Load verification system and load verification method

Abstract

To enable verification at a migration destination to be performed appropriately in a short time when data stored in a storage system or the like is migrated.SOLUTION: A load verification system for verifying load performance of a data storage area is provided with a verification performance metrics collecting unit 112 for collecting performance metrics of a data storage area to be verified with load and a volume thereof, an expected performance metrics data generating unit 111 for generating performance metrics data resulting in an expected load with respect to a volume to be verified with load, and an I / O pattern data generating unit 113 for generating input / output pattern data for generating a load resulting in the performance metrics data corresponding to the expected performance indicated by the expected performance metrics data. The input / output pattern is then reproduced by a reproduction unit based on data generated by the I / O pattern data generation unit 113 to collect the performance metrics as a result of applying a load to the volume to be verified with load.SELECTED DRAWING: Figure 1

Description

[Technical Field]

[0001] The present invention relates to a load verification system and a load verification method. [Background technology]

[0002] When migrating data stored in a storage system, it is common to first verify the performance of the volume in the actual environment of the destination. In this performance verification, a load is applied by data input / output (I / O) to the volume created in the target storage system, and it is confirmed whether the performance of the migrated volume and the impact on other business (applications) using the destination storage device meet the requirements.

[0003] Patent Document 1 describes a technique for constructing an environment equivalent to the migration destination when verifying the load on the migration destination. [Prior art documents] [Patent documents]

[0004] [Patent Document 1] International Publication No. 2015 / 049771 Summary of the Invention [Problem to be solved by the invention]

[0005] The migration destination volume shares shared resources (ports, processors, etc.) within the storage system with other volumes used for other purposes. Therefore, if a load is generated on the migration destination volume, the generated load will have a performance impact on the other volumes that use the shared resources. For this reason, when migrating data, it is necessary to verify the performance impact not only on the volume in question, but also on other resources within the device.

[0006] In a data migration, the volume usage trends are the same before and after the migration, so the trends in performance metrics that show the results for the load generated at the migration source will also be the same for the migration destination. Here, the performance metrics that were generated on the source volume can be known in advance. In other words, if the I / O that generates a specific performance metric can be reproduced, a load equivalent to that of the source can be reproduced on the migration destination volume, making it possible to verify the performance impact on other volumes on the migration destination device when that load occurs. However, if the device models (specifications) of the source storage system and the destination storage system are different, the performance metrics that show the performance impact when the exact same I / O occurs will be different values for each model.

[0007] Due to this characteristic, in the past, when the source and destination device models were different, it was impossible to identify the I / O that would generate the expected performance metrics for the destination volume, making reproduction impossible. Therefore, in order to reproduce trends, it was necessary to apply load to the destination volume by trial and error, measure the performance impact, and then check the performance impact on other volumes at the same time. Applying load by trial and error means, for example, applying load based on an I / O pattern extracted randomly from accumulated load data or based on specific rules.

[0008] The explanation so far has focused on the issues that arise in the case of storage systems, but similar problems arise when verifying data migration even when combining different types of data storage systems, known as a hybrid cloud solution.

[0009] In view of the above, it is an object of the present invention to provide a load verification system and a load verification method that can perform verification appropriately in a short time when migrating data stored in a storage system or the like. [Means for solving the problem]

[0010] In order to solve the above problems, for example, the configurations described in the claims are adopted. The present application includes multiple means for solving the above-mentioned problems, and one example is a load verification system that performs load performance verification of a specified data storage area, and is equipped with a verification performance metrics collection unit that acquires performance metrics of the data storage area and its volume that are the subject of load verification, an expected performance metrics data generation unit that generates performance metrics data that results in the expected load for the volume that is the subject of load verification, and an I / O pattern data generation unit that generates input / output pattern data that generates a load that results in performance metrics data equivalent to the expected performance indicated in the expected performance metrics data. and, I / O pattern data generator Based on the input / output pattern data generated in step 1, input / output is performed to the volume to be subjected to load verification, and the verification performance metrics collection unit collects performance metrics as a result of applying the load. [Effects of the Invention]

