JP2016522928A - Method, apparatus and system for establishing secondary indexes in a distributed storage system - Google Patents

Abstract

The present invention is a method for establishing a secondary index in a distributed storage system, comprising receiving user data transmitted by a user client, wherein the user data includes a key portion and a value portion; Determining a user table according to user data; writing user data into the user table; generating index data according to user data; writing index data into an index table corresponding to the user table; The user table and the index table are related to a method having a step of writing, which is established in one node of the distributed storage system and combined in a one-to-one manner. The present invention exists in that by combining the user child table and the index child table, two remote write operations across the nodes are reduced to one, greatly reducing the number of network requests, thereby improving performance. .

Description

  This application relates to distributed storage systems, and more particularly, to a method, apparatus and system for establishing secondary indexes in a distributed storage system.

  In the majority of distributed storage systems, a key-value type storage method has been accepted, i.e. the actual user data to be stored is stored in the value part, and then the key is configured to search for the corresponding value. Is done.

  For example, assume that a distributed storage system is used to store data in an online trading system, and Key and Value can be designed as follows.

Key: User code + trading time Value: Detailed information for trading When real user data is stored in a distributed storage system, the data of the same user code is stored in the dictionary order so that it is stored adjacently. Ordered. An indexing scheme in a distributed storage system can be used to quickly search for desired data in order to obtain all records having a user code as “xxxx” and transaction time as a period.

In a distributed storage system, the concept of a table is provided. If the user wants to write data to the distributed storage system, he / she should first build the data table according to his / her own requirements. However, a large amount of data may be included in one data table. In order to achieve distributed storage in a distributed storage system, the primary method is to generate multiple child tables by cutting the data table in the horizontal direction, and manage and maintain the multiple child tables. A child table can be defined as follows:
1. A child table is a cluster of a plurality of consecutive rows, and each child table has a key value range.
2. A database table is usually composed of one or more child tables.
3. The child table is the smallest unit of distributed storage and load balancing.
4). When increased to some extent, the child table is automatically split into two child tables.

  FIG. 1 is a schematic diagram of a data table. As shown in FIG. 1, the data table includes M child tables, and each child table occupies a key value range. The key to which each key value generated by the user corresponds can belong only to the key value range of one child table. The key value ranges of different child tables have no common part.

  Each child table needs to be assigned and can be uniquely assigned to only one managed area server. An area server is a node of a cluster of a distributed storage system, which is usually a physical server.

  FIG. 2 is a schematic diagram of the cluster of the distributed storage system. As shown in FIG. 2, the cluster of the distributed storage system includes three child servers (child table 1) stored in the three region servers and the first region server. , Child table 4 and child table 9), three child tables stored in the second area server (child table 2, child table 6 and child table 7), and three child tables stored in the third area server ( It has a child table 3, a child table 5 and a child table 8).

  In a distributed storage system, all user data is stored in the underlying file system in the form of key values. To retrieve user data, one way is to specify the correct key to perform the search correctly, and the other is to specify the previous part of the key to perform the data scan ambiguously. . However, both these two methods are based on the value of the key, and when the user wants to search for user data based on the state of the value, it is necessary to search the entire data table to find the record. The performance is very poor.

  In order to solve this problem, a secondary index method is provided. FIG. 3 is a schematic diagram of a conventional secondary index method. As shown in FIG. 3, the secondary index method is based on a client API (Application Programming Interface) package of the distributed storage system. In the secondary index method, another index table is used to store index values, and when a user writes data, two types of data can be generated by using a packaged client API, Is user data written to the user table, and the other is index data written to the index table.

  In an implementation of the present invention, the inventor has discovered that conventional secondary index schemes lead to poor performance because they are packaged in the client API. There are several other improved schemes, but to write the data, two remote write processes need to be sent between the client and the remote distributed storage system, writing the user table and the index table Therefore, it is basically impossible to improve the performance of write data.

  In view of the problems pointed out in the background art, the present invention is proposed.

  The main problem of embodiments of the present invention is to provide a method, apparatus and system for establishing secondary indexes in a distributed storage system in order to reduce the number of network requests.

