oracle, MySQL), file services, virtualization platforms, or specific business software. I / O monitoring of the storage volume or logical unit number (LUN) used by the business application can identify and track all I / O requests entering and leaving that storage resource.

[0027] The current I / O operation includes at least one I / O operation type among read operation, write operation, sequential operation, and random operation; the I / O performance data includes at least one of block size, throughput, and cache hit rate.

[0028] During I / O performance data collection, different collection periods can be selected based on business needs and storage system load. For example, for high-load real-time business systems, shorter collection periods (such as per second or per minute) can be set to capture performance changes in a timely manner; for low-load systems, longer collection periods (such as per hour) can be set. Furthermore, the granularity of the collected data can be adjusted according to analytical requirements. For instance, detailed data for each I / O operation can be collected, including operation type, block size, latency, etc.; or summary data, such as average throughput and cache hit rate per minute, can be collected.

[0029] Within a preset sampling period (e.g., 5 minutes), I / O operations are sampled in real time, and the following key I / O performance data are extracted: I / O operation volume (IOPS), throughput, cache hit rate, I / O latency, and I / O queue depth, etc.

[0030] I / O operations per second (IOPS) refers to the number of input / output operations per second, categorized separately for read and write operations (read IOPS and write IOPS). Throughput refers to the amount of data transferred per second, also categorized separately for read and write operations. Cache hit rate is the percentage of I / O requests that are successfully cached by the storage controller; this metric reflects the efficiency of the storage system in utilizing caching to accelerate I / O access, and is categorized separately for read and write operations. I / O latency is the time from when an I / O request is sent to when a response is received, used to evaluate the storage system's response speed. I / O queue depth is the average number of I / O requests waiting to be processed within the sampling period, reflecting the storage system's concurrent processing capacity.

[0031] The collected I / O operations are classified in multiple dimensions. The I / O operations generated by business applications are divided into current I / O operations such as read operations, write operations, sequential operations, and random operations according to their functions and access modes, so as to build a refined I / O behavior profile.

[0032] Based on read / write type, operations can be divided into read operations and write operations. Read operations refer to the process of retrieving data from a storage device. The performance of read operations directly affects the speed and efficiency of data retrieval, especially in scenarios such as database queries and file reading. Write operations refer to the process of writing data into a storage device. The performance of write operations is crucial for the timeliness of data updates and storage, and is commonly seen in scenarios such as log recording and file modification.

[0033] Based on access patterns, operations can be categorized into sequential operations and random operations. Sequential operations involve reading and writing data in a continuous sequence of addresses. Sequential operations typically offer higher throughput and are suitable for reading and writing large files, such as video stream storage and backup operations. Random operations involve reading and writing data across non-contiguous addresses. The performance of random operations is usually affected by the storage device's seek time and latency, and is commonly seen in database transaction processing and small file storage.

[0034] Classifying by block size allows you to calculate the block size of each I / O request and categorize it into different ranges, such as 4K, 8K, 16K, 32K, 64K, 128K, etc.

[0035] The collected raw data will undergo preprocessing, including data cleaning (removing outliers), data normalization (standardizing units), and data aggregation (summarizing by time window) to ensure the accuracy and consistency of the data.

[0036] This embodiment collects various I / O operation types (such as read / write, sequential / random) generated by business applications in the storage system and their corresponding performance data (such as block size, throughput, cache hit rate), and performs multi-dimensional classification and preprocessing on them, providing a comprehensive and accurate data foundation for subsequent I / O behavior modeling and performance evaluation, thereby realizing the quantitative analysis of complex low-level I / O operations.

[0037] S102. Calculate the data conversion ratio corresponding to the current I / O operation based on the baseline performance data and the I / O performance data.

[0038] Benchmark performance data are key reference values ​​used to standardize the performance of different I / O operations. These benchmark data are typically derived from standard testing environments for storage systems, ensuring that the performance of different I / O operations is measured under uniform conditions.

[0039] Benchmark performance data includes predefined performance metrics for unit resource consumption under standard I / O conditions. These predefined performance metrics can include benchmark IOPS, benchmark throughput, benchmark cache hit rate, etc. These predefined performance metrics can be obtained through rigorous performance testing in a standard testing environment, and an application I / O model matrix library can be constructed based on the test results.

[0040] In one embodiment, test performance data of storage devices of at least one device type under at least one I / O operation condition are collected; performance tests of preset standard I / O operations are performed in a standardized test environment to obtain benchmark test performance data; based on the benchmark test performance data, the test performance data of each storage device are normalized to obtain benchmark performance data corresponding to at least one device type; based on the benchmark performance data corresponding to each device type, the application I / O model matrix library is constructed.

[0041] In one embodiment, representative storage device models are selected as benchmark performance test objects based on the device types of common storage devices in the storage system and the business application scenarios. These storage device models can cover different storage media types (such as SSDs (Solid State Disks), HDDs (Hard Disk Drives), etc.) and storage architectures (such as SANs (Storage Area Networks), NAS (Network Attached Storage), etc.) to ensure that the benchmark performance data comprehensively reflects the performance characteristics of the storage system. For example, mainstream high-performance SSD models and mid-performance HDD models can be selected as test objects.

