WO2024066597A1

WO2024066597A1 - Data storage method and apparatus

Info

Publication number: WO2024066597A1
Application number: PCT/CN2023/104709
Authority: WO
Inventors: 王顺卓
Original assignee: 华为云计算技术有限公司
Priority date: 2022-09-29
Filing date: 2023-06-30
Publication date: 2024-04-04
Also published as: CN117827818A

Abstract

The present application relates to the field of data processing, and in particular to a data storage method and apparatus. The method comprises: providing a configuration interface, the configuration interface being used for a user to configure a load characteristic of a data table as a read-intensive load or a write-intensive load; when the configuration interface indicates that the load characteristic of the data table configured by the user is the read-intensive load, instructing a storage apparatus to store data in the data table according to a first index engine, the first index engine matching the read-intensive load; and when the configuration interface indicates that the load characteristic of the data table configured by the user is the write-intensive load, instructing the storage apparatus to store the data in the data table according to a second index engine, the second index engine matching the write-intensive load. According to the method, index engines of a data table can be configured by a user, so that the index engines of the data table match the load characteristics caused by a service, thereby improving the access performance of the data table.

Description

Data storage method and device

This application claims priority to the Chinese patent application filed with the State Intellectual Property Office of China on September 29, 2022, with application number 202211202134.2 and application name “A Data Storage Method and Device”, the entire contents of which are incorporated by reference in this application.

Technical Field

The present application relates to the field of data processing technology, and in particular to a data storage method and device.

Background technique

The database is a warehouse for storing data, storing a large amount of data. In order to ensure query efficiency, the data in the database needs to be organized according to a certain structure, that is, an index engine is used to store data. At present, the commonly used index engines for databases include log structured merge tree (LSM tree) and B+ tree.

Different access operations to the database bring different load characteristics to the database. Different index engines adapt to different load characteristics. For example, the log structure merge tree structure is more suitable for the load characteristics when the write operation frequency is high, and the B+ tree structure is more suitable for the load characteristics when the read operation frequency is high.

Different users may have different access operations to the database, and the load characteristics they bring are also different. Therefore, no matter which index engine the database uses, it is difficult to adapt to the load characteristics brought by the access operations of multiple users, resulting in a decrease in the access performance of the database.

Summary of the invention

The embodiments of the present application provide a data storage method and device, which can be configured by a user or adjust the index engine of a data table according to changes in the load characteristics of the data table.

In a first aspect, a data storage method is provided, which is applied to a control device in a storage system, wherein the storage system also includes a storage device, which stores a user's data table; the method includes: providing a configuration interface, the configuration interface being used for the user to configure the load characteristics of the data table as a read-intensive load or a write-intensive load; when the configuration interface indicates that the user configures the load characteristics of the data table as a read-intensive load, instructing the storage device to store the data in the data table according to a first index engine; the first index engine matches the read-intensive load; when the configuration interface indicates that the user configures the load characteristics of the data table as a write-intensive load, instructing the storage device to store the data in the data table according to a second index engine; the second index engine matches the write-intensive load.

Through this method, users can configure the index engine of their data table. Thus, users can adjust the index engine of the data table according to changes in the business served by the data table, so that the index engine matches the load characteristics caused by the changed business, thereby improving the access performance of the data table.

In a possible implementation, the first index engine includes at least a B+ tree structure, and the second index engine includes at least a log structure merged with a tree structure.

In this implementation, the B+ tree structure belongs to a read-friendly index engine, and the first index engine includes the B+ tree structure, which can improve the matching degree between the first index engine and read-intensive loads. The log structure merge tree structure belongs to a write-friendly index engine, and the second index engine includes the log structure merge tree structure, which can improve the matching degree between the second index engine and write-intensive loads.

In a possible implementation, before instructing the storage device to store the data in the data table according to the first index engine, the index engine in the data table is the second index engine; instructing the storage device to store the data in the data table according to the first index engine includes: instructing the storage device to first migrate the index engine in the data table from the second index engine to the third index engine, the structure of the third index engine being between the structure of the first index engine and the structure of the second index engine; and then instructing the storage device to migrate the index engine of the data table from the third index engine to the first index engine. The third index engine can be called a hybrid index engine.

In this implementation, the index engine of the data table can be switched from a write-friendly index engine to a hybrid index engine, and then from the hybrid index engine to a read-friendly index engine, thereby realizing a gradual switching of the index engine and avoiding the index engine change overhead caused by switching between index engines with large structural differences.

In one possible implementation, the data table has a local secondary index, and the configuration interface is also used for the user to configure the load characteristics of the local secondary index as a read-intensive load or a write-intensive load; when the configuration interface instructs the user to configure the load characteristics of the local secondary index as a read-intensive load, the storage device is instructed to store the data under the local secondary index according to the first index engine; when the configuration interface instructs the user to configure the load characteristics of the local secondary index as a write-intensive load, the storage device is instructed to store the data under the local secondary index according to the second index engine.

In this implementation, users can configure the index engine of the local secondary index in their data table. Thus, users can adjust the index engine of the local secondary index according to changes in the business served by the local secondary index, so that the index engine matches the load characteristics caused by the changed business, thereby improving the access performance of the local secondary index.

In a possible implementation, the data in the data table is stored in the form of key-value pairs.

In this implementation, the method can be applied to a key-value database, and can improve the access performance of the key-value database.

In a second aspect, a data storage method is provided, which is applied to a control device in a storage system, wherein the storage system also includes a storage device, which stores a data table; the method includes: when the index engine of the data table is a fourth index engine, the control device monitors the operation amplification of the data table, the operation amplification including read amplification of reading data from the data table or write amplification of writing data to the data table; when the operation amplification includes read amplification and the read amplification is greater than a first threshold value, instructing the storage device to store the data in the data table according to the fifth index engine; wherein the matching degree between the fifth index engine and the read-intensive load is greater than the matching degree between the fourth index engine and the read-intensive load; when the operation amplification includes write amplification and the write amplification is greater than a second threshold value, instructing the storage device to store the data in the data table according to the sixth index engine; wherein the matching degree between the sixth index engine and the write-intensive load is greater than the matching degree between the fourth index engine and the write-intensive load.

This method can monitor the read amplification or write amplification of the data table, and based on the read amplification or write amplification, determine whether the current index engine of the data table matches the load characteristics of the data table, as well as the direction of change of the index engine, so that the index engine can be adjusted in the direction of matching the load characteristics of the data table, so that the index engine matches the load characteristics of the data table, thereby improving the access performance of the data table.

In one possible implementation, operation amplification includes both read amplification and write amplification; when operation amplification includes read amplification and the read amplification is greater than a first threshold, instructing the storage device to store the data in the data table according to the fifth index engine, including: when the read amplification is greater than the first threshold and the write amplification is less than a third threshold, instructing the storage device to store the data in the data table according to the fifth index engine.

In this implementation, when read amplification is relatively large and write amplification is relatively small, the index engine of the data table can be adjusted towards a read-friendly index engine, thereby balancing read amplification and write amplification and improving the overall access performance of the data table.

In one possible implementation, operation amplification includes both read amplification and write amplification; when operation amplification includes write amplification, and the write amplification is greater than a second threshold, instructing the storage device to store the data in the data table according to the sixth index engine, including: when the write amplification is greater than the second threshold and the read amplification is less than the fourth threshold, instructing the storage device to store the data in the data table according to the sixth index engine.

In this implementation, when write amplification is relatively large and read amplification is relatively small, the index engine of the data table can be adjusted towards a write-friendly index engine, thereby balancing read amplification and write amplification and improving the overall access performance of the data table.

In a possible implementation, the fifth indexing engine includes at least a B+ tree structure, and the sixth indexing engine includes at least a log structure merged with a tree structure.

In this implementation, the B+ tree structure belongs to a read-friendly index engine, and the fifth index engine includes the B+ tree structure, which can improve the matching degree between the fifth index engine and read-intensive loads. The log structure merge tree structure belongs to a write-friendly index engine, and the sixth index engine includes the log structure merge tree structure, which can improve the matching degree between the sixth index engine and write-intensive loads.

In one possible implementation, the data table has a local secondary index LSI, and the operation amplification is the amplification generated by operating the data under the local secondary index; instructing the storage device to store the data in the data table according to the fifth index engine, including: instructing the storage device to store the data under the local secondary index according to the fifth index engine; or, instructing the storage device to store the data in the data table according to the sixth index engine, including: instructing the storage device to store the data under the local secondary index according to the sixth index engine.

In this implementation, the index engine of the local secondary index in the data table can be adjusted according to the operation amplification of the index engine of the local secondary index so that the index engine matches the load characteristics of the local secondary index, thereby improving the access performance of the local secondary index in the data table.

According to a third aspect, a data storage device is provided, which is configured in a control device in a storage system. The storage system also includes a storage device, which stores a user's data table. The data storage device includes: a providing module, which is used to provide a configuration interface, and the configuration interface is used for the user to configure the load characteristics of the data table as a read-intensive load or a write-intensive load; an indicating module, which is used to instruct the storage device to store the data in the data table according to a first index engine when the configuration interface indicates that the user configures the load characteristics of the data table as a read-intensive load; the first index engine matches the read-intensive load; the indicating module is also used to instruct the storage device to store the data in the data table according to a second index engine when the configuration interface indicates that the user configures the load characteristics of the data table as a write-intensive load; the second index engine matches the write-intensive load.

In one possible implementation, before instructing the storage device to store the data in the data table according to the first index engine, the index engine in the data table is the second index engine; the instruction module is also used to: instruct the storage device to first migrate the index engine in the data table from the second index engine to the third index engine, and the structure of the third index engine is between the structure of the first index engine and the structure of the second index engine; and then instruct the storage device to migrate the index engine of the data table from the third index engine to the first index engine.