According to an aspect of an embodiment of the present invention, a method for establishing a secondary index in a distributed storage system comprising:
Receiving user data transmitted by a user client, wherein the user data includes a key portion and a value portion;
Determining a user table according to the user data;
Writing the user data to the user table;
Generating index data according to the user data;
Writing the index data into an index table corresponding to the user table, wherein the user table and the index table are established in one node of the distributed storage system and combined in a one-to-one manner When,
Is provided.

According to another aspect of an embodiment of the present invention, an apparatus for establishing a secondary index in a distributed storage system, comprising:
A receiving unit configured to receive user data sent by a user client, wherein the user data includes a key portion and a value portion;
A determination unit configured to determine a user table according to the user data;
A first writing unit configured to write the user data to the user table;
A generating unit configured to generate index data according to the user data;
A second writing unit configured to write the index data to an index table corresponding to the user table, wherein the user table and the index table are established in one node of a distributed storage system, in a one-to-one manner A second writing unit coupled;
A device is provided.

According to yet another aspect of an embodiment of the present invention, a distributed storage system is at least one region server, each region server having at least one user table and at least one user table. Storing at least one corresponding index table, wherein each range key of the user table is the same as that of the index table corresponding to the user table; and
The secondary index establishment device described above;
A distributed storage system is provided.

  The effect of the embodiment of the present invention is that by combining the user child table and the index child table, the two remote write processes across the nodes are reduced to one, the number of network requests is greatly reduced, and the performance is improved accordingly. It exists in the point to do.

  Specific embodiments of the present invention are described in detail below with reference to the following description and the accompanying drawings, pointing out the manner in which the principles of the invention are utilized. It should be understood that the implementation of the invention is not limited in scope. Rather, the invention includes all changes, modifications and equivalents falling within the spirit and conditions of the appended claims.

  Features described and / or illustrated with respect to one embodiment may be utilized in one or more other embodiments in the same or similar manner and / or in connection with or instead of features of other embodiments. Good.

  As used herein, the term “comprising” is used to indicate the presence of a stated feature, integer, step or component, but one or more other features, integers, steps, components or these It should be emphasized that it does not exclude the presence or addition of other groups.

The drawings are included to provide a further understanding of the invention, which forms a part of the specification and illustrates the preferred embodiment of the invention and is used in conjunction with the description to explain the principles of the invention. The same elements are denoted by the same reference numerals throughout the drawings.
FIG. 1 is a schematic diagram of a data table. FIG. 2 is a schematic diagram of a cluster of a distributed storage system. FIG. 3 is a schematic diagram of a conventional secondary index method. FIG. 4 is a schematic diagram of a distributed storage system according to an embodiment of the present invention. FIG. 5 is a flowchart of a method for establishing a secondary index according to an embodiment of the present invention. FIG. 6a is a schematic diagram of an embodiment of a user child table. FIG. 6b is a schematic diagram of an embodiment of an index data table. FIG. 6c is a schematic diagram of another embodiment of the index data table. FIG. 7 is a schematic diagram of a user child table and its corresponding index child table stored in one region server. FIG. 8 is a schematic diagram of a secondary index establishment device. FIG. 9 is a schematic diagram of a distributed storage system.

  These and other features of embodiments of the present invention will become apparent through the following description with reference to the drawings. These examples are illustrative only and are not intended to limit the invention. For easy understanding of the principles and embodiments of the present invention by those skilled in the art, the description of the embodiments of the present invention is provided by taking an online transaction system as a specific example, but the embodiments of the present invention are such systems. It should be understood that the invention is not limited to.

  FIG. 4 is a schematic diagram of a distributed storage system according to an embodiment of the present invention. As shown in FIG. 4, the distributed storage system includes two area servers, that is, an area server 41 and an area server 42. Each area server has a user table and an index table. The user table and index table established in one region server are combined in a one-to-one manner, that is, the key range of the user table and the index table are the same. Each user table may be divided into a plurality of user child tables, and each index table may be divided into a plurality of index child tables. One of the plurality of index child tables corresponds to one of the plurality of user child tables, and the number of the plurality of user child tables in one region server is equal to the number of the plurality of index child tables in the same region server. Are the same. The user child table and index child table established in one region server are combined in a one-to-one manner, that is, the key range of each user child table is the same as that of its corresponding index child table. In the embodiment of the present invention, the user table 411 and the index table 412 are established in the area server 41, and the user table 421 and the index table 422 are established in the area server 42.