[0042] To ensure the accuracy and repeatability of benchmark performance data, a standardized test environment needs to be built. This standardized test environment uses server hardware similar to that of a real-world business environment, ensuring that the configuration of CPU, memory, network interfaces, etc., meets the requirements of storage performance testing. Simultaneously, it must be ensured that the storage devices in the test environment are in good working order, free from hardware failures or performance bottlenecks. Standard storage performance testing tools should be installed and configured. These tools should have rich test parameter configuration options and be able to simulate I / O operations under various real-world business scenarios. Furthermore, the versions of the testing tools should be consistent to avoid deviations in test results due to software version differences. If the test involves network storage (such as SAN or NAS), it is necessary to ensure that network bandwidth and latency meet standard testing requirements. Network bandwidth should be high enough to avoid the impact of network bottlenecks on storage performance test results; at the same time, network latency should be kept at a low level to ensure that the test results accurately reflect the performance of the storage devices.

[0043] Based on the typical application scenarios and business requirements of the storage system, define baseline I / O operation types and test parameters. These baseline I / O operation types and test parameters should comprehensively cover various I / O behaviors of the storage system in actual operation. For example, baseline I / O operation types may include read operations, write operations, sequential operations, random operations, etc., and test parameters may include I / O operations with different block sizes, cache hit rate, etc.

[0044] After defining the standardized test environment, baseline I / O operation types, and test parameters, perform performance tests and record the data.

[0045] Specifically, the storage device is formatted and initialized to ensure it is in a consistent state at the start of the test, eliminating the influence of historical data on the test results. The parameters of the performance testing tool are configured according to the test requirements, such as I / O block size, test time, and number of concurrent threads, ensuring that the tool configuration matches the predefined test parameters.

[0046] In a standardized testing environment, run testing tools to perform benchmark I / O operation tests individually for each storage device model. Record the storage system's performance data under different I / O operation types, such as IOPS, throughput, latency, and cache hit rate. During the testing process, maintain a stable testing environment and avoid external interference affecting the test results. To ensure the reliability of the test results, multiple repetitions are usually required, and the average value is taken as the final benchmark performance data. The number of repetitions should be determined based on actual needs and the stability of the testing environment; for example, more than three repetitions may be performed.

[0047] After testing, the obtained benchmark performance data is analyzed and verified to ensure its accuracy and reliability. Specifically, the test data is checked for outliers or errors, such as data deviations caused by hardware failure, network jitter, or incorrect testing tool configuration. Outlier data should be marked and removed, and the test should be repeated to obtain accurate data.

[0048] Performance data under different test conditions are normalized to analyze the performance differences of storage devices under different I / O operation conditions.

[0049] Specifically, a representative and stable I / O operation type is selected as the baseline I / O operation. For example, a 4KB random read operation with a 90% cache hit rate is selected as the baseline. For each storage device model and each I / O operation condition, the measured unit performance data is divided by the unit performance data of that device under the baseline I / O operation condition. Based on the baseline performance data, the test performance data of different storage device models under different I / O operation conditions are normalized to form baseline performance data corresponding to different storage device models. The baseline performance data includes at least one standardized performance indicator, which may include standardized block size, standardized cache hit rate, standardized throughput, etc.

[0050] The normalization calculation formula can be expressed as:

[0051] Normalization analysis can verify the accuracy of the obtained benchmark performance data. If the test results deviate significantly from known performance, the test environment and parameters should be re-examined to ensure the correctness of the testing process.

[0052] The obtained benchmark performance data is stored in a database or configuration file to obtain the application I / O model matrix library as shown in Table 1: Table 1

[0053] In one embodiment, the application I / O model matrix library is a collection of performance data for various storage devices under different I / O operation conditions. It organizes the data in matrix form; for example, each row can represent a type of storage device, and each column can represent a specific I / O operation condition (such as block size, cache hit rate, sequential / random operation type, etc.). Each element in the application I / O model matrix is ​​a baseline performance indicator (such as IOPS, throughput, latency, etc.) for that storage device under the corresponding I / O condition. This allows for the construction of a unified business I / O performance profiling system, achieving a standardized conversion from "device performance" to "application performance."

[0054] In the application I / O model matrix library, the benchmark performance data storage content can include device model, test conditions, benchmark performance metrics, test time, and test environment. The device model includes the specific model of the storage device, such as brand, series, and model number. Test conditions can include parameters such as I / O operation type, block size, cache hit rate, and number of concurrent threads during the test. Benchmark performance metrics can include benchmark IOPS, benchmark throughput, benchmark latency, and benchmark cache hit rate. Test time refers to the time the benchmark performance data was acquired, for subsequent updates and calibration. Test environment data includes information such as hardware configuration, software version, and network conditions during the test to ensure the repeatability of test results.

[0055] Understandably, the benchmark performance data in the application I / O model matrix library should be updated and maintained regularly to ensure its accuracy and timeliness. Specifically, benchmark performance tests should be re-performed periodically based on storage device usage and business needs. For newly released storage device models or technology updates, testing and updating of benchmark performance data should be conducted promptly. The update cycle can be determined based on actual needs; for example, a comprehensive update every six months to one year.