[0011] This invention can provide input / output pattern data that reproduces the expected performance metrics for the volume being subjected to load verification. This makes it possible to reproduce the load generated in the source environment in the destination environment, shortening the verification period and enabling the timely provision of application services to be operated. Problems, configurations, and effects other than those described above will become apparent from the following description of the embodiments. [Brief explanation of the drawings]

[0012] [Figure 1] 1 is a configuration diagram showing an example of the overall load verification system according to an embodiment of the present invention; [Figure 2] 2 is a block diagram showing an example of the configuration of an I / O pattern data generating unit according to an embodiment of the present invention. FIG. [Figure 3] 1 is a block diagram showing an example of the configuration of a load verification management system according to an embodiment of the present invention; [Figure 4] FIG. 2 is a diagram showing an example of the configuration of performance metrics data according to an embodiment of the present invention. [Figure 5]FIG. 3 is a diagram showing an example of the configuration of I / O pattern data according to an embodiment of the present invention. [Figure 6] FIG. 10 is a diagram illustrating an example of a process for generating load data according to an embodiment of the present invention. [Figure 7] 10 is a flowchart illustrating an example of processing by an expected performance metrics data generation unit according to an embodiment of the present invention. [Figure 8] 10 is a flowchart illustrating an example of processing by an I / O pattern data generation unit according to an embodiment of the present invention. [Figure 9] 10 is a flowchart illustrating an example of processing by a load data generating unit according to an embodiment of the present invention. [Figure 10] 10 is a flowchart illustrating an example of processing by a load simulator according to an embodiment of the present invention. [Figure 11] 10 is a flowchart illustrating an example of processing by a verification performance metric collector according to an embodiment of the present invention. DETAILED DESCRIPTION OF THE INVENTION

[0013] A load verification system according to an embodiment of the present invention (hereinafter referred to as "this embodiment") will be described below with reference to the accompanying drawings.

[0014] [Load verification system configuration] FIG. 1 shows the overall configuration of the load verification system of this example. The load verification system of this example includes a load verification management system 100 and a load simulator 19, and performs verification of a storage device 10 to be verified that is installed in a data center DC1. The storage device 10 to be verified is a data migration destination. The storage device 10 to be verified comprises a plurality of volumes 10a to 10n (n is an arbitrary integer), each of which has a predetermined data storage area. The load simulator 19 is installed in the data center DC1.

[0015] The storage device 10 under verification has an interface 11 and a port 12, and simulated input / output pattern data from a reproducing unit 19a of a load simulator 19 is supplied to the volumes 10a to 10n under verification via the port 12. Note that the input / output pattern data and input / output data will be referred to as I / O pattern data and I / O data in the following explanation. Furthermore, the performance metrics data actually measured in each of the volumes 10a to 10n is transmitted from the interface 11 to the load verification management system 100 via a predetermined LAN (Local Area Network) 181.

[0016] Performance metrics data for storage (volume) includes various data that indicate the performance of the storage, such as read throughput, write throughput, number of read request operations per unit time, number of write request operations per unit time, read latency, write latency, utilization rate, etc. These performance metrics vary not only depending on the performance of the device (storage) itself, which is the data storage area, but also on the environment in which the storage is installed.

[0017] The load verification management system 100 receives a command from a user terminal 190, which is an external terminal. of The target performance metrics data is input, I / O pattern data is generated, and the generated I / O pattern data is supplied to the load simulator 19. Based on the data, the load simulator 19 applies a load to the volume to be verified according to the simulated I / O pattern data. but The load verification management system 100 collects performance metrics data actually measured when a load is applied, and performs load verification.

[0018] Furthermore, a plurality (numerous) of storage devices 20, 30, 40, etc. are connected to the load verification management system 100 via a network as storage devices from which data is migrated. For example, the storage device 20 installed in a data center DC2 and the storage devices 30, 40 installed in data centers DCn (n is an arbitrary integer) are connected via a WAN (Wide Area Network) 182, and the load verification management system 100 collects performance metrics data from each storage device. The plurality of storage devices 20, 30, 40, etc. are storage devices of different formats. Alternatively, even if they are the same storage device, it is preferable that they are storage devices installed in different environments.