  When the user table 411 includes the user child table 4111 and the user child table 4112, and the index table 412 includes the index child table 4121 and the index child table 4122, in the embodiment of the present invention, the user child table 4111 and the index child table 4121 are included. Are combined in a one-to-one manner, that is, the key range of the user child table 4111 and that of the index child table 4121 are the same. Similarly, the user child table 4112 and the index child table 4122 are combined in a one-to-one manner.

  On the other hand, the user table 412 has a user child table 4211 and a user child table 4212. The index table 422 has an index child table 4221 and an index child table 4222. Similarly, in the embodiment of the present invention, the user child table 4211 is provided. The index child table 4221 is coupled by a one-to-one method, and the user child table 4212 and the index child table 4222 are coupled by a one-to-one method.

  In an embodiment of the present invention, the user table and the index table are established in the same node (region server), combined by a one-to-one method, and when user data is written by the user, the user side is the data related to the user table. Need to send only the data related to the index table, i.e. only one process of writing the user table is required, the other process of writing the index table is avoided, the data of the index table Is writable by a call in the process, can improve the performance of the data writing stage, and is particularly suitable for a user scene that requires high data writing performance.

  Preferred embodiments of the invention are described below with reference to the drawings.

Example 1
Embodiments of the present invention provide a method for establishing a secondary index in a distributed storage system. FIG. 5 is a flowchart of the method, and with reference to FIG.
Step 501: Receiving user data sent by a user client, wherein the user data includes a key portion and a value portion;
Step 502: determining a user table according to the user data;
Step 503: writing user data to the user table;
Step 504: generating index data according to the user data;
Step 505: writing index data to an index table corresponding to the user table, wherein the user table and the index table are established in one node of the distributed storage system and combined in a one-to-one manner;
Have

  In step 501, each table has a key range from the start key to the end key, and the range key of user data is usually in the range from the start key to the end key of the table. In general, in embodiments of the present invention, user data includes a key portion and a value portion, where the key portion of user data may be divided into two portions, one being a range key and the other being a specific key. The range key corresponds to any key from the start key to the end key and is used to indicate which user table the user data belongs to, and a specific key is used to uniquely indicate the user data Is done. Thus, according to the range key of the key portion of the user data, the distributed storage system can determine which user table is required to store the user data.

  In step 503, after determining the user table, the distributed storage system can execute a process of writing user data, and the method of writing user data is the same as in the prior art and will not be further described.

  In step 504, the distributed storage system can generate index data by utilizing user data achievable by in-process calls, thus avoiding other processing requests of the distributed storage system for index data from user clients. A method of generating index data by a distributed storage system can save network resources and can be more efficient compared to a conventional user client API packaging scheme.

  The generated index data has a key part, the key part of the index data has two parts, a range key and a specific key, and the index data range key is used to indicate the index table to which the index data belongs. The specific key of the index data is used to uniquely indicate the index data. In the embodiment of the present invention, since the key range of the user table is the same as that of the index table, the index data range key corresponds to the user data range key, or the index data range key is the start of the user table. The key may be adopted, or the end key of the user table may be adopted as the range key of the index data.

  In a specific example of step 504, the distributed storage system extracts the range key of the key portion of the user data as the range key of the key portion of the index data. Details are described below.

  In step 505, since the user table and the index table have a one-to-one relationship, the distributed storage system can write the index data to the index table corresponding to the user table.

  Each user table may be divided into a plurality of user child tables, and each index table may be divided into a plurality of index child tables. In step 502, the distributed storage system further determines a user child table according to the user data, writes the user data to the user child table in step 504, and in step 505 stores the index data in the index child table corresponding to the user child table. Write.