[0056] This embodiment constructs a standardized application I / O model matrix library, normalizing the performance data of different storage devices under various I / O operation conditions into a unified benchmark performance index, thereby achieving a standardized conversion from "device performance" to "application performance".

[0057] In actual business operations, based on the real-time collected I / O performance data, the corresponding benchmark performance data is found by applying the I / O model matrix library, which is used to calculate the data conversion ratio of the real-time collected I / O performance data, thereby achieving dynamic I / O fitting.

[0058] Furthermore, the baseline performance data corresponding to the current I / O operation is queried from the pre-built application I / O model matrix library; based on the baseline performance data and the I / O performance data, the data conversion ratio of the current I / O operation is calculated.

[0059] In one embodiment, device information of the currently detected storage device, including device name and model, can be queried from the current storage system. Based on the I / O performance data corresponding to the current I / O operation, characteristic parameters such as block size, cache hit rate, sequential / random operation type, and read / write ratio are extracted.

[0060] Based on device information and feature parameters extracted from I / O performance data, the current I / O operation is matched with the I / O operation types defined in the application I / O model matrix library to determine the corresponding I / O operation type. For example, if real-time data indicates that the current I / O operation is a 4KB random read operation with a cache hit rate of 90%, then it is matched as the corresponding I / O operation type.

[0061] Connect to the database or data storage system that stores the application I / O model matrix library, access the application I / O model matrix library, and prepare for a data query operation. Based on the determined I / O operation type corresponding to the current I / O operation, construct a precise query statement. This query statement should include key fields such as device model, block size, cache hit rate, sequential / random operation type, and read / write ratio to ensure the accuracy of the query results. Run the query statement to retrieve benchmark performance data from the application I / O model matrix library that perfectly matches the I / O operation type corresponding to the current I / O operation. The benchmark performance data includes key performance indicators such as IOPS, throughput, latency, and resource consumption per unit (I / O CPU).

[0062] Perform data preprocessing on the baseline performance data and I / O performance data corresponding to the current I / O operation, such as data normalization, unit unification, and data cleaning.

[0063] Data normalization is used to normalize the collected I / O performance data (such as actual IOPS, throughput, and cache hit rate) to the dimension of the baseline performance data. For example, if the baseline performance data is based on a 4KB block size, while the actual collected I / O data is in 8KB blocks, then the 8KB performance data needs to be converted into equivalent 4KB performance data.

[0064] Unit standardization is used to ensure that the units of benchmark performance data and real-time acquired I / O performance data are consistent. For example, if benchmark IOPS is in "times per second" while real-time IOPS is in "times per minute", then the real-time data needs to be converted to "times per second".

[0065] Data cleaning is used to remove outliers or noisy data from real-time I / O performance data. For example, it removes extremely high or low performance data points caused by brief network jitter or hardware failure to ensure the reliability of calculation results.

[0066] In one embodiment, based on the I / O performance data, the actual performance index value of the target performance index is determined; based on the benchmark performance data, the benchmark performance index value of the target performance index is determined; and based on the ratio of the actual performance index value to the benchmark performance index value, the data conversion ratio is determined.

[0067] After data preprocessing, select appropriate target performance metrics for conversion ratio calculation based on analysis needs. These metrics can include one or more of the following: IOPS, throughput, latency, and unit resource consumption. For example, if the focus is on the response speed of the storage system, latency can be selected as the calculation metric.

[0068] Based on the selected performance metrics, the data conversion ratio is calculated using the following formula:

[0069] Among them, the real-time performance indicators are derived from the real-time collected I / O performance data, and the benchmark performance indicators are derived from the benchmark performance data.

[0070] Taking unit resource consumption as an example, the actual unit resource consumption value is extracted from the real-time collected I / O performance data, and the baseline unit resource consumption value is extracted from the baseline performance data. The above data conversion ratio calculation formula is then applied to calculate:

[0071] If multiple performance metrics need to be considered comprehensively, a weighted average method can be introduced to calculate the overall data conversion ratio, thereby obtaining a more comprehensive performance evaluation. Taking throughput and other performance metrics as examples:

[0072] in, These are weighting coefficients, which can be adjusted according to actual needs.

[0073] The calculated data conversion ratio needs to be calibrated and verified to ensure its accuracy and reliability.

[0074] Specifically, the calculated data conversion ratio can be calibrated by comparing it with a storage system or test environment with known performance. For example, the accuracy of the calculation results can be verified by using standard performance testing tools (such as FIO) on a storage system with known performance.

[0075] Cross-validation can be used to compare the calculated data conversion ratio with actual performance. For example, actual business load tests can be conducted to verify whether the data conversion ratio accurately reflects the performance differences of the storage system.

[0076] This embodiment collects I / O performance data in real time during business operation, dynamically matches and obtains benchmark performance data by combining it with a pre-built application I / O model matrix library, and realizes dynamic fitting and normalization quantification of actual I / O behavior by calculating the data conversion ratio, thereby transforming the complex and ever-changing real-time I / O load into standardized and comparable performance indicators.