[0019] To explain the specific configuration of the load verification management system 100, the load verification management system 100 includes an expected performance metrics data generation unit 111, a verification performance metrics collection unit 112, an I / O pattern data generation unit 113, a load data storage unit 114, an I / O pattern data storage unit 115, and a performance metrics data storage unit 116.

[0020] The expected performance metrics data generation unit 111 generates expected performance metric data based on the target performance metric data input from the user terminal 190. In the user terminal 190, the performance metric data may be obtained by user input, or may be obtained automatically by a program installed in the user terminal 190. The expected performance metric data generated by the expected performance metrics data generation unit 111 is supplied to the I / O pattern data generation unit 113 . The verification performance metrics collector 112 collects performance metrics data transmitted via the LAN 181 from the storage device 10 under verification in the data center DC1.

[0021] The I / O pattern data generation unit 113 generates and outputs I / O pattern data that matches expected performance metrics data, using the data stored in the load data storage unit 114, the I / O pattern data storage unit 115, and the performance metrics data storage unit 116. A detailed configuration for generating I / O pattern data will be explained in FIG. 2. The I / O pattern data output by the I / O pattern data generator 113 is supplied to the load simulator 19 via the LAN 181 .

[0022] In the load simulator 19, the reproducing unit 19a generates simulated I / O pattern data based on the I / O pattern data supplied from the I / O pattern data generating unit 113. The simulated I / O pattern data generated by the reproducing unit 19a is supplied to the volumes 10a to 10n to be verified via the port 12.

[0023] [Configuration of I / O pattern data generation unit] FIG. 2 shows the configuration of the I / O pattern data generator 113. The I / O pattern data generation unit 113 includes a load data generation unit 201 , a performance metrics data collection unit 202 , an I / O pattern data collection unit 203 , and an I / O pattern data output unit 204 .

[0024] The performance metrics data collection unit 202 collects performance metrics data from the storage device that is the data transfer source. The I / O pattern data collection unit 203 collects I / O pattern data from the storage device that is the data transfer source. In the example of FIG. 2, performance metrics data and I / O pattern data are collected from the storage device 20 installed in the data center DC2. and I / O pattern data collection unit 203 also collects performance metrics data and I / O pattern data from storage devices 30, 40 and the like installed in data centers in other environments.

[0025] The performance metrics data of the multiple storage devices collected by the performance metrics data collector 202 is stored in the performance metrics data storage unit 116 . FIG. 4 shows an example of the configuration of performance metrics data collected by the performance metrics data collector 202 and stored in the performance metrics data storage unit 116. The performance metrics data stored in the performance metrics data storage unit 116 has one record consisting of a device identifier indicating the storage device configuration, a target volume identifier, the time the data was collected, and the actually measured performance metrics data.

[0026] The I / O pattern data of the plurality of storage devices collected by the I / O pattern data collection unit 203 is stored in the I / O pattern data storage unit 115 . The performance metrics data stored in the performance metrics data storage unit 116 and the I / O pattern data stored in the I / O pattern data storage unit 115 are read out by the load data generation unit 201 . FIG. 5 shows an example of the configuration of I / O pattern data collected by the I / O pattern data collection unit 203 and stored in the I / O pattern data storage unit 115. As shown in FIG. Stored in the I / O pattern data storage unit 115 。 One record of I / O pattern data is a set of a device identifier indicating the device configuration, a target volume identifier, the time when the data was collected, and the actually measured I / O pattern data.

[0027] Returning to the explanation of Figure 2, the load data generation unit 201 maps each stored data, generates load data, and stores the generated load data in the load data storage unit 114. When generating this load data, the load data generation unit 201 acquires I / O pattern data from the I / O pattern data storage unit 115, and also acquires performance metrics data corresponding to the collection time of the acquired I / O pattern data from the performance metrics data storage unit 116. Then, the load data generation unit 201 generates load data for a combination of each acquired data, and stores the generated load data in the load data storage unit 114.