In one possible implementation, the data table has a local secondary index, and the configuration interface is also used for the user to configure the load characteristics of the local secondary index as a read-intensive load or a write-intensive load; the indication module is also used to: when the configuration interface instructs the user to configure the load characteristics of the local secondary index as a read-intensive load, instruct the storage device to store the data under the local secondary index according to the first index engine; when the configuration interface instructs the user to configure the load characteristics of the local secondary index as a write-intensive load, instruct the storage device to store the data under the local secondary index according to the second index engine.

In a fourth aspect, a data storage device is provided, which is configured in a control device of a storage system, wherein the storage system further comprises a storage device, a storage device A storage device stores a data table; the data storage device includes: a monitoring module, which is used to control the device to monitor the operation amplification of the data table when the index engine of the data table is a fourth index engine, the operation amplification including read amplification of reading data from the data table or write amplification of writing data to the data table; an indication module, which is used to instruct the storage device to store the data in the data table according to a fifth index engine when the operation amplification includes read amplification and the read amplification is greater than a first threshold value; wherein the matching degree between the fifth index engine and the read-intensive load is greater than the matching degree between the fourth index engine and the read-intensive load; the indication module is also used to instruct the storage device to store the data in the data table according to a sixth index engine when the operation amplification includes write amplification and the write amplification is greater than a second threshold value; wherein the matching degree between the sixth index engine and the write-intensive load is greater than the matching degree between the fourth index engine and the write-intensive load.

In a possible implementation, the operation amplification includes both read amplification and write amplification; the indication module is used to: when the read amplification is greater than a first threshold and the write amplification is less than a third threshold, instruct the storage device to store the data in the data table according to the fifth index engine.

In a possible implementation, the operation amplification includes both read amplification and write amplification; the indication module is used to: when the write amplification is greater than the second threshold and the read amplification is less than the fourth threshold, instruct the storage device to store the data in the data table according to the sixth index engine.

In one possible implementation, the data table has a local secondary index LSI, and the operation amplification is the amplification generated by operating the data under the local secondary index; the indication module is used to: instruct the storage device to store the data under the local secondary index according to the fifth index engine; or, instruct the storage device to store the data under the local secondary index according to the sixth index engine.

In a fifth aspect, a computing device cluster is provided, comprising at least one computing device, each computing device comprising a processor and a memory; the processor of the at least one computing device is used to execute instructions stored in the memory of the at least one computing device, so that the computing device cluster performs the method provided in the first aspect or the method provided in the second aspect.

In a sixth aspect, a computer program product comprising instructions is provided. When the instructions are executed by a computing device cluster, the computing device cluster executes the method provided in the first aspect or the method provided in the second aspect.

In a seventh aspect, a computer-readable storage medium is provided, comprising computer program instructions. When the computer program instructions are executed by a computing device cluster, the computing device cluster executes the method provided in the first aspect or the method provided in the second aspect.

The data storage method and device provided in the embodiments of the present application allow the user to configure the index engine of the data table, or adjust the index engine of the data table according to the load characteristics of the data table, so that the index engine of the data table matches the load characteristics of the data table, thereby improving the access performance of the data table.

BRIEF DESCRIPTION OF THE DRAWINGS

FIG1A is a schematic diagram of the structure of a log structure merge tree;

FIG1B is a schematic diagram of the structure of a B+ tree;

FIG2 is a schematic diagram of the structure of a storage system provided in an embodiment of the present application;

FIG3 is a schematic diagram of the structure of an initial load characteristic configuration submodule provided in an embodiment of the present application;

FIG4 is a flow chart of a data storage solution provided in an embodiment of the present application;

FIG5 is a schematic diagram of an index engine provided in an embodiment of the present application;

FIG6 is a flow chart of a data storage solution provided in an embodiment of the present application;

FIG7 is a flow chart of a data storage solution provided in an embodiment of the present application;

FIG8 is a flow chart of a data storage method provided in an embodiment of the present application;

FIG9 is a flow chart of a data storage method provided in an embodiment of the present application;

FIG10 is a schematic diagram of the structure of a data storage device provided in an embodiment of the present application;

FIG11 is a schematic diagram of the structure of a data storage device provided in an embodiment of the present application;

FIG12 is a schematic diagram of the structure of a computing device provided in an embodiment of the present application;

FIG13 is a schematic diagram of the structure of a computing device cluster provided in an embodiment of the present application;

FIG14 is a schematic diagram of the structure of a computing device cluster provided in an embodiment of the present application;

FIG15 is a schematic diagram of the structure of a computing device provided in an embodiment of the present application;

FIG16 is a schematic diagram of the structure of a computing device cluster provided in an embodiment of the present application;

FIG. 17 is a schematic diagram of the structure of a computing device cluster provided in an embodiment of the present application.

Detailed ways

The technical solutions in the embodiments of the present application will be described below in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments. Among them, in the embodiments of the present application, "multiple" means "at least two".

A key-value store stores data in the form of key-value pairs, where the key is the unique identifier of the data. Value refers to the data content, which can be anything from simple to complex composite objects. Key-value databases store data in a simple form, provide distributed processing capabilities, and have advantages such as fast response. In addition, key-value databases are non-relational databases (not only SQL, NoSQL) and can handle large-scale data storage. Therefore, key-value databases have been widely used in computer systems, especially in the field of cloud storage.

The key-value database can serve multiple users. Exemplarily, the user can be a tenant of cloud storage. The user can query data in the database through an index engine. Among them, the index engine is a data storage structure for efficiently querying data. In the embodiment of the present application, the index engine can also be called a data storage structure.

Specifically, users can create data tables in the key-value database to store and manage their data. The data table uses a local primary index (LPI) to establish a key-to-value mapping, and uses an index engine to store the key-to-value mapping relationship to store the data in the data table. The key is used to identify data and can be the name of the data. Value represents data and can be defined according to business needs, such as a student's grades in a subject or multiple subjects. Using a local primary index to establish a key-to-value mapping means storing keys and values in accordance with the local primary index data storage method.

In some embodiments, a data table may be composed of one or more data partition instances. A partition instance may be used to store data in a specified key range. The data in the specified key range may be referred to as data in the partition instance. A partition instance may establish a mapping from key to value in the corresponding key range through a local primary index to store data in the partition instance. When a data table is composed of multiple data partition instances, the index engines used by the local primary indexes of the multiple data partition instances may be the same, that is, multiple data partition instances in the same data table may use the same index engine to store data. Therefore, the index engine of the local primary index of the data partition instance in the data table may be referred to as the index engine of the data table.

In some embodiments, a value may include multiple pieces of information (for example, a value is an object of complex composite information). In order to improve query efficiency, a local secondary index (LSI) may be constructed to establish a mapping relationship between a certain piece of information among the multiple pieces of information and a key. The piece of information may be referred to as a sub-value. The local secondary index also uses an index engine to store the mapping relationship between the sub-value and the key.

When querying the value corresponding to a certain key or keys, you can first filter the key in the local secondary index by the sub-value corresponding to the key. Then, query the value in the local primary index by the filtered key, thereby improving the query efficiency. Take the data table or data partition example that stores student scores as an example, where the key is the student's student ID, and the value includes the student's scores in multiple subjects such as language, mathematics, English, physics, and chemistry. The local primary index constructs a mapping from student ID to scores in multiple subjects. The local secondary index constructs a mapping from Chinese scores to student IDs, that is, the sub-value is the Chinese score. It can be set that the user needs to query the scores of each subject for students whose Chinese scores are greater than 90 points. Then, in the local secondary index, you can first filter out the student IDs corresponding to the Chinese scores greater than 90 points, and then, in the local primary index, query the value according to the filtered student IDs, so as to obtain the scores of each subject for students whose Chinese scores are greater than 90 points. Therefore, the query efficiency can be improved through the local secondary index.

For a data table, one or more local secondary indexes can be constructed. The corresponding sub-values in different local secondary indexes can be different or partially the same.

In the following description, the local primary index may be referred to as the primary index, and the local secondary index may be referred to as the secondary index. In addition, when no special distinction is made between the local primary index and the local secondary index, they may be referred to as indexes.

The above example introduces indexes in a key-value database. Next, we will introduce the load characteristics of a key-value database.

Access operations to a database may include read operations and write operations, and accordingly, load characteristics may include read-intensive load, write-intensive load, and mixed load. Mixed load is a load characteristic between read-intensive load and write-intensive load, which indicates that the frequency of read operations and write operations is not much different, or in other words, the read operations and write operations are relatively balanced.

Among them, when the ratio of the occurrence frequency of read operations to the occurrence frequency of write operations is greater than a threshold value A1, the load characteristic is specifically a read-intensive load. Threshold A1 is a preset value. Exemplarily, threshold A1 is greater than or equal to 4. In one example, threshold A1 is 9. When the ratio of the occurrence frequency of write operations to the occurrence frequency of read operations is greater than threshold A2, the load characteristic is a write-intensive load. Threshold A2 is a preset value. Exemplarily, threshold A2 is greater than or equal to 4. In one example, threshold A2 is 9. When the ratio of the occurrence frequency of read operations to the occurrence frequency of write operations is not greater than threshold A1, and the ratio of the occurrence frequency of write operations to the occurrence frequency of read operations is not greater than threshold A2, the load characteristic is a mixed load.

Among them, the access operation to the index, that is, the access operation performed in the index engine used by the index, will bring load characteristics to the index. Different access operations cause the index to carry different load characteristics. Among them, the load characteristics brought by the access operation to the local primary index are called the primary index load characteristics, and the load characteristics brought by the access operation to the local secondary index are called the secondary index load characteristics.

The load characteristics of a data table depend on the type of business it serves and the type of user behavior (such as user access to the data table and querying data in the data table using local secondary indexes). During peak hours (such as 8pm to 10pm), data tables are read and written frequently, and the load characteristics at this time are mixed loads. During low business peak hours (such as 0am to 6am), users dump and back up the data in the data table, and the data in the data table is read frequently. At this time, the data table is read-intensive. For data tables serving the data backup business, their load characteristics are usually write-intensive. Among them, when using the data table serving the data backup business to restore data for the data table that has lost data, a large amount of data needs to be read from the data table serving the data backup business, and the load characteristics of the data table serving the data backup business are converted to read-intensive loads.