[0077] S103. Based on the data conversion ratio corresponding to the current I / O operation and the I / O operation quantity (IOPS) of the current I / O operation, calculate the baseline total I / O of the service application under standard I / O conditions.

[0078] In one embodiment, the data conversion ratio can be the data conversion ratio corresponding to a certain type of I / O performance data in the current I / O operation, or it can be the data conversion ratio corresponding to multiple I / O performance data in the current I / O operation, or it can be the data conversion ratio after weighted fusion of multiple I / O performance data in the current I / O operation.

[0079] In one embodiment, a data conversion ratio can be calculated based on actual application requirements, a data conversion ratio mapping table can be constructed, and an I / O operation statistics table can be constructed. The data conversion ratio mapping table records the data conversion ratio corresponding to each I / O operation type in the current I / O operation. ,in Indicates the first I / O operations. These data conversion ratios are calculated by comparing actual I / O performance with baseline performance, used to standardize the actual performance of different I / O operations to baseline conditions. The I / O operation statistics table records the average IOPS value of each I / O operation type within the sampling period. This reflects the frequency of I / O operations in current business applications.

[0080] Before calculating the baseline I / O total, the input data is first validated and preprocessed. First, the data conversion ratio mapping table and the I / O operation statistics table are checked for completeness to ensure they contain the same I / O operation types, guaranteeing data alignment. If incomplete data is found, the missing I / O operation types are recorded, and measures are taken to supplement data or exclude that type of I / O operation, depending on the actual situation.

[0081] Next, the IOPS values ​​in the I / O operation statistics table are checked for reasonableness, and obviously abnormal data points, such as negative values ​​or values ​​exceeding the theoretical performance limits of the device, are removed. These abnormal values ​​may be caused by data acquisition errors or hardware failures, and removing these values ​​can avoid interference with subsequent calculations.

[0082] Finally, the data is uniformly converted into a standard format to ensure the accuracy of subsequent calculations. This includes verifying whether the data types are correct, such as ensuring that IOPS values ​​are numeric and I / O operation type identifiers are string types, and performing necessary type conversions on the data.

[0083] After data verification and preprocessing are completed, the calculation of baseline I / O quantities begins.

[0084] Further, based on the data conversion ratio and the I / O operation amount corresponding to the current I / O operation, at least one baseline I / O amount corresponding to the current I / O operation is calculated; the at least one baseline I / O amount corresponding to the current I / O operation is accumulated to obtain the total baseline I / O amount.

[0085] For each I / O operation type in the current I / O operation, the following calculation formula is applied:

[0086] Among them, the reference I / O quantity Indicates the first The baseline I / O operation quantity corresponding to the I / O operation type under standard I / O conditions; It is the first The data conversion ratio for I / O operation types is derived from the data conversion ratio mapping table; It is the first The actual I / O operation volume for each I / O operation type is obtained from the I / O operation volume statistics table.

[0087] Iterate through the data conversion ratio mapping table to find the I / O operation type and read the corresponding value from the table. The value is read from the I / O operation statistics table. The value is calculated, and then a multiplication operation is performed to obtain the baseline I / O quantity for that type of I / O operation. The calculation result can be stored in a temporary result table, recording information such as the I / O operation type, baseline I / O quantity, and calculation time, for subsequent summary calculations and traceability.

[0088] After calculating the baseline I / O quantity for a single type of I / O operation, the baseline I / O quantities for all I / O operation types are read from the temporary result table, and an accumulation operation is performed to sum them up. The values ​​are summed to calculate the total baseline I / O. The summation calculation model is as follows:

[0089] Here, ∑ represents summing over all I / O operation types.

[0090] The calculation results can be formatted, retaining a preset number of decimal places and adding units (such as "times / second"). The final baseline I / O total can be stored in the results table, recording information such as the business application identifier, the baseline I / O total, and the calculation time. This baseline I / O total reflects the overall performance of the business application under standard I / O conditions and can be used for subsequent performance evaluation, resource optimization, and billing model calculation.

[0091] This embodiment calculates the baseline I / O total under standard I / O conditions by multiplying the actual performance (IOPS) of various I / O operations in business applications with their corresponding data conversion ratios and summing them up. This realizes the unified transformation of complex loads with different I / O modes and frequencies into quantifiable and comparable performance indicators.

[0092] S104. Calculate the performance density of the storage system based on the number of beneficiaries and the total baseline I / O.

[0093] In one embodiment, the baseline I / O total reflects the total I / O operations of the business application under standard I / O conditions. The data source for the benefit count can be the storage system's management module, the business application's configuration information, or the billing system, ensuring its accuracy and real-time performance.

[0094] Further, based on the storage resource usage of the storage system, the benefit number is calculated, wherein the storage resource usage includes at least one of storage capacity, number of storage volumes, or storage service duration; based on the baseline I / O total and the benefit number, the performance density of the storage system is calculated, and the performance density is used to characterize the efficiency of the storage system in meeting the performance requirements of business applications.

[0095] Benefit count is a key metric for measuring the actual usage of storage resources by business applications, reflecting the efficiency of the storage system in meeting business needs. Benefit count can be determined in various ways, including by storage capacity (e.g., TB), number of storage volumes, or service duration (e.g., hours).