[0028] FIG. 6 shows an example of a processing state in which the load data generating unit 201 generates load data. The load data generation unit 201 selects as mapping targets those tables prepared in the performance metrics data storage unit 116 and the I / O pattern data storage unit 115 that have the same value for "device identifier-volume identifier-collection time". The load data generation unit 201 then generates load data in which the "device configuration-performance metrics data-I / O pattern data" of the corresponding mapping target is one record, and stores the generated load data in a table in the load data storage unit 114.

[0029] Returning to the explanation of Figure 2, the I / O pattern data output unit 204 searches the stored data in the load data storage unit 114 based on input indicating a set of device configuration-performance metrics data from the user terminal 190, reads out the corresponding I / O pattern data, and outputs it to the load simulator 19 (Figure 1).

[0030] [Example of hardware configuration for load verification management system] FIG. 3 shows an example of the hardware configuration of the load verification management system 100. As shown in FIG. The load verification management system 100 shown in FIG. 3 is an example configured by a computer. That is, the load verification management system 100 configured as a computer comprises a processor 101, a memory 102, a storage device 103, and an input / output device 104, and each component is connected by a bus line to enable data transfer between them.

[0031] Then, the processor 101 executes the program stored in the storage device 103, and the storage device 103 stores the following: Expected performance metrics data generation unit 111, Verification performance metrics collection unit 112, and an I / O pattern data generation unit 113. A part of the storage area of the storage device 103 is used to configure a load data storage unit 114, an I / O pattern data storage unit 115, and a performance metrics data storage unit .

[0032] Note that the load verification management system 100 configured as a computer having a processor 101 is just an example, and other configurations are also possible. For example, part or all of the load verification management system 100 may be realized by dedicated hardware such as an FPGA (Field Programmable Gate Array) or an ASIC (Application Specific Integrated Circuit). The load verification management system 100 may also be a so-called cloud computer provided on a server accessible via the Internet or the like. In this case, each storage unit, such as the load data storage unit 114, is also stored on the server. Alternatively, each storage unit, such as the load data storage unit 114, may be provided on the server, and a computer constituting the load verification management system 100 may read the stored data from the server and perform the load verification management process.

[0033] [Processing performed by the expected performance metrics data generation unit] FIG. 7 is a flowchart showing an example of processing performed by the expected performance metrics data generation unit 111. First, the expected performance metrics data generator 111 receives as input from the user terminal 190 the target of metrics to be generated (step S11). Then, the expected performance metrics data generation unit 111 performs an expected performance metrics data generation process for generating performance metrics data indicating the load result expected for the target of load verification in response to the received input (step S12). The performance metrics data generated in step S12 is supplied to the I / O pattern data generation unit 113.

[0034] [Processing performed by the I / O pattern data generator] FIG. 8 is a flowchart showing an example of processing performed by the I / O pattern data generating unit 113. The I / O pattern data generation unit 113 receives the expected performance metrics data supplied from the expected performance metrics data generation unit 111 as an input (step S21). Then, the I / O pattern data generation unit 113 performs an I / O pattern data generation process for generating I / O pattern data that will generate corresponding performance metrics data for the storage system that is the target of load verification for the received performance metrics data (step S22). In other words, the I / O pattern data generation unit 113 extracts I / O pattern data that will be performance metrics data equivalent to the expected performance metrics data from the various load data stored in the load data storage unit 114, and generates this I / O pattern data.

[0035] [Processing performed by the load data generator] FIG. 9 is a flowchart showing an example of processing performed by the load data generating unit 201. The load data generation unit 201 acquires I / O pattern data from the I / O pattern data storage unit 115 (step S31). Then, the load data generation unit 201 acquires performance metrics data corresponding to the collection time of the I / O pattern data acquired in step S31 from the performance metrics data storage unit 116 (step S32). Thereafter, the load data generating unit 201 generates load data for each combination of the acquired data (step S33). Then, the load data generating unit 201 registers (stores) the load data generated in step S33 in the load data storage unit 114 (step S34).