Local secondary indexes provide fast access to data in a data table. Different local secondary indexes may record different sub-values in a data table. Therefore, the data in the data table can be queried through different sub-values. Therefore, the load characteristics of different local secondary indexes in a data table depend on the user's query behavior. For example, for a data table containing two local secondary indexes, the local secondary indexes are recorded as LSI 1 and LSI 2 respectively. In a certain period of time, if the user only accesses the data table with the sub-value recorded in LSI 1 as the query condition, LSI 1 is a mixed load and LSI 2 is a write-intensive load. Conversely, LSI 1 is a write-intensive load and LSI 2 is a mixed load.

It is not difficult to understand that the access operation for the secondary index is not static. Therefore, the load characteristics of the same secondary index at different times may also be different.

Therefore, both the load characteristics of the primary index and the load characteristics of the secondary index usually change dynamically.

In addition, in the following description, when no special distinction is made between the primary index load characteristics and the secondary index load characteristics, they may be referred to as data table load characteristics for short.

Some index engines are suitable for read operations, which makes the read magnification small, the read latency low, and the read operation experience good. This type of index engine can be called a read-friendly index engine. Some index engines are suitable for write operations, which makes the write magnification small, the write latency low, and the write operation experience good. This type of index engine can be called a write-friendly index engine.

Specifically, the read-friendly index engine uses append to write data, which has fast writing speed, low writing latency and high writing performance. However, the append writing method delays the update of data in the data table, which results in multiple historical versions of key values being retained in the data table, increasing read amplification and causing high read latency.

Among them, the log structured merge tree (LSM tree) structure is a typical write-friendly index engine. The LSM tree stores data in the form of logs. As shown in FIG1A , the LSM tree includes a data area and an index area (manifest). Among them, the data area is the area in the LSM tree where data is stored. Among them, the data area can be located on a hard disk to achieve persistent storage of data. The data area includes multiple storage layers from top to bottom (C1 layer, C2 layer, C3 layer, C4 layer, C5 layer, and C6 layer as shown in FIG1A ). Among them, the storage space of the upper layers in the multiple storage layers is smaller, and the storage space of the lower layers is larger. When writing data, the data is first written to the top layer, that is, the C1 layer. When the amount of data in the C1 layer reaches the preset value D1, the data of the C1 layer and the next layer of C1 (that is, the C2 layer) are merged (compaction), and the merged data is transferred to the C2 layer. When the amount of data in layer C2 reaches the preset value D2, the data in layer C2 and the next layer (C3) of C2 are merged, and the merged data is transferred to layer C3, and so on, so that the old data is continuously transferred to the lower layer and the new data can be continuously written to the upper layer. In addition, the index area can be used to accelerate the location of which storage layer a certain key value is in.

The read-friendly index engine uses a timely update method to store data. In a read-friendly index engine, only a single key value is retained. Therefore, the read-friendly index engine is conducive to data reading and is more suitable for read operations. When writing data to a read-friendly index engine, it is necessary to read the key value of the previous version, update the key value, and then write the updated key value, resulting in a large write overhead. Therefore, the read-friendly index engine is not suitable for write operations, resulting in write amplification.

Among them, the B+ tree structure is a typical read-friendly index engine. As shown in Figure 1B, the data area consists of leaf nodes, and key values are stored in the leaf nodes in an orderly manner. The B+ tree structure has an index area, which is used to accelerate the location of which leaf node a certain key value is in.

Since the load characteristics of a data table may change from read-intensive to write-intensive, or from write-intensive to read-intensive, whether a read-friendly or write-friendly index engine is used for a data table, there may be a mismatch between the index engine and the load characteristics, which will increase access latency, affect business operations, and user experience.

In view of the above situation, the present application provides a data storage solution that can provide an index engine configuration interface so that users can configure the index engine of the data table at any time. Then, the data in the data table can be stored according to the data storage structure configured by the user. It is not difficult to understand that the load characteristics of the data table are related to the business served by the data table. For example, for a data table serving a data backup business, its load characteristics are usually write-intensive loads. Among them, when using a data table serving a data backup business to restore data for a data table with lost data, a large amount of data needs to be read from the data table serving the data backup business, and the load characteristics of the data table serving the data backup business are converted to read-intensive loads. The user can set or predict what kind of business the data table serves at what time. In this way, the user can configure an index engine that matches the load characteristics brought by the switched business when or before the data table switches the business it serves, so that the data table can provide better access performance and reduce read or write delays.

Next, an example is given of the data storage solution provided in the embodiment of the present application.

First, the present application embodiment provides a storage system 100 that can be used to implement a data storage solution. As shown in FIG. 2 , the storage system 100 includes a control device 110 and a storage device 120 .

The storage device 120 is a device or equipment for persistently storing data. In some embodiments, the storage device 120 may be a hard disk, such as a solid state disk (SSD). In other embodiments, the storage device 120 may be other forms of devices or equipment with a persistent data storage function. The embodiments of the present application do not specifically limit the specific implementation form of the storage device 120.

As shown in FIG. 2 , the storage device 120 may include multiple data tables such as a data table T1 and a data table T2. Among them, the data table T1 and the data table T2 may belong to a database. The database may specifically be a key-value database. Among them, some of the multiple data tables belong to one user, and some of the data tables may belong to different users. Among them, as shown in FIG. 2 , a data table, such as the data table T1, may include a local primary index. In some embodiments, the data table T1 may also include a local secondary index B1 and a local secondary index B2.

The control module 110 is a module or component with data processing capabilities. In some embodiments, the control module 110 may be a physical device, such as a server or a processor. In some embodiments, the control module 110 may be a virtual device, such as a virtual machine (VM) or a container. The embodiment of the present application does not specifically limit the specific implementation form of the control module 110.

The control module 110 is used to control or adjust the index engine of the data table in the storage device 120. As shown in FIG2 , the control module 110 may include a configuration module 111 and a processing module 112. The configuration module 111 may be used for the user to configure the load characteristics of the data table. The processing module 112 may configure an index engine that matches the load characteristics based on the load characteristics configured by the user, and instruct the storage device 120 to store data according to the index engine. More specifically, the configuration module 111 may provide a configuration interface to the user so that the user can input the load characteristics through the configuration interface. The processing module 112 may obtain the load characteristics input by the user from the configuration module 111, and configure an index engine that matches the load characteristics, thereby instructing the storage device 120 to store the data in the user's data table according to the index engine.

In some embodiments, as shown in FIG. 2 , the configuration module 111 may include an initial load characteristic configuration submodule 111A, and the processing module 112 may include an index engine initialization submodule 112A. The initial load characteristic configuration submodule 111A may provide an initial load characteristic configuration interface to the user when the user creates a data table in the database 121. The user may input the initial load characteristic through the configuration interface. The index engine initialization submodule 112A may configure an index engine that matches the initial load characteristic based on the initial load characteristic. Afterwards, the index engine initialization submodule 112A may instruct the storage device 120 to use the configured index engine as the initialization index engine for the data table newly created by the user.

In some embodiments, as shown in FIG3 , the initial load characteristic configuration submodule 111A may include a main index initial load characteristic configuration submodule 111A1. The main index initial load characteristic configuration submodule 111A1 may provide a main index initial load characteristic configuration interface to the user when the user creates a data table in the database 121. The user may input the main index initial load characteristic through the configuration interface. The index engine initialization submodule 112A may configure an index engine that matches the main index initial load characteristic based on the main index initial load characteristic. Afterwards, the index engine initialization submodule 112A may instruct the storage device 120 to use the configured index engine as the initial main index of the data table newly created by the user.

In some embodiments, as shown in FIG3 , the initial load characteristic configuration submodule 111A may include a secondary index initial load characteristic configuration submodule 111A2. The secondary index initial load characteristic configuration submodule 111A2 may provide a secondary index initial load characteristic configuration interface to the user when the user creates a secondary index in the data table. The user may input the secondary index initial load characteristic through the configuration interface. The index engine initialization submodule 112A may configure an index engine that matches the secondary index initial load characteristic based on the secondary index initial load characteristic. Afterwards, the index engine initialization submodule 112A may instruct the storage device 120 to use the configured index engine as the initial secondary index of the data table newly created by the user.

Returning to FIG. 2 , in some embodiments, the configuration module 111B may include a load characteristic adjustment submodule 111B, and the processing module 112 may include an index engine adjustment submodule 112B. The load characteristic adjustment submodule 111B may provide a load characteristic adjustment interface to the user after the data table is created. The user may input a new load characteristic through the adjustment interface. The index engine adjustment submodule 112B may configure an index engine that matches the new load characteristic based on the new load characteristic, and instruct the storage device 120 to migrate the index engine of the user data table to the index engine that matches the new load characteristic, or instruct the storage device 120 to switch the index engine of the user data table to an index engine that matches the new load characteristic.

Among them, in the embodiments of the present application, migration can be understood as a gradual change. For example, a first structure, a second structure and a third structure are set, wherein the first structure and the second structure are quite different, and the third structure is between the first structure and the second structure, that is, the difference between the first structure and the third structure, and the difference between the second structure and the third structure are both smaller than the difference between the first structure and the second structure. The migration of the first structure to the second structure is specifically that the first structure is first switched to the third structure, and then the third structure is switched to the second structure. Thereby, the index engine change overhead caused by switching between structures with large differences can be avoided.

In some embodiments, as shown in FIG. 2 , the control module 110 further includes a load characteristic sensing module 113. The load characteristic sensing module 113 can sense the load characteristic of the data table. Specifically, the load characteristic sensing module 113 can sense the operation amplification of the operation on the data table. When the operation amplification is greater than a preset threshold, it can be determined that the current index engine of the data table does not match the current load characteristic of the data table, and the index engine needs to be adjusted. And the type of the current load characteristic can be determined according to the specific type of the operation.