[0096] For example, if the benefit is determined by storage capacity, it is calculated in units of the actual used storage capacity (e.g., TB). For instance, if a business application uses 10TB of storage capacity, the benefit can be expressed as 10TB. Alternatively, if the benefit is determined by the number of storage volumes, it is calculated in units of the number of storage volumes. For instance, if a business application uses 5 storage volumes, the benefit can be expressed as 5. Or, if the benefit is determined by service duration, it is calculated in units of the duration of storage service usage (e.g., hours). For instance, if a business application uses storage services for 100 hours, the benefit can be expressed as 100 hours.

[0097] Depending on the specific application requirements, the benefit count can also be a comprehensive calculation result of the above factors (storage capacity, number of storage volumes, and service duration). For example, the benefit count can be expressed as the product of storage capacity, number of storage volumes, and service duration, i.e. .

[0098] Furthermore, the number of beneficiaries is adjusted based on the priority or quality of service (QoS) requirements of the business applications to reflect the differences in the actual storage resource needs of different business applications. For example, the number of beneficiaries for high-priority business applications can be multiplied by a weight greater than 1, while the number of beneficiaries for low-priority business applications can be multiplied by a weight less than 1.

[0099] Performance density is a key metric for measuring the efficiency of a storage system in meeting the performance requirements of business applications, and is used for quantitative evaluation of storage system performance. Performance density is calculated using a baseline total I / O and the benefited number of TB. Before calculating performance density, ensure that the units of the baseline total I / O and the benefited number are consistent for accurate calculation. For example, if the baseline total I / O is 238,000 IOPS and the benefited number is 10TB, then the performance density is 23,800 IOPS / TB.

[0100] The formula for calculating performance density can be expressed as:

[0101] The baseline I / O total is measured in IOPS (Input / Output Operations Per Second), while the benefit is measured in storage capacity (TB), number of storage volumes, or service duration (hours). Performance density can be measured in IOPS / TB, IOPS / volume, or IOPS / hour, depending on the unit of the benefit.

[0102] This embodiment calculates the performance density by dividing the baseline I / O total by the benefited number (such as storage capacity, volume number, or service duration), thereby establishing a correlation between the performance of the storage system and resource consumption, and achieving a precise quantitative assessment of the service efficiency of storage resources.

[0103] S105. Based on the performance density, determine the performance level of the storage system to evaluate the performance of the storage system.

[0104] The performance levels of a storage system are defined based on its actual application scenarios and business requirements. Based on these defined performance levels, performance density is then divided into different performance tiers. For example, high performance, medium performance, and low performance; or Tier 1, Tier 2, and Tier 3, etc. The specific rules for classifying performance tiers can be determined according to the actual application scenarios and business needs.

[0105] Different performance levels correspond to different performance density thresholds. Taking high performance, medium performance, and low performance as examples, the high performance level corresponds to a performance density higher than a preset high threshold, the medium performance level corresponds to a performance density within the medium threshold range, and the low performance level corresponds to a performance density lower than a preset low threshold. The performance level evaluation thresholds can be dynamically adjusted based on the storage system type, business type, and user needs.

[0106] In one embodiment, the performance density is compared with at least one preset performance level evaluation threshold to obtain a comparison result; based on the comparison result, the performance level corresponding to the storage system is determined.

[0107] The calculated performance density is compared with predefined performance level evaluation thresholds to determine the performance level of the storage system. Specifically, the current performance density value of the storage system is read and compared with the high, medium, and low thresholds in the performance level standard. Based on the comparison results, the performance level to which the storage system belongs is determined.

[0108] For example, if the performance density value is higher than the high threshold, the storage system can be rated as high performance; if the performance density value is within the medium threshold range, the storage system can be rated as medium performance; and if the performance density value is lower than the low threshold, the storage system can be rated as low performance.

[0109] Determining performance levels can be a continuous monitoring and adjustment process. Performance density can be recalculated periodically, and the storage system's performance level can be reassessed based on the latest performance density values. Within the storage system, built-in monitoring modules can automatically perform performance density recalculation and performance level reassessment, analyze performance level trends, and predict the storage system's performance future.

[0110] To better manage storage resources, performance density trends can be predicted, allowing for proactive adjustments to storage resource allocation. Specifically, historical performance density data is collected, recording performance density values ​​at different points in time. This historical data can be obtained from the storage system's database or log files. Data analysis tools (such as linear regression and time series analysis) are used to predict performance density trends, constructing a performance density trend chart. For example, a linear regression model can be used to predict performance density changes over a future period. Based on these performance density trends, future performance requirements of the storage system are predicted, and storage strategies and billing standards are adjusted accordingly. For instance, if performance density continues to decline, resource expansion or hardware upgrades can be planned in advance, and billing standards can be submitted accordingly.

[0111] In one embodiment, a performance evaluation report for the storage system can be generated based on the performance level. This performance analysis report may include the storage system's unique identifier, I / O characteristic distribution map, performance density value, performance density trend map, performance level evaluation results, and optimization suggestions or improvement measures provided based on the performance level.

[0112] Performance evaluation reports can be applied to the management and optimization of storage systems. Specifically, storage resource allocation can be dynamically adjusted based on performance levels; for example, high-priority services can be migrated to high-performance storage systems, while low-priority services can be migrated to low-performance storage systems.