  FIG. 6a is a schematic diagram of an embodiment of a user child table, in which certain user data is stored in the user child table, and like the user data, the user child table also has a key portion and a value portion, The key portion of the user child table is used to store the key portion of the user data, and the value portion of the user child table is used to store the value portion of the user data. As shown in FIG. 6a, the key portion of the user child table has one column named “Key” and the value portion of the user child table has one column named “Seller Code”. However, it should be understood that the embodiments are not limited to this, and in other embodiments, the value portion of the user child table can also have multiple columns. As shown in FIG. 6a, the user data includes a key portion and a value portion. In the embodiment, the key portion of the user data is “buyera1201207011000” and the value portion of the user data is “sellerc1”. The key part of the user data has two parts, one is a range key and the other is a specific key. In the example shown in FIG. 6a, the range key is “buyera” and the specific key is “2012077011000”. FIG. 6b is a schematic diagram of an embodiment of an index child table, in which the index child table corresponds to the user child table shown in FIG. 6a, where the key range of the index child table is that of the user child table. Means the same. As shown in FIG. 6b, the index data is generated according to user data stored in the user child table as shown in FIG. 6a. As shown in FIG. 6b, the index data also includes a key portion and a value portion, the value portion is null, and the key portion of the index data has a range key and a specific key. The index data range key is “buyera”, which is the start key of the user table in FIG. 6 a or the user data range key in FIG. 6 a. Furthermore, the specific key of the index data is “cellerc1 buyer 1201207011000”, where “sellerc1” is the value of the index column in the value part of the user data, and “buyera 1201207011000” is the key part of the user data. As shown in FIG. 6b, the key portion of the index data in this embodiment is “buyera seller1 buyera1201207011000”.

  In other embodiments, the column name of the index column of the user child table can also be included in the index data as shown in FIG. 6c, and unlike the embodiment of FIG. It is included in the index data arranged between “buyera” and “sellerc1”, and the key part of the index data in the present embodiment is “buyera sellercode sellerc1 buyera1201207011000”.

  In one embodiment of the present invention, a user data range key (or a user child table start key or a user child table end key), a user data value portion, and a user data key portion are sequentially arranged in the index data. . In another embodiment of the present invention, the value portion of the user data includes an index column value.

  In the embodiments shown in FIGS. 6b and 6c, the value portion of the index child table is empty, i.e., the index data includes only the key portion, but the embodiments are not limited to this, and in other embodiments, the index data It should be understood that the value portion of the data can be filled with other useful information according to actual requirements.

  In one embodiment of the present invention, as shown in FIG. 4, the user child table and the index child table corresponding to the user child table are stored in the same region server.

  For an easy understanding of the embodiments of the present invention, a detailed description with reference to the drawings is given below.

There are three user child tables in the user table, and the key range of each user child table is
User child table 1: [a, b);
User child table 2: [b, c);
User child table 3: [c, d)
Assume that

Accordingly, there are three index child tables in the index table, and the key range of each index child table is the same as that of the user child table to which the index child table corresponds.
Index child table 1: [a, b);
Index child table 2: [b, c);
Index child table 3: [c, d)
That is, the user child table 1 corresponds to the index child table 1, the user child table 2 corresponds to the index child table 2, and the user child table 3 corresponds to the index child table 3.

  As described above, since all data is stored in the distributed storage system in the order of the dictionary, when the user data of the key “a0001” is written by the user, the user data is stored in the [a, Obviously, user data should be written to the user child table 1 because it belongs to the key range of b). According to an embodiment of the present invention, user data was written to the user child table 1 as described in steps 501 to 503.

  In the embodiment, the corresponding index data is generated and stored in the index child table 1 as described in steps 504 to 505. That is, the corresponding index data also belongs to the key range [a, b) of the index child table 1.

  According to an embodiment of the present invention, the user data range key “a” is placed at the start of the key portion of the index child table to ensure that the generated index data belongs to the corresponding index child table.

  In the embodiment, the structure of the index data is defined as the corresponding user data range key + user data value part (index column name or index column name and value) + user data key part. This is merely a specific example, and the present invention is not limited to this. For example, except for the corresponding user data range key, the index column name position, the index column value, and the user data key portion can be any combination of these.