Specifically, the operation may include a read operation and a write operation, and correspondingly, the operation amplification includes read amplification and write amplification. Among them, the operation amplification refers to the ratio of the amount of data actually operated to the amount of data required to be operated, the read amplification refers to the ratio of the amount of data actually read to the amount of data required to be read, and the write amplification refers to the ratio of the amount of data actually written to the amount of data required to be written.

The load characteristic perception module 113 can perceive the read amplification of the read operation on the data table. Among them, when the read amplification is greater than the preset threshold value A3, it can be determined that the structure of the current index engine of the data table does not match the current load characteristics, and the index engine of the data table needs to be adjusted in the direction of the read-friendly index engine. The load characteristic perception module 113 can perceive the write amplification of the write operation on the data table. In an example, when the read amplification is greater than the preset threshold value A3, and the write amplification is less than the preset threshold value A4, it can be determined that the structure of the current index engine of the data table does not match the current load characteristics of the data table, and the index engine of the data table needs to be adjusted in the direction of the read-friendly index engine. Among them, adjusting the index engine of the data table in the direction of the read-friendly index engine means making the structure of the adjusted index engine more suitable or more matching the read-intensive load than the structure of the index engine before the adjustment.

The threshold value A3 and the threshold value A4 may be preset based on experience or experiments. In one example, the threshold value A3 may be 20, and the threshold value A4 may be 10. In another example, the threshold value A3 may be 30, and the threshold value A4 may be 15. In yet another example, the threshold value A3 may be 40, and the threshold value A4 may be 20. The present embodiment of the application does not specifically limit the threshold value A3 and the threshold value A4.

Among them, the load characteristic perception module 113 can perceive the write amplification of the write operation on the data table. When the write amplification is greater than the preset threshold value A5, it can be determined that the current index engine of the data table does not match the current load characteristics of the data table, and the index engine of the data table needs to be adjusted in the direction of a write-friendly index engine. In one example, when the write amplification is greater than the preset threshold value A5 and the read amplification is less than the preset threshold value A6, it can be determined that the current index engine of the data table does not match the current load characteristics of the data table, and the index engine of the data table needs to be adjusted in the direction of a write-friendly index engine. Among them, adjusting the index engine of the data table in the direction of a write-friendly index engine means making the structure of the adjusted index engine more suitable or more matching to write-intensive loads than the structure of the index engine before the adjustment.

Threshold A5 and threshold A6 can be preset based on experience or experiments. In one example, threshold A5 can be 20, and threshold A6 can be 10. In another example, threshold A5 can be 30, and threshold A6 can be 15. In another example, threshold A5 can be 40, and threshold A6 can be 20. And so on. The embodiments of the present application do not specifically limit threshold A5 and threshold A6.

The above example introduces the storage system 100. Next, the data storage solution provided by the embodiment of the present application is introduced by example in combination with the storage system 100. In which, the data table T1 can be set to correspond to the user 200, and the data table T1 is used to store the service data of the user 200.

Referring to Figure 4, the control module 110 can execute step 401 to provide a configuration interface to the user 200. The configuration interface is used for the user to input the load characteristics of the data table. The load characteristics input by the user through the configuration interface can be read-intensive load, write-intensive load, or mixed load. In other words, the configuration interface is used for the user to configure whether the load characteristics of the data table T1 are read-intensive load, write-intensive load, or mixed load. Exemplarily, when the user does not input the load characteristics to the configuration interface, that is, when the control module 110 does not receive the load characteristics input by the user, the control module 110 can confirm that the load characteristics configured by the user are the default characteristics. In one example, the default characteristic can be a mixed load.

The user 200 may execute step 403 to input the load characteristic E1, wherein the load characteristic E1 may specifically be a read-intensive load, a write-intensive load, or a mixed load.

In some embodiments, the storage system 100 may further include a client (not shown) located at the user 200 side. In step 401, the control device 100 may provide a configuration interface to the client. The user may input on the client to input the load characteristic E1 to the configuration interface. The above only illustrates the method for the user to input the load characteristic to the configuration interface, and does not constitute a limitation. Other methods supported by the prior art may also be used to implement the user inputting the load characteristic to the configuration interface, which will not be described one by one here.

In some embodiments, when the user 200 creates the data table T1, the control device 110 may provide a configuration interface to the user 200. The configuration interface at this time is used for the user 200 to configure the initialization load characteristic. That is, the load characteristic E1 is the initialization load characteristic.

In some embodiments, during the use of the data table T1, the control device 110 may provide a configuration interface to the user 200. The configuration interface at this time is used for the user 200 to adjust the load characteristics. In other words, the load characteristic E1 is a load characteristic actively adjusted by the user. It is not difficult to understand that the user can instruct the data table T1 to serve different businesses in different time periods. The load characteristics caused by different businesses are different. In this way, the user can input the load characteristics caused by the changed business through the configuration interface according to the changes in the business served by the data table T1. That is, the user can actively adjust the load characteristics.

In some embodiments, the configuration interface can be used by the user to configure the load characteristics of the primary index and/or the load characteristics of the secondary index. That is, the load characteristic E1 can be the load characteristics of the primary index, the load characteristics of the secondary index, or the load characteristics of the primary index and the secondary index at the same time.

In some embodiments, the configuration interface may specifically be an application programming interface (API).

In one example, the configuration interface is specifically used for the user 200 to configure the initialization load characteristics of the primary index. The function of the configuration interface can be InitTableStore(workloadType type). The parameter workloadType type can take values from read, write, and default. When the parameter value of workloadType type is read, the load characteristic E1 is a read-intensive load. When the parameter value of workloadType type is write, the load characteristic E1 is a write-intensive load. When the parameter value of workloadType type is default, the load characteristic E1 is a mixed load.

In one example, the configuration interface is specifically used for user 200 to configure the initialization load characteristics of the secondary index, and the function of the configuration interface may be InitIndexStore(workloadType type). The parameter workloadType type takes values from read, write, and default. When the parameter value of workloadType type is read, the load characteristic E1 is a read-intensive load. When the parameter value of workloadType type is write, the load characteristic E1 is a write-intensive load. When the parameter value of workloadType type is default, the load characteristic E1 is a mixed load.

In one example, the configuration interface is used by user 200 to adjust the load characteristics, and the function of the configuration interface may be ChangeStore(workloadType oldType, workloadType newType). The parameter workloadType oldType represents the load characteristics before adjustment, and the parameter workloadType newType represents the load characteristics after adjustment. The load characteristic E1 is the load characteristic after adjustment, that is, the parameter workloadType newType represents the load characteristic E1. The parameters workloadType oldType and workloadType newType can both take values from read, write, and default. As described above, when the value is read, the load characteristic is a read-intensive load. When the value is write, the load characteristic is a write-intensive load. When the value is default, the load characteristic E1 is a mixed load.

Continuing to refer to FIG. 4 , the control device 110 may execute step 405 to determine an index engine E11 that matches the load characteristic E1. When the load characteristic E1 is a read-intensive load, the index engine E11 is a read-friendly index engine. When the load characteristic E1 is a write-intensive load, the index engine E11 is a write-friendly index engine. When the load characteristic E1 is a mixed load, the index engine E11 is a mixed index engine. The structure of the mixed index engine is between the structure of the read-friendly index engine and the structure of the write-friendly index engine.

In some embodiments, an index engine may be constructed based on a B+ tree structure and a log structure merge tree structure. The read-friendly index engine includes at least a B+ tree structure, and the write-friendly index engine includes at least a log structure merge tree structure. The hybrid index engine may include both a B+ tree structure and a log structure merge tree structure, wherein the B+ tree structure is located at the lower layer of the log structure merge tree structure.

In FIG5 , from left to right, multiple index engines are shown in sequence, wherein the read performance of the multiple index engines is enhanced from left to right, and the write performance is enhanced from right to left. The leftmost index engine can be used as a read-friendly index engine, the rightmost index engine can be used as a write-friendly index engine, and the middle index engine can be used as a hybrid index engine.

In an illustrative example, as shown in FIG5 , the read-friendly index engine is specifically a B+ tree structure. The write-friendly index engine is composed of a B+ tree structure and a log-structured merge tree structure, wherein the B+ tree structure is located at the lower layer of the log-structured merge tree structure. The hybrid index engine is also composed of a B+ tree structure and a log-structured merge tree structure, and the B+ tree structure is located at the lower layer of the log-structured merge tree structure. Among them, compared with the write-friendly index engine, the log-structured merge tree structure in the hybrid index engine has fewer storage layers. That is to say, the log-structured merge tree structure in the write-friendly index engine has a storage layer with a larger number of layers, and the log-structured merge tree structure in the hybrid index engine has a storage layer with a smaller number of layers.

As mentioned above, when writing data, the data is first written to the top layer, that is, the C1 layer. When the amount of data in the C1 layer reaches the preset value D1, the data in the C1 layer and the data in the next layer of the C1 layer (that is, the C2 layer) are merged (compaction), and the merged data is transferred to the C2 layer. When the amount of data in the C2 layer reaches the preset value D2, the data in the C2 layer and the data in the next layer of the C2 layer (that is, the C3 layer) are merged, and the merged data is transferred to the C3 layer, and so on. Therefore, the more storage layers there are in the log structure merge tree structure, the better the data aggregation effect of the upper layer of the B+ tree, and the less data is written to the B+ tree structure, so write amplification can be reduced and write latency can be reduced.

The more storage layers there are in the log-structured merge tree structure, the more historical versions of the data may be, which leads to read amplification and increases read latency. Therefore, in order to reduce read amplification and lower read latency, it is necessary to reduce the number of storage layers in the log-structured merge tree structure. In this way, by adjusting the number of storage layers in the log-structured merge tree structure, index engines with different read and write performances are constructed, that is, a read-friendly index engine, a write-friendly index engine, and a hybrid index engine are constructed. Among them, the number of storage layers of the log-structured merge tree structure in the hybrid index engine can be reduced or increased, so that the hybrid index engine tends to be read-friendly or write-friendly.