[0113] The storage system can be optimized based on the optimization recommendations in the performance evaluation report, such as by adjusting caching strategies, increasing storage bandwidth, or upgrading storage devices to improve performance. Specifically, the caching strategy of the storage system should be adjusted according to the recommendations in the performance evaluation report. First, if the performance evaluation report recommends increasing cache capacity, this can be achieved by adding cache modules or adjusting cache allocation. For example, increasing the cache capacity from 10GB to 20GB. Second, the cache allocation strategy can be adjusted based on the I / O characteristics of the business applications. For example, for applications with high read frequency, increase the read cache ratio; for applications with high write frequency, increase the write cache ratio. Finally, a suitable caching algorithm can be selected based on the characteristics of the business applications, such as FIFO (First Input First Output) or an adaptive caching algorithm.

[0114] In addition, storage systems can be optimized through hardware and software upgrades. For example, hardware optimization can be achieved by adding bandwidth modules, upgrading network interfaces, upgrading storage media and storage architecture, while software optimization can be achieved by optimizing data transmission protocols and load balancing.

[0115] This embodiment determines the performance level of the storage system by comparing the calculated performance density with a preset performance level evaluation threshold, and generates a performance evaluation report containing optimization suggestions based on this, thereby achieving accurate classification and quantitative evaluation of the performance status of the storage system.

[0116] Furthermore, most mainstream storage services currently use static billing based on device type, which fails to dynamically reflect the performance requirements of actual business loads, leading to both resource waste and performance bottlenecks. To address this technical issue, this embodiment proposes a performance-level-based billing method for storage systems. This method aims to dynamically adjust billing standards based on the actual performance of the storage system, thereby achieving a fair, transparent, and efficient billing mechanism.

[0117] In one embodiment, a billing standard corresponding to the performance level of the storage system is determined; and the billing amount of the storage system is calculated based on the billing standard and the resource usage of the storage system.

[0118] The performance-level-based billing method for storage systems proposed in this embodiment can adjust the billing strategy according to the performance level, charging higher fees for high-performance storage systems and offering preferential prices for low-performance storage systems. To achieve accurate billing of storage systems, this solution proposes a dynamic billing method based on performance levels.

[0119] First, storage systems are categorized into different performance levels based on their performance density. Performance density is calculated by comprehensively considering the I / O operations and actual benefits (such as storage capacity and usage time) of the storage system, and it accurately reflects the efficiency of the storage system in meeting business needs.

[0120] For example, if the performance levels of a storage system are classified into high-performance, medium-performance, and low-performance levels: if the performance density of the storage system is greater than or equal to the high-performance threshold (e.g., 20,000 IOPS / TB), then the storage system's performance level is determined to be high-performance; if the performance density of the storage system is less than or equal to the low-performance threshold (e.g., 10,000 IOPS / TB), then the storage system's performance level is determined to be low-performance; if the performance density of the storage system is greater than the low-performance threshold but less than the high-performance threshold (e.g., between 10,000 IOPS / TB and 20,000 IOPS / TB), then the storage system's performance level is determined to be medium-performance.

[0121] Based on the performance level of the storage system, a billing standard corresponding to that performance level is determined. The billing standard is set according to the performance of the storage system, with different performance levels corresponding to different billing standards. This billing standard includes the unit price corresponding to different performance levels; for example, the unit price for a high-performance level is higher than the unit price for a medium-performance level, and the unit price for a medium-performance level is higher than the unit price for a low-performance level.

[0122] For example, for high-performance storage systems, the billing standard can be set at a unit price of 1 yuan / TB / hour; for medium-performance storage systems, the billing standard can be set at a unit price of 0.8 yuan / TB / hour; and for low-performance storage systems, the billing standard can be set at a unit price of 0.5 yuan / TB / hour.

[0123] The billing amount is calculated based on the storage system's resource usage (including storage capacity and usage time) and the corresponding billing standard. The formula for calculating the billing amount can be expressed as:

[0124] For example, assuming a storage system has a high performance rating, a storage capacity of 10TB, and is used for 100 hours, the billing amount would be calculated as follows:

[0125] Billing standards can be dynamically adjusted based on the type of storage system, business type, and user needs. For example, for high-priority business applications, the unit price can be appropriately increased through a weighted average to reflect their demand for high performance.

[0126] For example, if the storage system in the example above serves high-priority services, the unit price can be increased to a weighted average of 1.2, and the billing amount will be adjusted as follows:

[0127] To better adapt to changes in storage system performance, billing standards can also be dynamically adjusted based on trends in performance density. For example, if a forecast indicates a decrease in performance density, the system will adjust the billing standards in advance to reflect the real-time performance of the storage system.

[0128] Specifically, the billing standard is dynamically adjusted based on the changing trend of performance density. For example, if performance density continues to decline, the unit price can be adjusted from 1 yuan / TB / hour to 0.8 yuan / TB / hour. The billing standard is updated based on the forecast results to ensure that the billing model always matches the actual performance of the storage system.