  FIG. 7 is a schematic diagram of a user child table and its corresponding index child table stored in one region server. As shown in FIG. 7, the key range of the user child table 1 is [buyera, buyerb), and thus the key range of the corresponding index child table 1 is also [buyera, buyerb). According to an embodiment of the present invention, three records of user data are written to the user child table 1, and thus three records of index data are generated and written to the index child table 1. For the third record, the key part of the user data is “buyera120120208101100”, the value part of the user data is “sellerc3”, and therefore the key part of the index data is “buyera sellercode seller3 buyera1201208101100”, and thus the user Data range keys are written in the key portion of the index data, and when the user wishes to obtain all transaction records for seller code “sellerc3”, these records are stored adjacent to the index child table. Therefore, it is easy to read these records, and by decoding the data, the user supports the user child table according to the user data. To obtain the desired data by Ji, it is possible to obtain user data "buyera1201208101100".

  According to an embodiment of the present invention, by combining the user child table and the index child table, two remote write processes across the nodes are reduced to one, and data reading is no longer required between nodes. And the number of network requests is greatly reduced and performance improves accordingly.

  Embodiments of the present invention further provide a secondary index establishment device, and the principle of the device for solving the problem is that of the method of establishing a secondary index in Embodiment 1 as described in Embodiment 2 below. For the specific implementation thereof, the implementation of the method of Example 1 can be referred to. Similarities are not further explained.

Example 2
Embodiments of the present invention further provide a secondary index establishment device. FIG. 8 is a schematic diagram of the apparatus, and with reference to FIG.
A receiving unit 81 configured to receive user data transmitted by a user client, wherein the user data includes a key portion and a value portion;
A determination unit 82 configured to determine a user table according to user data;
A first writing unit 83 configured to write user data to a user table;
A generating unit 84 configured to generate index data according to user data;
A second writing unit 85 configured to write index data to an index table corresponding to a user table, wherein the user table and the index table are established in one node of the distributed storage system and combined in a one-to-one manner A second writing unit 85;
Have

  In the embodiment, the key range of the user table and that of the index table are the same.

  In an embodiment, the user table has a plurality of user child tables, the index table has a plurality of index child tables, and the number of the plurality of user child tables is the same as the number of the plurality of index child tables, The key range of the user child table is the same as that of its corresponding index child table.

  In the embodiment, the first writing unit 83 includes a first determination module 831 and a first writing module 832, and the first determination module 831 is configured to determine the user child table according to the user data. 832 is configured to write user data to the user child table.

  In an embodiment, the second writing unit 85 has a second writing module 851, and the second writing module is configured to write index data to an index child table corresponding to the user child table.

  In an embodiment, the generation unit 84 includes a second determination module 841 and a generation module 842, the second determination module 841 being configured to determine a range key for index data and a specific key for index data, The range key has a user data range key, a user child table start key or a user child table end key, the specific key of the index data has a value portion of user data and a key portion of user data, The generation module 842 is configured to generate the index data according to the index data range key and the index data specific key. Here, the second determination module 841 can determine the range key of the index data by extracting the range key of the user data, the start key of the user child table, or the end key of the user child table. Here, the value part of the user data adopted in the index data is related to the value of the index column, and the index data further has the name of the index column in the user child table.

  In the embodiment, the name of the index column, the value of the index column, and the key part of the user data are sequentially arranged in the index data.

  According to an embodiment of the present invention, by combining the user child table and the index child table, two remote write processes across the nodes are reduced to one, and data reading is no longer a process that requires data between nodes. , The number of network requests is greatly reduced and performance is improved accordingly.

Example 3
Embodiments of the present invention further provide a distributed storage system. FIG. 9 is a schematic diagram of the system. Referring to FIG. 9, the system includes at least one area server 91 and a secondary index establishment device 92.
Each region server stores at least one user table and at least one index table corresponding to the at least one user table, and each range key of the user table is the same as that of the index table corresponding to the user table. Yes,
The secondary index establishment device 92 can be realized by a device as described in the embodiment, and the contents in the second embodiment are included here and will not be further described.

  Here, the storage form of the user table and the corresponding index table stored in the same area server 91 is the same as that described in FIG. 4, the contents of which are included here, and will not be described further.

  In a distributed storage system, child table migration is usually done in consideration of region server operation or load balancing during damage, and when the child table is migrated to another region server, the user child table and index child The tables are migrated together.