[0036] [Processing performed by the load simulator] FIG. 10 is a flowchart showing an example of processing performed by the reproducing unit 19a of the load simulator 19. The load simulator 19 receives as input the I / O pattern data transmitted from the I / O pattern data generation unit 113 (step S41). Then, the reproduction unit 19a of the load simulator 19 simulates an I / O pattern to a volume to be subjected to load verification from the input pattern data, and executes processing to apply input and output loads (step S42).

[0037] [Processing performed by the verification performance metrics collection unit] FIG. 11 is a flowchart showing an example of processing performed by the verification performance metrics collector 112. The verification performance metrics collection unit 112 performs a verification performance metrics collection process to receive the performance metrics data of the actually measured storage system and all volumes in the storage device 10 from the storage device 10 that is the load verification target (step S 51 ). The load verification management system 100 compares the performance metrics data received by the verification performance metrics collection unit 112 with the expected performance metrics data generated by the expected performance metrics data generation unit 111 to verify whether the storage device 10 is performing as expected. The verification result is displayed on the user terminal 190, for example.

[0038] [Effects of the load verification system] As explained above, the load verification system of this example can provide input / output pattern data that reproduces the expected performance metrics for the volume being verified. This means that the load generated in the source environment can be reproduced in the destination environment, shortening the verification period and enabling the timely provision of application services to be operated.

[0039] In other words, in the past, when the models of the source storage device and the destination storage device were different, it was impossible to identify the I / O pattern data that would generate the expected performance metrics for the destination volume, making reproduction impossible.For this reason, in the past, in order to reproduce trends, a load was applied to the destination volume using I / O pattern data extracted randomly or based on specific rules from load data accumulated by trial and error, and the performance impact was measured, and the performance impact of other volumes at the same time was confirmed.

[0040] In contrast, in the case of the load verification system of this example, combinations of performance metrics data and I / O pattern data are collected in advance from multiple storage devices 20, 30, 40, etc. that have different formats and installation environments, and the I / O pattern data that matches the expected performance is extracted and simulated on the storage device 10 to be verified. Therefore, the load verification system of this embodiment can reduce the time required for I / O load verification using a large number of test patterns, thereby achieving the effect of shortening the verification period.

[0041] Furthermore, in the case of the load verification system of this example, the load verification management system 100 is equipped with a load data storage unit 114, an I / O pattern data storage unit 115, and a performance metrics data storage unit 116, and is able to extract I / O pattern data that matches expected performance from the load data, thereby enabling verification against various expected performances.

[0042] [Variations] The embodiment examples described so far have been described in detail to clearly explain the present invention, and are not necessarily limited to those having all of the configurations described. For example, in the above-described embodiment, a storage device installed in a data center was used as the device having the data storage area to be verified, but it may also be used to verify a data storage area of a predetermined volume provided by a cloud solution or hybrid cloud solution. In other words, when data storage areas provided in various devices such as servers connected via the Internet are used as the data storage areas, the volumes may be used as the verification targets.

[0043] In this case, the storage devices 20, 30, 40, . . . that collect performance metrics data and I / O pattern data may also be volumes provided by a cloud solution or a hybrid cloud solution.

[0044] Furthermore, when verifying a storage device 10 installed in a data center as shown in FIG. 1, it is preferable to install the load simulator 19 in the same location as the storage device 10 and supply simulated I / O pattern data, but the load simulator 19 may also be installed in a location other than the location where the data storage area is installed.

[0045] In the above-described embodiment, the I / O pattern data generation unit 113 stores I / O pattern data that serve as various performance metrics from multiple storage devices 20, 30, 40, etc., and obtains I / O pattern data that provides equivalent performance from the stored data. Alternatively, the I / O pattern data generation unit 113 may generate I / O pattern data that corresponds to expected performance metrics by performing some kind of calculation processing. In the above embodiment, the processing up to the point where the verification performance metrics collector 112 collects performance metrics as a result of applying a load has been described, but the load verification management system 100 may also perform verification based on the collected data. Alternatively, the collected performance metrics may be sent to the user terminal 190, and verification may be performed on the user terminal 190 side.