[0129] This embodiment directly links the performance level of the storage system to the billing standard and calculates the billing amount based on resource usage (storage capacity and usage time), realizing the transformation from "pay-per-device" to "pay-per-performance". This enables the dynamic reflection of the performance requirements of the actual business load, effectively solving the problem of resource waste or performance bottlenecks coexisting, and realizing a fair, transparent and efficient billing mechanism.

[0130] This embodiment provides a storage performance analysis method. This method collects I / O performance data generated by business applications, achieving accurate perception from device performance to application performance. It accurately reflects the storage performance requirements of business applications in the actual operating environment, ensuring the real-time nature and relevance of the data. By calculating the data conversion ratio, the collected actual performance data is compared with predefined benchmark performance data. Complex and diverse non-benchmark I / O operation types are normalized into standard benchmark I / O units, quantifying the performance differences between different I / O operation types and avoiding evaluation bias caused by the diversity of I / O operation types. By combining the data conversion ratio with the actual I / O operation volume (IOPS), the benchmark total I / O of the business application under standard I / O conditions is calculated, achieving standardization of performance evaluation. This provides a unified quantitative indicator for the overall performance evaluation of the storage system, allowing the storage performance of different business applications to be compared under the same benchmark conditions, improving the comparability and consistency of evaluation results. By calculating performance density, the benchmark total I / O is correlated with the benefit number, considering not only the performance output of the storage system but also the actual resource usage. By calculating performance density, the efficiency of a storage system in meeting actual business needs can be more accurately assessed, thus avoiding resource waste or inadequacy that may result from solely relying on performance metrics. Determining performance levels based on performance density enables precise grading and assessment of the storage system's performance status. This not only allows users to quickly understand the storage system's performance level but also enables the development of corresponding optimization strategies based on different performance levels. Performance density-based tiered assessment more accurately reflects the storage system's performance in real-world business scenarios, thereby providing strong support for resource planning and performance optimization.

[0131] Please see Figure 3 , Figure 3 This is a schematic diagram of a current embodiment of a storage performance analysis device provided in this application, which is used to perform the aforementioned storage performance analysis method.

[0132] like Figure 3 As shown, the storage performance analysis device 200 includes: a data acquisition module 201, a conversion ratio calculation module 202, a baseline I / O total calculation module 203, a performance density calculation module 204, and a storage performance evaluation module 205.

[0133] Data acquisition module 201 is used to acquire I / O performance data generated by at least one current I / O operation corresponding to the business application in the storage system; The conversion ratio calculation module 202 is used to calculate the data conversion ratio corresponding to the current I / O operation based on the baseline performance data corresponding to the current I / O operation and the I / O performance data; The baseline I / O total calculation module 203 is used to calculate the baseline I / O total of the service application under standard I / O conditions based on the data conversion ratio corresponding to the current I / O operation and the I / O operation quantity (IOPS) of the current I / O operation. The performance density calculation module 204 is used to calculate the performance density of the storage system based on the number of benefits and the total baseline I / O. Storage performance evaluation module 205 is used to determine the performance level of the storage system based on the performance density, so as to evaluate the performance of the storage system.

[0134] It should be noted that those skilled in the art will understand that, for the sake of convenience and brevity, the specific working processes of the above-described apparatus and modules can be referred to the corresponding processes in the aforementioned storage performance analysis method embodiments, and will not be repeated here.

[0135] The apparatus provided in the above embodiments can be implemented as a computer program, which can be used in, for example... Figure 4 It runs on the computer device shown.

[0136] Please see Figure 4 , Figure 4 This is a schematic block diagram of a computer device provided in an embodiment of this application. The computer device may be a server.

[0137] See Figure 4 The computer device includes a processor, memory, and network interface connected via a system bus, wherein the memory may include non-volatile storage media and internal memory.

[0138] Non-volatile storage media can store operating systems and computer programs. These computer programs include program instructions that, when executed, cause the processor to perform any storage performance analysis method.

[0139] The processor provides computing and control capabilities, supporting the operation of the entire computer device.

[0140] Internal memory provides an environment for the execution of computer programs on non-volatile storage media. When these computer programs are executed by a processor, the processor can perform any storage performance analysis method.

[0141] This network interface is used for network communication, such as sending assigned tasks. Those skilled in the art will understand that... Figure 4The structure shown is merely a block diagram of a portion of the structure related to the present application and does not constitute a limitation on the computer device to which the present application is applied. Specific computer devices may include more or fewer components than those shown in the figure, or combine certain components, or have different component arrangements.

[0142] It should be understood that the processor can be a Central Processing Unit (CPU), but it can also be other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. Among these, a general-purpose processor can be a microprocessor or any conventional processor.

[0143] In one embodiment, the processor is configured to run a computer program stored in memory to perform the following steps: Collect I / O performance data generated by at least one current I / O operation corresponding to the business application in the storage system; Calculate the data conversion ratio corresponding to the current I / O operation based on the baseline performance data and the I / O performance data corresponding to the current I / O operation; Based on the data conversion ratio corresponding to the current I / O operation and the I / O operation quantity (IOPS) of the current I / O operation, calculate the baseline total I / O of the service application under standard I / O conditions; Calculate the performance density of the storage system based on the number of beneficiaries and the total baseline I / O; Based on the performance density, the performance level of the storage system is determined to evaluate the performance of the storage system.