  According to an embodiment of the present invention, by combining the user child table and the index child table, two remote write processes across the nodes are reduced to one, and data reading is no longer a process that requires data between nodes. , The number of network requests is greatly reduced and performance is improved accordingly.

  The preferred embodiment of the present invention has been described above with reference to the drawings. Many features and advantages of an embodiment will be apparent from the detailed description, and thus, it is the appended claims that cover all such features and advantages of the embodiment within its true spirit and scope Intended by. Further, since numerous modifications and changes will readily occur to those skilled in the art, it is not desired to limit embodiments of the invention to the exact configuration and process illustrated and described, and thus, all suitable modifications and changes It may be argued that equivalence belongs within its scope.

  It should be understood that each part of the present invention may be realized by hardware, software, firmware or a combination thereof. In the above embodiment, the steps or methods may be implemented by software or firmware stored in memory and executed by a suitable instruction execution system. For example, if it is implemented by hardware, it, like the other embodiments, has the following techniques known in the art: a logic gate circuit for implementing the logic function of the data signal Any one or more of discrete logic circuits, application specific integrated circuits with appropriate synthesized logic gate circuits, programmable gate arrays (PGA) and field programmable gate arrays (FPGAs), etc. It may be realized by a combination.

  A description or block of any process or method in a flowchart or other manner may have one or more modules, segments, or parts to implement the code of executable instructions for a step in a particular logical function or process. The scope of the preferred embodiments of the present invention may have other implementations, and the functions perform functions according to related functions in a substantially simultaneous manner or in the opposite order. May be implemented in a manner different from that shown or described, and should be understood by those skilled in the art to which this invention belongs.

  The logic and / or steps illustrated in the flowcharts or described in other manners herein may include, for example, an instruction execution system, device, or apparatus (a system having a computer, a system having a processor, or an instruction execution system, device, or apparatus). Logic that may be implemented on any computer-readable medium for use by, for example, other systems capable of extracting instructions from and executing with the instructions, or for use with an instruction execution system, device, or apparatus It may be understood as a sequenced list of executable instructions to implement the function. As used herein, a “computer-readable medium” can be any device that can store, store, communicate, propagate, or transmit a program for use by an instruction execution system, device, or apparatus. Or with an instruction execution system, device or apparatus. The computer readable medium may be, for example, without limitation, an electronic, magnetic, optical, electromagnetic, infrared, or semiconductor system, device, apparatus, or propagation medium. More detailed examples (non-comprehensive list) of computer readable media include the following: electrical connections (electronic devices) having one or more wires, portable computer hardware boxes (magnetic devices), random access Memory (RAM) (electronic device), read-only memory (ROM) (electronic device), erasable programmable read-only memory (EPROM or flash memory) (electronic device), optical fiber (optical device) and portable compact disc read-only You may have a memory (CDROM) (optical device). Further, when the program is obtained electronically through optical scanning of paper or other suitable media, and then compiling, interpreting or processing in other suitable manner as required, computer storage The medium may be paper or other suitable medium on which the program can be printed, after which the program is stored in computer memory.

  The above text description and drawings illustrate various features of the present invention. It should be understood that one skilled in the art may prepare computer code suitable for performing each of the steps and processes described above and illustrated in the drawings. The terminals, computers, servers, networks, etc. described above may be of any type, and the computer code may also be prepared according to the disclosure contained herein to implement the invention by utilizing a device. Should be understood.

  Particular embodiments of the present invention have been disclosed herein. Those skilled in the art will readily recognize that the present invention is applicable in other environments. In fact, there are many examples and implementations. The appended claims are not intended to limit the scope of the invention to the specific embodiments described above. Further, any reference to “a device to do” is a description of the device plus function to explain the elements and claims, and any reference that does not use the “device to do” reference. An element is not desired to be understood as a device plus function element even if the word “device” is included in the claim.

  While the particular preferred embodiment has been shown and the invention has been described, it will be apparent that equivalent modifications and changes will occur to those skilled in the art upon reading and understanding the description and drawings. In particular, unless otherwise noted for the various functions performed by the above elements (parts, assemblies, devices, configurations, etc.), the elements perform the functions of the exemplary embodiments shown in the present invention with respect to the configuration. Although the terminology used to describe these elements (including references to “devices”) may differ from any element that performs a particular function of these elements. Furthermore, while specific features of the invention have been described with respect to only one or more of the illustrated embodiments, such features may be considered in any given or effective aspect of a particular application, It may be combined with one or more other features of other embodiments as desired.