Returning to FIG. 4 , when the index engine E11 is determined in step 405, the control device 110 may execute step 407 to instruct the storage device 120 to store the data in the data table T1 according to the index engine E11. Specifically, as shown in FIG. 4 , the control device 110 may execute step 4071 to send instruction information to the storage device 120, wherein the instruction information may include the identifier of the index engine E11. The storage device 120 may execute step 4072 to store the data in the data table T1 according to the index engine E11 in response to the instruction information.

When the load characteristic E1 is the load characteristic of the primary index, the control device 110 instructs the storage device 120 to store the data in the entire data table T1 according to the index engine E11, or to store the data in the data partition instance corresponding to the primary index. Specifically, the instruction information may include the identifier of the data table T1 or the identifier of the data partition instance while including the identifier of the index engine E11. Thus, the storage device 120 stores the data in the entire data table T1 or the data partition instance according to the identifier of the data table T1 or the identifier of the data partition instance according to the index engine E11, or stores the data in the data partition instance corresponding to the primary index.

When the load characteristic E1 is the load characteristic of the secondary index, the control 110 instructs the storage device 120 to store the data under the secondary index according to the index engine E11. Specifically, the instruction information may include the identifier of the secondary index while including the identifier of the index engine E11, so that the storage device 120 stores the data under the secondary index according to the index engine E11 according to the identifier of the secondary index.

In some embodiments, before executing step 407, the index engine of data table T1 is index engine E12. Among them, index engine E12 is a read-friendly index engine, and index engine E11 is a write-friendly index engine; or, index engine E11 is a read-friendly index engine, and index engine E12 is a write-friendly index engine. In this case, in step 407, the control device 110 may instruct the storage device 120 to first store the data in data table T1 according to the hybrid index engine, and then store the data in data table T1 according to index engine E11. The storage device 120 may first change the index engine of data table T1 from index engine E12 to a hybrid index engine, and then change the index engine of data table T1 from a hybrid index engine to index engine E1. In this way, the index engine for storing data is gradually migrated from index engine E12 to index engine E11, which can reduce the overhead of changing the index engine.

In the data storage solution provided in the embodiment of the present application, the user can configure the index engine of his data table, so that when the business served by the data table changes, the user can adjust the index engine of the data table in a timely manner or at any time, so that the index engine matches the load characteristics caused by the changed business, thereby improving the access performance of the data table.

In conjunction with the storage system 100 shown in Fig. 2, the embodiment of the present application further provides a data storage solution. Next, the solution is introduced by way of example.

As shown in FIG6 , the control device 110 may perform step 601 to monitor the operation amplification of the data table T1. Exemplarily, step 601 may be performed periodically. The average value of the operation amplification monitored in an execution cycle may be used as the operation amplification of the execution cycle. The execution cycle of step 601 may be preset. In one example, the execution cycle of step 601 may be 10 minutes. In another example, the execution cycle of step 601 may be 20 minutes. And so on.

The operation amplification may include read amplification of data read from the data table T1 or write amplification of data written to the data table T1. The operation amplification may also include read amplification of data read from the data table T1 and/or write amplification of data written to the data table T1.

In addition, in the following description, unless otherwise specified, read amplification refers to read amplification of reading data from the data table T1, and write amplification refers to write amplification of writing data to the data table T1.

Continuing to refer to FIG. 6 , the control device 100 may execute step 603 to determine whether the operation amplification is greater than the threshold y1 .

In some embodiments, the operation amplification includes read amplification, and the threshold y1 includes a threshold A3. In step 603, it can be determined whether the read amplification is greater than the threshold A3. If the read amplification is greater than the threshold A3, it can be determined that the current index engine of the data table T1 does not match the current load characteristics of the data table T1, and the index engine of the data table T1 needs to be adjusted in the direction of a read-friendly index engine.

In an illustrative example of this embodiment, operation amplification includes read amplification and write amplification, and threshold y1 includes threshold A3 and threshold A4. In step 603, it can be determined whether the read amplification is greater than threshold A3, and whether the write amplification is less than threshold A4. If the read amplification is greater than threshold A3, and the write amplification is less than threshold A4, it can be determined that the current index engine of data table T1 does not match the current load characteristics of data table T1, and it is necessary to adjust the index engine of data table T1 in the direction of a read-friendly index engine.

The threshold values A3 and A4 may be specifically described above and will not be described in detail here.

In some embodiments, the operation amplification includes write amplification, and the threshold y1 includes a threshold A5. In step 603, it can be determined whether the write amplification is greater than the threshold A5. If the write amplification is greater than the threshold A5, it can be determined that the current index engine of the data table T1 does not match the current load characteristics of the data table T1, and the index engine of the data table T1 needs to be adjusted in the direction of a write-friendly index engine.

In an illustrative example of this embodiment, operation amplification includes write amplification and read amplification, and threshold y1 includes threshold A5 and threshold A6. In step 603, it can be determined whether the write amplification is greater than threshold A5, and whether the read amplification is less than threshold A6. If the write amplification is greater than threshold A5, and the read amplification is less than threshold A6, it can be determined that the current index engine of data table T1 does not match the current load characteristics of data table T1, and it is necessary to adjust the index engine of data table T1 in the direction of a write-friendly index engine.

The specific details of the threshold A5 and the threshold A6 can be referred to the above description, which will not be described again here.

Continuing to refer to FIG. 6 , the control device 110 may execute step 605 to adjust the index engine of the data table T1 to the index engine E21 in the direction of reducing the operation amplification.

Specifically, when it is determined in step 603 that the index engine of data table T1 needs to be adjusted in the direction of a read-friendly index engine, then index engine E21 is an index engine that is more suitable or more conducive to read operations than the current index engine of data table T1, that is, the matching degree between index engine E21 and the read-intensive load is greater than the matching degree between the current index engine of data table T1 and the read-intensive load. In other words, the read amplification of data table T1 when the data of data table T1 is stored according to index engine E21 is less than the current read amplification of data table T1.

In some embodiments, as described above, the number of storage layers of the log structure merge tree in the index engine can be adjusted to make the index engine more read-friendly or write-friendly. The index engine is more read-friendly. In this way, when it is determined in step 603 that the index engine of data table T1 needs to be adjusted in the direction of a read-friendly index engine, an index engine with N fewer storage layers than the current index engine structure of data table T1 can be used as index engine E21. In other words, the structure of index engine E21 has N fewer storage layers than the current index engine structure of data table T1. The storage layer refers to the storage layer of the log structure merge tree, and N is an integer greater than or equal to 1. The value of N can be preset. In one example, N is 1, 2, or 3, etc.

When it is determined in step 603 that the index engine of data table T1 needs to be adjusted in the direction of a write-friendly index engine, then index engine E21 is an index engine that is more suitable or more conducive to write operations than the current index engine of data table T1, that is, the matching degree between index engine E21 and the write-intensive load is greater than the matching degree between the current index engine of data table T1 and the write-intensive load. In other words, the write amplification of data table T1 when the data of data table T1 is stored according to index engine E21 is less than the current write amplification of data table T1.

In some embodiments, as described above, the index engine can be made more read-friendly or write-friendly by adjusting the number of storage layers of the log-structured merge tree in the index engine. Among them, when the number of storage layers of the log-structured merge tree in the index engine increases, the index engine is more write-friendly. In this way, when it is determined in step 603 that the index engine of data table T1 needs to be adjusted in the direction of a write-friendly index engine, an index engine with M more storage layers than the structure of the current index engine of data table T1 can be used as index engine E21. That is, compared with the current index engine of data table T1, the structure of index engine E21 has M more storage layers than the structure of the current index engine of data table T1. Among them, the storage layer refers to the storage layer of the log-structured merge tree, and M is an integer greater than or equal to 1. Among them, the M values can be preset. In one example, M is 1, 2, or 3, etc.

The control device 110 may also execute step 607 to instruct the storage device 120 to store the data in the data table T1 according to the index engine E21.

6, the control device 110 may execute step 6071 to send indication information to the storage device 120, wherein the indication information may include the identifier of the index engine E21. The storage device 120 may execute step 6072 to store the data in the data table T1 according to the index engine E21 in response to the indication information.

Among them, when the operation amplification monitored in step 601 is the operation amplification of the primary index, the control device 110 instructs the storage device 120 to store the data in the entire data table T1 according to the index engine E21, or to store the data in the data partition instance corresponding to the primary index. Specifically, the instruction information may include the identifier of the data table T1 or the identifier of the data partition instance while including the identifier of the index engine E21, so that the storage device 120 may store the data in the entire data table T1 according to the identifier of the data table T1 or the identifier of the data partition instance, or to store the data in the data partition instance corresponding to the primary index according to the index engine E21.

When the operation amplification monitored in step 601 is the operation amplification of the local secondary index of the data table T1, the control 110 instructs the storage device 120 to store the data under the local secondary index according to the index engine E21. Specifically, the instruction information may include the identifier of the local secondary index as well as the identifier of the index engine E21, so that the storage device 120 stores the data under the local secondary index according to the index engine E21 according to the identifier of the local secondary index.

Continuing to refer to FIG. 6 , after step 607, the control device 110 may execute step 601 and step 603 again. When the operation amplification is not greater than the threshold value y1, the index engine for adjusting the data table T1 may be stopped. When the operation amplification is greater than the threshold value y1, steps 605 and 607 may be executed again. For details, please refer to the above description, which will not be repeated here.

In this way, through iterative execution of step 601 to step 607 , the index engine of data table T1 can be dynamically adjusted so that the index engine of data table T1 matches the load characteristics of data table T1 as much as possible, thereby reducing operation amplification.