[0144] The embodiments of this application also provide a computer-readable storage medium storing a computer program, the computer program including program instructions, and the processor executing the program instructions to implement any of the storage performance analysis methods provided in the embodiments of this application.

[0145] The computer-readable storage medium may be an internal storage unit of the computer device described in the foregoing embodiments, such as the hard disk or memory of the computer device. The computer-readable storage medium may also be an external storage device of the computer device, such as a plug-in hard disk, SmartMediaCard (SMC), SecureDigital (SD) card, or FlashCard equipped on the computer device.

[0146] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any person skilled in the art can easily conceive of various equivalent modifications or substitutions within the technical scope disclosed in this application, and these modifications or substitutions should all be covered within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

Claims

1. A storage performance analysis method, characterized in that, The method includes: Collect I / O performance data generated by at least one current I / O operation corresponding to the business application in the storage system; Calculate the data conversion ratio corresponding to the current I / O operation based on the baseline performance data and the I / O performance data corresponding to the current I / O operation; Based on the data conversion ratio corresponding to the current I / O operation and the I / O operation amount of the current I / O operation, calculate the baseline total I / O of the business application under standard I / O conditions; Calculate the performance density of the storage system based on the number of beneficiaries and the total baseline I / O; Based on the performance density, the performance level of the storage system is determined to evaluate the performance of the storage system.

2. The storage performance analysis method according to claim 1, characterized in that, The step of calculating the data conversion ratio corresponding to the current I / O operation based on the baseline performance data and the I / O performance data includes: Query the benchmark performance data corresponding to the current I / O operation from the pre-built application I / O model matrix library; Based on the baseline performance data and the I / O performance data, the data conversion ratio of the current I / O operation is calculated.

3. The storage performance analysis method according to claim 2, characterized in that, The step of calculating the data conversion ratio of the current I / O operation based on the baseline performance data and the I / O performance data includes: Based on the I / O performance data, determine the actual performance index value of the target performance index; Based on the benchmark performance data, determine the benchmark performance index value of the target performance index; The data conversion ratio is determined based on the ratio of the actual performance index value to the benchmark performance index value.

4. The storage performance analysis method according to claim 2, characterized in that, Before querying the benchmark performance data corresponding to the current I / O operation from the pre-built application I / O model matrix library, the process also includes: Collect test performance data of storage devices of at least one device type under at least one I / O operation condition; Performance tests are performed on preset standard I / O operations in a standardized testing environment to obtain benchmark performance data. Based on the benchmark performance data, the test performance data of each of the storage devices are normalized to obtain the benchmark performance data corresponding to at least one device type. Based on the benchmark performance data corresponding to each of the aforementioned device types, the application I / O model matrix library is constructed.

5. The storage performance analysis method according to claim 1, characterized in that, The step of calculating the baseline total I / O of the service application under standard I / O conditions based on the data conversion ratio corresponding to the current I / O operation and the I / O operation amount of the current I / O operation includes: Based on the data conversion ratio and the I / O operation amount corresponding to the current I / O operation, calculate at least one baseline I / O amount corresponding to the current I / O operation; The total amount of the reference I / O is obtained by summing up at least one reference I / O quantity corresponding to the current I / O operation.

6. The storage performance analysis method according to claim 1, characterized in that, The calculation of the performance density of the storage system based on the number of beneficiaries and the baseline I / O total includes: The number of beneficiaries is calculated based on the storage resource usage of the storage system, wherein the storage resource usage includes at least one of storage capacity, number of storage volumes, or storage service duration. The performance density of the storage system is calculated based on the baseline I / O total and the benefit number.

7. The storage performance analysis method according to claim 1, characterized in that, After determining the performance level of the storage system based on the performance density, the process further includes: Based on the performance level of the storage system, a billing standard corresponding to the performance level is determined; The billing amount for the storage system is calculated based on the billing standard and the resource usage of the storage system.

8. A storage performance analysis device, characterized in that, The storage performance analysis device includes: The data acquisition module is used to collect I / O performance data generated by at least one current I / O operation corresponding to the business application in the storage system; The conversion ratio calculation module is used to calculate the data conversion ratio corresponding to the current I / O operation based on the baseline performance data and the I / O performance data corresponding to the current I / O operation; The baseline I / O total calculation module is used to calculate the baseline I / O total of the service application under standard I / O conditions based on the data conversion ratio corresponding to the current I / O operation and the I / O operation quantity (IOPS) of the current I / O operation. The performance density calculation module is used to calculate the performance density of the storage system based on the number of benefits and the total baseline I / O. A storage performance evaluation module is used to determine the performance level of the storage system based on the performance density, so as to evaluate the performance of the storage system.

9. A computer device, characterized in that, The computer device includes a processor, a memory, and a computer program stored in the memory and executable by the processor, wherein when the computer program is executed by the processor, it implements the steps of the storage performance analysis method as described in any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a computer program, wherein when the computer program is executed by a processor, it implements the steps of the storage performance analysis method as described in any one of claims 1 to 7.

Images