Claims (19)

A method for establishing a secondary index in a distributed storage system, comprising:
Receiving user data transmitted by a user client, wherein the user data includes a key portion and a value portion;
Determining a user table according to the user data;
Writing the user data to the user table;
Generating index data according to the user data;
Writing the index data into an index table corresponding to the user table, wherein the user table and the index table are established in one node of the distributed storage system and combined in a one-to-one manner When,
Having a method.   The method of claim 1, wherein the key range of the user table is the same as that of the index table. The user table has a plurality of user child tables, the index table has a plurality of index child tables,
The number of the plurality of user child tables is the same as the number of the plurality of index child tables;
The method of claim 2, wherein the key range of each user child table is the same as that of its corresponding index child table. Writing the user data to the user table comprises:
Determining a user child table according to the user data;
Writing the user data to the user child table;
Have
Writing the index data to the index table comprises:
4. The method of claim 3, comprising writing the index data to an index child table corresponding to the user child table. The step of generating index data according to the user data includes:
Determining a range key of the index data and a specific key of the index data, wherein the range key of the index data is a range key of the user data, a start key of the user child table or a key of the user child table; Determining an end key, and wherein the particular key of the index data comprises a value portion of the user data and a key portion of the user data;
Generating the index data according to a range key of the index data and a specific key of the index data;
The method of claim 4, comprising: Determining a range key of the index data;
6. The method according to claim 5, comprising the step of extracting a range key of the user data, a start key of the user child table or an end key of the user child table.   The method according to claim 5, wherein a value portion of the user data adopted for the index data is related to a value of an index column.   The method of claim 7, wherein the index data further comprises an index column name in the user child table.   The method of claim 8, wherein the name of the index column, the value of the index column, and the key portion of the user data are sequentially arranged in the index data. A secondary index establishment device,
A receiving unit configured to receive user data sent by a user client, wherein the user data includes a key portion and a value portion;
A determination unit configured to determine a user table according to the user data;
A first writing unit configured to write the user data to the user table;
A generating unit configured to generate index data according to the user data;
A second writing unit configured to write the index data to an index table corresponding to the user table, wherein the user table and the index table are established in one node of a distributed storage system, in a one-to-one manner A second writing unit coupled;
A secondary index establishing device.   The apparatus of claim 10, wherein a key range of the user table and that of the index table are the same. The user table has a plurality of user child tables, the index table has a plurality of index child tables;
12. The number of the plurality of user child tables is the same as the number of the plurality of index child tables, and the key range of each user child table is the same as that of its corresponding index child table. Equipment. The first writing unit includes:
A first determination module configured to determine a user child table according to the user data;
A first writing module configured to write the user data to the user child table;
Have
13. The apparatus of claim 12, wherein the second writing unit comprises a second writing module configured to write the index data to an index child table corresponding to the user child table. The generating unit is
A second determination module configured to determine a range key of the index data and a specific key of the index data, the range key of the index data being the range key of the user data, the start of the user child table A second determination module having a key or an end key of the user child table, wherein the specific key of the index data comprises a value portion of the user data and a key portion of the user data;
A generation module configured to generate index data according to the index data range key and the index data specific key;
14. The apparatus of claim 13, comprising:   The apparatus of claim 14, wherein the second determination module is configured to extract a range key of the user data, a start key of the user child table, or an end key of the user child table.   The apparatus of claim 14, wherein a value portion of the user data employed in the index data is associated with a value of the index column.   The apparatus of claim 16, wherein the index data further comprises an index column name in the user child table.   The apparatus according to claim 17, wherein the name of the index column, the value of the index column, and the key part of the user data are sequentially arranged in the index data. A distributed storage system,
At least one region server, each of the region servers storing at least one user table and at least one index table corresponding to the at least one user table, wherein each range key of the user table is: An area server identical to that of the index table corresponding to the user table;
A secondary index establishment device according to any one of claims 10 to 18,
A distributed storage system.

Images