The data storage solution provided in the embodiment of the present application can sense the dynamic changes of the operation amplification of the data table, and dynamically adjust the index engine of the data table according to the dynamic changes of the operation amplification, so that the index engine matches the load characteristics of the data table, thereby improving the access performance of the data table.

As shown in Figure 7, the control device 110 can execute step 701 to monitor the read and write operations of data table T1 of data table T1. Exemplarily, step 601 can be executed periodically. Among them, the read and write operations include read operations and write operations. In step 701, the total number of read operations and the total number of write operations in an execution cycle can be monitored to obtain the monitoring result of the execution cycle. That is, the monitoring result includes the total number of read operations and the total number of write operations of data table T1 in the monitoring cycle. The execution cycle can be preset. In one example, the execution cycle of step 701 can be 1 hour. In another example, the execution cycle of step 701 can be two hours. And so on.

The control device 110 may execute step 703 and obtain the load characteristic E3 according to the monitoring result. When the ratio of the total number of read operations to the total number of write operations is greater than the threshold value A1, the read-intensive load is used as the load characteristic E3. When the ratio of the total number of write operations to the total number of read operations is greater than the threshold value A2, the write-intensive load is used as the load characteristic E3. For the specifics of the threshold values A1 and A2, please refer to the above description and will not be repeated here.

Next, the control device 110 may execute step 705 to migrate to an indexing engine matching the load characteristic E3 to obtain an indexing engine E31.

When the load characteristic E3 is a read-intensive load, the structure of the index engine E31 is less than the structure of the current index engine of the data table T1. N storage layers. The storage layer refers to the storage layer of the log structure merge tree, and N is an integer greater than or equal to 1. The value of N can be preset. In one example, N is 1, 2, or 3, etc.

When the load characteristic E3 is a write-intensive load, the structure of the index engine E31 has M more storage layers than the structure of the current index engine of the data table T1. The storage layer refers to the storage layer of the log structure merge tree, and M is an integer greater than or equal to 1. The value of M can be preset. In one example, M is 1, 2, or 3, etc.

Then, the control device 110 may execute step 707 to instruct the storage device 120 to store the data in the data table T1 according to the index engine E31. Among them, step 707 may include step 7071, sending instruction information to the storage device. Step 707 may also include step 7072, storing the data in the data table T1 according to the index engine E31. For details, please refer to the above introduction to step 407, step 4071, and step 4072 in Figure 4, which will not be repeated here.

The data storage solution provided in the embodiment of the present application can sense the dynamic changes in the load characteristics of the data table, and dynamically adjust the index engine of the data table according to the dynamic changes in the load characteristics, so that the index engine matches the load characteristics of the data table, thereby improving the access performance of the data table.

Based on the data storage solution described above, the embodiment of the present application provides a data storage method. It can be understood that the method is combined with the data storage solution described above, and the specific execution process of the relevant steps in the method can refer to the execution process of the corresponding steps in the data storage solution.

The method is applied to a control device in a storage system (eg, control device 110 in storage system 100), the storage system further includes a storage device (eg, storage device 120 in storage system 100), and the storage device stores a user data table. As shown in FIG8 , the method includes the following steps.

Step 801, providing a configuration interface, wherein the configuration interface is used for the user to configure the load characteristics of the data table as read-intensive load or write-intensive load. For details, please refer to the above description of step 401 in FIG. 4 .

Step 803a, when the configuration interface indicates that the user configures the load characteristic of the data table as a read-intensive load, instruct the storage device to store the data in the data table according to the first index engine; the first index engine matches the read-intensive load. For details, please refer to the above description of steps 403 to 407 in Figure 4.

Step 803b, when the configuration interface indicates that the user configures the load characteristic of the data table as a write-intensive load, instruct the storage device to store the data in the data table according to the second index engine; the second index engine matches the write-intensive load. For details, please refer to the above description of steps 403 to 407 in Figure 4.

In some embodiments, the first indexing engine includes at least a B+ tree structure, and the second indexing engine includes at least a log structure merge tree structure.

In some embodiments, before instructing the storage device to store the data in the data table according to the first index engine, the index engine in the data table is the second index engine; instructing the storage device to store the data in the data table according to the first index engine includes: instructing the storage device to first migrate the index engine in the data table from the second index engine to the third index engine, the structure of the third index engine being between the structure of the first index engine and the structure of the second index engine; and then instructing the storage device to migrate the index engine of the data table from the third index engine to the first index engine. For details, please refer to the above description of step 407 in Figure 4.

In some embodiments, the data table has a local secondary index, and the configuration interface is also used for the user to configure the load characteristics of the local secondary index as a read-intensive load or a write-intensive load; when the configuration interface instructs the user to configure the load characteristics of the local secondary index as a read-intensive load, the storage device is instructed to store the data under the local secondary index according to the first index engine; when the configuration interface instructs the user to configure the load characteristics of the local secondary index as a write-intensive load, the storage device is instructed to store the data under the local secondary index according to the second index engine. For details, please refer to the above description of step 407 in Figure 4.

Through the data storage method provided in the embodiment of the present application, users can configure the index engine of their data table, so that when the business served by the data table changes, the user can adjust the index engine of the data table in a timely manner or at any time, so that the index engine matches the load characteristics caused by the changed business, thereby improving the access performance of the data table.

The embodiment of the present application also provides a data storage method, which is applied to a control device in a storage system (e.g., the control device 110 in the storage system 100), wherein the storage system further includes a storage device (e.g., the storage device 120 in the storage system 100), and the storage device stores a data table. As shown in FIG9 , the method includes the following steps.

In step 901, the control device monitors the operation amplification of the data table when the index engine of the data table is the fourth index engine, and the operation amplification includes read amplification of reading data from the data table or write amplification of writing data to the data table. For details, please refer to the above description of step 601 in FIG. 6.

Step 903a, when the operation amplification includes the read amplification, and the read amplification is greater than a first threshold, instructing the storage device to The fifth index engine stores the data in the data table; wherein the matching degree between the fifth index engine and the read-intensive load is greater than the matching degree between the fourth index engine and the read-intensive load. For details, please refer to the above description of steps 603 to 607 in FIG. 6 .

Step 903b, when the operation amplification includes the write amplification, and the write amplification is greater than the second threshold, instruct the storage device to store the data in the data table according to the sixth index engine; wherein the matching degree between the sixth index engine and the write-intensive load is greater than the matching degree between the fourth index engine and the write-intensive load. For details, please refer to the above description of steps 603 to 607 in Figure 6.

In some embodiments, the operation amplification includes both the read amplification and the write amplification; when the operation amplification includes the read amplification, and the read amplification is greater than the first threshold, instructing the storage device to store the data in the data table according to the fifth index engine, including: when the read amplification is greater than the first threshold, and the write amplification is less than the third threshold, instructing the storage device to store the data in the data table according to the fifth index engine. For details, please refer to the above description of steps 603 to 607 in Figure 6.

In some embodiments, the operation amplification includes both the read amplification and the write amplification; when the operation amplification includes the write amplification, and the write amplification is greater than the second threshold, instructing the storage device to store the data in the data table according to the sixth index engine, including: when the write amplification is greater than the second threshold, and the read amplification is less than the fourth threshold, instructing the storage device to store the data in the data table according to the sixth index engine. For details, please refer to the above description of steps 603 to 607 in Figure 6.

In some embodiments, the fifth index engine includes at least a B+ tree structure, and the sixth index engine includes at least a log structure merge tree structure.

In some embodiments, the data table has a local secondary index LSI, the operation amplification is the amplification generated by operating the data under the local secondary index; the instructing the storage device to store the data in the data table according to the fifth index engine includes: instructing the storage device to store the data under the local secondary index according to the fifth index engine; or, the instructing the storage device to store the data in the data table according to the sixth index engine includes: instructing the storage device to store the data under the local secondary index according to the sixth index engine. For details, please refer to the above introduction to step 607 in Figure 6.

The data storage method provided in the embodiment of the present application can sense the dynamic changes of operation amplification of a data table, and dynamically adjust the index engine of the data table according to the dynamic changes of operation amplification, so that the index engine matches the load characteristics of the data table, thereby improving the access performance of the data table.

The embodiment of the present application provides a data storage device 1000. The device 1000 can be configured in a control device in a storage system, and the storage system also includes a storage device, and the storage device stores a user's data table. As shown in FIG. 10 , the device 1000 includes:

A module 1010 is provided, for providing a configuration interface, wherein the configuration interface is used for the user to configure the load characteristic of the data table as a read-intensive load or a write-intensive load;

The instructing module 1020 is used to instruct the storage device to store the data in the data table according to the first indexing engine when the configuration interface indicates that the user configures the load characteristic of the data table as a read-intensive load; the first indexing engine matches the read-intensive load;

The indication module 1020 is also used to instruct the storage device to store the data in the data table according to the second index engine when the configuration interface indicates that the user configures the load characteristics of the data table as a write-intensive load; the second index engine matches the write-intensive load.

Among them, both the providing module 1010 and the indicating module 1020 can be implemented by software or by hardware. Exemplarily, the implementation of the providing module 1010 is introduced below by taking the providing module 1010 as an example. Similarly, the implementation of the indicating module 1020 can refer to the implementation of the providing module 1010.

As an example of a software functional unit, the module 1010 provided may include code running on a computing instance. Among them, the computing instance may include at least one of a physical host (computing device), a virtual machine, and a container. Further, the above-mentioned computing instance may be one or more. For example, the module 1010 provided may include code running on multiple hosts/virtual machines/containers. It should be noted that the multiple hosts/virtual machines/containers used to run the code may be distributed in the same region (region) or in different regions. Furthermore, the multiple hosts/virtual machines/containers used to run the code may be distributed in the same availability zone (AZ) or in different AZs, each AZ including one data center or multiple data centers with close geographical locations. Among them, usually a region may include multiple AZs.