[0046] 1 to 3, only control lines and information lines that are considered necessary for explanation are shown, and not all control lines and information lines in the product are necessarily shown. In reality, it can be assumed that almost all components are interconnected. Furthermore, the processing flows shown in the flowcharts of FIGS. 7 to 11 are also examples, and as long as the processing results are the same, the order of some of the processing may be changed or multiple processes may be executed simultaneously.

[0047] The load verification management system 100 described in this embodiment can be configured by installing a load verification management program in a general computer, as described in Fig. 3. In this case, the program for realizing each processing function may be stored in a storage device within the computer, or may be stored and transferred on a recording medium such as an external memory, IC card, SD card, or optical disk. [Explanation of symbols]

[0048] 10...storage device (storage device to be verified), 10a to 10n...volume, 11...interface, 12...port, 19...load simulator, 19a...reproduction unit, 20, 30...storage device, 100...load verification management system, 101...processor, 102...memory, 103...storage device, 104...input / output device, 111...expected performance metrics data generation unit, 112...verification performance metrics collection unit, 113...I / O pattern data generation unit, 114...load data storage unit, 115...I / O pattern data storage unit, 116...performance metrics data storage unit, 181...LAN, 182...WAN, 190...user terminal, 201...load data generation unit, 202...performance metrics data collection unit, 203...I / O pattern data collection unit, 204...I / O pattern data output unit, DC1 to DCn...data center

Claims

1. A load verification system that performs load performance verification of a predetermined data storage area, a verification performance metrics collection unit that acquires performance metrics of the data storage area and its volume that are the target of load verification; an expected performance metrics data generation unit that generates performance metrics data that results in an expected load for the volume that is the target of load verification; an I / O pattern data generation unit that generates input / output pattern data that generates a load that results in performance metrics data equivalent to the expected performance indicated by the expected performance metrics data, Based on the input / output pattern data generated by the I / O pattern data generation unit, input / output is performed to the volume that is the load verification target, and the verification performance metrics collection unit collects performance metrics as a result of applying the load. Load verification system.

2. a performance metrics data collection unit that collects performance metrics of components used to execute input / output processing for a plurality of data storage areas; a performance metrics data storage unit that stores the performance metrics data collected by the performance metrics data collection unit; an I / O pattern data collection unit that monitors input / output data in a plurality of data storage areas and collects input / output pattern data; an I / O pattern data storage unit that stores the input / output pattern data collected by the I / O pattern data collection unit; a load data generation unit that generates load data in which components of the data storage area, their performance metrics, and input / output pattern data are mapped based on the data stored in the performance metrics data storage unit and the I / O pattern data storage unit; a load data storage unit that stores the load data generated by the load data generation unit; The load verification system of claim 1 .

3. The expected performance metrics data generation unit generates expected performance metrics data based on target performance metric data input from an external terminal. The load verification system of claim 1 .

4. The data storage area is a storage device having a predetermined volume. The load verification system of claim 1 .

5. Further, a load simulator is provided for reproducing input / output to the volume that is the object of load verification based on the input / output pattern data generated by the I / O pattern data generating unit. The load verification system of claim 1 .

6. The data storage area is a data storage area of a predetermined volume provided by a hybrid cloud solution. The load verification system of claim 1 .

7. A load verification method in which a computer performs load performance verification on a data storage area having a predetermined volume provided by a storage device or a hybrid cloud solution, The process executed by the computer is a verification performance metrics collection process for acquiring performance metrics of the data storage area and its volume that are the load verification target; an expected performance metrics data generation process for generating performance metrics data that results in an expected load for the volume that is the target of load verification; an I / O pattern data generation process for generating I / O pattern data that generates a load that results in performance metrics data equivalent to the expected performance indicated by the expected performance metrics data; a reproduction process for reproducing input / output to the volume that is the load verification target based on the input / output pattern data generated by the I / O pattern data generation process, The computer collects performance metrics as a result of applying a load to the volume that is the target of load verification by the verification performance metrics collection process. Load validation method.

Images