Similarly, multiple hosts/virtual machines/containers used to run the code can be distributed in the same virtual private cloud (VPC) or in multiple VPCs. Usually, a VPC is set up in a region. For cross-region communication between two VPCs in the same region and between VPCs in different regions, a communication gateway needs to be set up in each VPC to achieve interconnection between VPCs through the communication gateway.

As an example of a hardware functional unit, the providing module 1010 may include at least one computing device, such as a server, etc. Alternatively, the providing module 1010 may also be implemented using an application-specific integrated circuit (ASIC). Or a device implemented by a programmable logic device (PLD), etc. The PLD may be a complex programmable logical device (CPLD), a field-programmable gate array (FPGA), a generic array logic (GAL), or any combination thereof.

The multiple computing devices included in the providing module 1010 may be distributed in the same region or in different regions. The multiple computing devices included in the providing module 1010 may be distributed in the same AZ or in different AZs. Similarly, the multiple computing devices included in the providing module 1010 may be distributed in the same VPC or in multiple VPCs. The multiple computing devices may be any combination of computing devices such as servers, ASICs, PLDs, CPLDs, FPGAs, and GALs.

It should be noted that, in other embodiments, the providing module 1010 can be used to execute any step in the method shown in FIG8, and the indicating module 1020 can be used to execute any step in the method shown in FIG8. The steps that the providing module 1010 and the indicating module 1020 are responsible for implementing can be specified as needed, and the full functions of the data storage device 1000 are realized by implementing different steps in the method shown in FIG8 by the providing module 1010 and the indicating module 1020 respectively.

The embodiment of the present application also provides a data storage device 1100. The device 1100 can be configured in a control device in a storage system, and the storage system also includes a storage device, and the storage device stores a data table. As shown in FIG. 11 , the device 1100 includes:

A monitoring module 1110, configured for the control device to monitor, when the index engine of the data table is the fourth index engine, the operation amplification of the data table, wherein the operation amplification includes a read amplification of reading data from the data table or a write amplification of writing data to the data table;

an indication module 1120, configured to indicate, when the operation amplification includes the read amplification and the read amplification is greater than a first threshold, the storage device to store the data in the data table according to a fifth index engine; wherein the matching degree between the fifth index engine and the read-intensive load is greater than the matching degree between the fourth index engine and the read-intensive load;

The indication module 1120 is also used to instruct the storage device to store the data in the data table according to the sixth index engine when the operation amplification includes the write amplification and the write amplification is greater than a second threshold; wherein the matching degree between the sixth index engine and the write-intensive load is greater than the matching degree between the fourth index engine and the write-intensive load.

The monitoring module 1110 and the indicating module 1120 can be implemented by software or hardware. The implementation of the monitoring module 1110 and the indicating module 1120 can refer to the implementation of the providing module 1010, which is described above and will not be repeated here.

It should be noted that, in other embodiments, the monitoring module 1110 can be used to execute any step in the method shown in FIG9 , and the indicating module 1120 can be used to execute any step in the method shown in FIG9 . The steps that the monitoring module 1110 and the indicating module 1120 are responsible for implementing can be specified as needed, and the monitoring module 1110 and the indicating module 1120 respectively implement different steps in the method shown in FIG9 to implement all functions of the data storage device 1100 .

The present application also provides a computing device 1200. As shown in FIG. 12 , the computing device 1200 includes: a bus 1202, a processor 1204, a memory 1206, and a communication interface 1208. The processor 1204, the memory 1206, and the communication interface 1208 communicate with each other through the bus 1202. The computing device 1200 may be a server or a terminal device. It should be understood that the present application does not limit the number of processors and memories in the computing device 1200.

The bus 1202 may be a peripheral component interconnect (PCI) bus or an extended industry standard architecture (EISA) bus, etc. The bus may be divided into an address bus, a data bus, a control bus, etc. For ease of representation, FIG. 12 is represented by only one line, but does not mean that there is only one bus or one type of bus. The bus 1202 may include a path for transmitting information between various components of the computing device 1200 (e.g., the memory 1206, the processor 1204, and the communication interface 1208).

Processor 1204 may include any one or more of a central processing unit (CPU), a graphics processing unit (GPU), a microprocessor (MP), or a digital signal processor (DSP).

The memory 1206 may include a volatile memory (volatile memory), such as a random access memory (RAM). The memory 1206 may also include a non-volatile memory (non-volatile memory), such as a read-only memory (ROM), a flash memory, a hard disk drive (HDD), or a solid state drive (SSD).

The memory 1206 stores executable program codes, and the processor 1204 executes the executable program codes to respectively implement the functions of the providing module 1010 and the indicating module 1020, thereby implementing the method shown in Figure 8. That is, the memory 1206 stores instructions for executing the method shown in Figure 8.

The communication interface 1208 uses a transceiver module such as, but not limited to, a network interface card or a transceiver to implement communication between the computing device 1200 and other devices or a communication network.

The present application also provides a computing device cluster. The computing device cluster includes at least one computing device. The computing device may be a server, such as a central server, an edge server, or a local server in a local data center. In some embodiments, the computing device cluster includes at least one computing device. The computing device can also be a terminal device such as a desktop computer, a laptop computer or a smart phone.

As shown in Fig. 13, the computing device cluster includes at least one computing device 1200. The memory 1206 in one or more computing devices 1200 in the computing device cluster may store the same instructions for executing the method shown in Fig. 8 .

In some possible implementations, the memory 1206 of one or more computing devices 1200 in the computing device cluster may also respectively store some instructions for executing the method shown in Figure 8. In other words, the combination of one or more computing devices 1200 may jointly execute instructions for executing the method shown in Figure 8.

It should be noted that the memory 1206 in different computing devices 1200 in the computing device cluster may store different instructions, which are respectively used to execute part of the functions of the apparatus 1000. That is, the instructions stored in the memory 1206 in different computing devices 1200 may implement the functions of one or more of the providing module 1010 and the indicating module 1020.

In some possible implementations, one or more computing devices in the computing device cluster can be connected via a network. The network can be a wide area network or a local area network, etc. FIG. 14 shows a possible implementation. As shown in FIG. 14 , two computing devices 1200A and 1200B are connected via a network. Specifically, the network is connected via a communication interface in each computing device. In this type of possible implementation, the memory 1206 in the computing device 1200A stores instructions for executing the functions of the providing module 1010. At the same time, the memory 1206 in the computing device 1200B stores instructions for executing the functions of the indicating module 1020.

It should be understood that the functionality of the computing device 1200A shown in FIG14 may also be accomplished by multiple computing devices 1200. Similarly, the functionality of the computing device 1200B may also be accomplished by multiple computing devices 1200.

The embodiment of the present application also provides another computing device cluster. The connection relationship between the computing devices in the computing device cluster can be similar to the connection mode of the computing device cluster described in Figures 13 and 14. The difference is that the memory 1206 in one or more computing devices 1200 in the computing device cluster can store the same instructions for executing the method shown in Figure 8.

The present application also provides a computing device 1500. As shown in FIG. 15 , the computing device 1500 includes: a bus 1502, a processor 1504, a memory 1506, and a communication interface 1508. The processor 1504, the memory 1506, and the communication interface 1508 communicate with each other through the bus 1502. The computing device 1500 can be a server or a terminal device. It should be understood that the present application does not limit the number of processors and memories in the computing device 1500.

The implementations of the bus 1502 , the processor 1504 , the memory 1506 , and the communication interface 1508 may refer to the implementations of the bus 1202 , the processor 1204 , the memory 1206 , and the communication interface 1208 , respectively.

The memory 1506 stores executable program codes, and the processor 1504 executes the executable program codes to respectively implement the functions of the aforementioned monitoring module 1110 and the indication module 1120, thereby implementing the method shown in Figure 9. That is, the memory 1506 stores instructions for executing the method shown in Figure 9.

The communication interface 1508 uses a transceiver module such as, but not limited to, a network interface card or a transceiver to implement communication between the computing device 1500 and other devices or communication networks.

The embodiment of the present application also provides a computing device cluster. The computing device cluster includes at least one computing device. The computing device can be a server, such as a central server, an edge server, or a local server in a local data center. In some embodiments, the computing device can also be a terminal device such as a desktop computer, a laptop computer, or a smart phone.

As shown in Fig. 16, the computing device cluster includes at least one computing device 1500. The memory 1506 in one or more computing devices 1500 in the computing device cluster may store the same instructions for executing the method shown in Fig. 9.

In some possible implementations, the memory 1506 of one or more computing devices 1500 in the computing device cluster may also respectively store some instructions for executing the method shown in Figure 9. In other words, the combination of one or more computing devices 1500 may jointly execute instructions for executing the method shown in Figure 9.

It should be noted that the memory 1506 in different computing devices 1500 in the computing device cluster may store different instructions, which are respectively used to execute part of the functions of the apparatus 1100. That is, the instructions stored in the memory 1506 in different computing devices 1500 may implement the functions of one or more modules in the monitoring module 1110 and the indication module 1120.

In some possible implementations, one or more computing devices in the computing device cluster can be connected via a network. Wherein, the network can be a wide area network or a local area network, etc. Figure 17 shows a possible implementation. As shown in Figure 17, two computing devices 1500A and 1500B are connected via a network. Specifically, the network is connected via a communication interface in each computing device. In this type of possible implementation, the memory 1506 in the computing device 1500A stores instructions for executing the functions of the monitoring module 1110. At the same time, the memory 1506 in the computing device 1500B stores instructions for executing the functions of the indication module 1120.

It should be understood that the functions of the computing device 1500A shown in FIG. 17 may also be performed by multiple computing devices 1500. The functionality of 1500B may also be performed by multiple computing devices 1500 .

The embodiment of the present application also provides another computing device cluster. The connection relationship between the computing devices in the computing device cluster can be similar to the connection mode of the computing device cluster described in Figures 16 and 17. The difference is that the memory 1506 in one or more computing devices 1500 in the computing device cluster can store the same instructions for executing the method shown in Figure 9.

The embodiment of the present application also provides a computer program product including instructions. The computer program product may be software or a program product including instructions that can be run on a computing device or stored in any available medium. When the computer program product is run on at least one computing device, the at least one computing device executes the method shown in FIG8 .

The embodiment of the present application also provides a computer program product including instructions. The computer program product may be software or a program product including instructions that can be run on a computing device or stored in any available medium. When the computer program product is run on at least one computing device, the at least one computing device executes the method shown in FIG. 9 .

The embodiment of the present application also provides a computer-readable storage medium. The computer-readable storage medium can be any available medium that can be stored by a computing device or a data storage device such as a data center containing one or more available media. The available medium can be a magnetic medium (e.g., a floppy disk, a hard disk, a tape), an optical medium (e.g., a DVD), or a semiconductor medium (e.g., a solid-state hard disk). The computer-readable storage medium includes instructions that instruct the computing device to execute the method shown in Figure 8.

The embodiment of the present application also provides a computer-readable storage medium. The computer-readable storage medium can be any available medium that can be stored by a computing device or a data storage device such as a data center containing one or more available media. The available medium can be a magnetic medium (e.g., a floppy disk, a hard disk, a tape), an optical medium (e.g., a DVD), or a semiconductor medium (e.g., a solid-state hard disk). The computer-readable storage medium includes instructions that instruct the computing device to execute the method shown in Figure 9.

Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not cause the essence of the corresponding technical solutions to deviate from the protection scope of the technical solutions of the embodiments of the present invention.

Claims

A data storage method, characterized in that it is applied to a control device in a storage system, wherein the storage system further comprises a storage device, and the storage device stores a user's data table; the method comprises:

Providing a configuration interface, wherein the configuration interface is used for the user to configure the load characteristics of the data table as a read-intensive load or a write-intensive load;

When the configuration interface instructs the user to configure the load characteristic of the data table as a read-intensive load, instructing the storage device to store the data in the data table according to a first index engine; the first index engine matches the read-intensive load;

When the configuration interface instructs the user to configure the load characteristic of the data table as a write-intensive load, the storage device is instructed to store the data in the data table according to a second index engine; the second index engine matches the write-intensive load.
The method according to claim 1 is characterized in that the first index engine includes at least a B+ tree structure, and the second index engine includes at least a log structure merged tree structure.
The method according to claim 1 or 2, characterized in that before instructing the storage device to store the data in the data table according to the first index engine, the index engine in the data table is the second index engine;

The instructing the storage device to store the data in the data table according to the first index engine includes:

Instructing the storage device to first migrate the index engine in the data table from the second index engine to a third index engine, wherein the structure of the third index engine is between the structure of the first index engine and the structure of the second index engine;

Then, the storage device is instructed to migrate the index engine of the data table from the third index engine to the first index engine.
The method according to any one of claims 1 to 3, characterized in that the data table has a local secondary index, and the configuration interface is further used for the user to configure the load characteristics of the local secondary index as a read-intensive load or a write-intensive load;

When the configuration interface indicates that the user configures the load characteristic of the local secondary index as a read-intensive load, instructing the storage device to store the data under the local secondary index according to the first index engine;

When the configuration interface instructs the user to configure the load characteristic of the local secondary index as a write-intensive load, the storage device is instructed to store the data under the local secondary index according to the second index engine.
A data storage method, characterized in that it is applied to a control device in a storage system, wherein the storage system further comprises a storage device, and the storage device stores a data table; the method comprises:

The control device monitors the operation amplification of the data table when the index engine of the data table is a fourth index engine, the operation amplification including read amplification of reading data from the data table or write amplification of writing data to the data table;

When the operation amplification includes the read amplification, and the read amplification is greater than a first threshold, instructing the storage device to store the data in the data table according to a fifth index engine; wherein the matching degree between the fifth index engine and the read-intensive load is greater than the matching degree between the fourth index engine and the read-intensive load;

When the operation amplification includes the write amplification and the write amplification is greater than a second threshold, the storage device is instructed to store the data in the data table according to a sixth index engine; wherein the matching degree between the sixth index engine and the write-intensive load is greater than the matching degree between the fourth index engine and the write-intensive load.
The method according to claim 5, characterized in that the operational amplification includes both the read amplification and the write amplification;

When the operation amplification includes the read amplification, and the read amplification is greater than a first threshold, instructing the storage device to store the data in the data table according to the fifth index engine, including:

When the read amplification is greater than the first threshold and the write amplification is less than a third threshold, the storage device is instructed to store the data in the data table according to the fifth index engine.
The method according to claim 5 or 6, characterized in that the operational amplification includes both the read amplification and the write amplification;

When the operation amplification includes the write amplification, and the write amplification is greater than a second threshold, instructing the storage device to store the data in the data table according to the sixth index engine, including:

When the write amplification is greater than the second threshold and the read amplification is less than a fourth threshold, the storage device is instructed to store the data in the data table according to the sixth index engine.
The method according to any one of claims 5-7 is characterized in that the fifth index engine includes at least a B+ tree structure, and the sixth index engine includes at least a log structure merged with a tree structure.
The method according to any one of claims 5 to 8, characterized in that the data table has a local secondary index LSI, and the operation amplification is the amplification generated by operating the data under the local secondary index;

The instructing the storage device to store the data in the data table according to the fifth index engine includes: instructing the storage device to store the data under the local secondary index according to the fifth index engine; or

The instructing the storage device to store the data in the data table according to the sixth index engine includes: instructing the storage device to store the data under the local secondary index according to the sixth index engine.
A data storage device, characterized in that it is a control device configured in a storage system, the storage system further comprising a storage device, the storage device storing a user's data table; the data storage device comprises:

A module is provided, which is used to provide a configuration interface, wherein the configuration interface is used for the user to configure the load characteristics of the data table as a read-intensive load or a write-intensive load;

an indication module, configured to, when the configuration interface indicates that the user configures the load characteristic of the data table as a read-intensive load, indicate the storage device to store the data in the data table according to a first index engine; the first index engine matches the read-intensive load;

The indication module is also used to instruct the storage device to store the data in the data table according to the second index engine when the configuration interface instructs the user to configure the load characteristics of the data table as a write-intensive load; the second index engine matches the write-intensive load.
The data storage device according to claim 10 is characterized in that the first index engine includes at least a B+ tree structure, and the second index engine includes at least a log structure merged tree structure.
The data storage device according to claim 10 or 11, characterized in that before instructing the storage device to store the data in the data table according to the first index engine, the index engine in the data table is the second index engine; and the instruction module is further used to:

Instructing the storage device to first migrate the index engine in the data table from the second index engine to a third index engine, wherein the structure of the third index engine is between the structure of the first index engine and the structure of the second index engine;

Then, the storage device is instructed to migrate the index engine of the data table from the third index engine to the first index engine.
The data storage device according to any one of claims 10 to 12, characterized in that the data table has a local secondary index, and the configuration interface is further used for the user to configure the load characteristics of the local secondary index as a read-intensive load or a write-intensive load; the indication module is also used to:

When the configuration interface indicates that the user configures the load characteristic of the local secondary index as a read-intensive load, instructing the storage device to store the data under the local secondary index according to the first index engine;

When the configuration interface instructs the user to configure the load characteristic of the local secondary index as a write-intensive load, the storage device is instructed to store the data under the local secondary index according to the second index engine.
A data storage device, characterized in that it is a control device configured in a storage system, the storage system further includes a storage device, the storage device stores a data table; the data storage device includes:

a monitoring module, used for the control device to monitor the operation amplification of the data table when the index engine of the data table is the fourth index engine, the operation amplification including read amplification of reading data from the data table or write amplification of writing data to the data table;

an indication module, configured to indicate, when the operation amplification includes the read amplification and the read amplification is greater than a first threshold, the storage device to store the data in the data table according to a fifth index engine; wherein the matching degree between the fifth index engine and the read-intensive load is greater than the matching degree between the fourth index engine and the read-intensive load;

The indication module is also used to instruct the storage device to store the data in the data table according to a sixth index engine when the operation amplification includes the write amplification and the write amplification is greater than a second threshold; wherein the matching degree between the sixth index engine and the write-intensive load is greater than the matching degree between the fourth index engine and the write-intensive load.
The data storage device according to claim 14, characterized in that the operational amplification includes both the read amplification and the write amplification;

The indication module is used for: when the read amplification is greater than the first threshold and the write amplification is less than a third threshold, indicating the storage device to store the data in the data table according to the fifth index engine.
The data storage device according to claim 14 or 15, characterized in that the operational amplification includes both the read amplification and the write amplification;

The indication module is used for: when the write amplification is greater than the second threshold and the read amplification is less than a fourth threshold, indicating the storage device to store the data in the data table according to the sixth index engine.
The data storage device according to any one of claims 14 to 16 is characterized in that the fifth index engine includes at least a B+ tree structure, and the sixth index engine includes at least a log structure merged tree structure.
The data storage device according to any one of claims 14 to 17, characterized in that the data table has a local secondary index LSI, and the operation amplification is the amplification generated by operating the data under the local secondary index;

The indication module is used for:

instructing the storage device to store the data under the local secondary index according to the fifth index engine; or,

Instruct the storage device to store the data under the local secondary index according to the sixth index engine.
A computing device cluster, characterized in that it includes at least one computing device, each computing device includes a processor and a memory;

The processor of the at least one computing device is configured to execute instructions stored in the memory of the at least one computing device, so that the computing device cluster executes the method according to any one of claims 1 to 4 or the method according to any one of claims 5 to 9.
A computer program product comprising instructions, characterized in that when the instructions are executed by a computing device cluster, the computing device cluster executes the method according to any one of claims 1 to 4 or the method according to any one of claims 5 to 9.
A computer-readable storage medium, characterized in that it includes computer program instructions. When the computer program instructions are executed by a computing device cluster, the computing device cluster executes the method according to any one of claims 1 to 4 or the method according to any one of claims 5 to 9.