Distributed B+ tree index system and building method

Abstract

The invention relates to a distributed B+ tree index system and a building method. The distributed B+ tree index system is characterized by comprising a master server, a transaction server cluster, anindex server cluster and a version control server; the transaction server cluster comprises a plurality of transaction servers, and the index server cluster comprises a plurality of index servers; the master server is in charge of managing META data and conducts load balance scheduling on the index server cluster; the transaction server cluster is in charge of transaction control of index data access in a distributed file system; and the index server cluster is in charge of managing, reading and writing the index data in the distributed file system. The invention provides an index transactionmechanism with fine granularity and low network flow. As the basic transaction operation in the distributed B+ tree index system is based on key assignment granularity of a B+ tree and only dozens ofbytes of B+ tree index key assignment pairs are needed to be transmitted when being the basic transaction is operated, the distributed B+ tree index system effectively realizes the transaction function of the index data in a concurrent environment.

Description

technical field [0001] The invention relates to a data index system and a construction method, in particular to a distributed B+ tree index system and a construction method used in a cloud storage environment. Background technique [0002] With the development of Internet technology and the advent of the Web2.0 era, due to the need to process and calculate massive data in the application process, such an application environment puts forward more requirements for the scalability, reliability and efficiency of the Web application data management system. high demands. With the maturity of cloud environment technology, cloud storage system shows more and more obvious advantages in today's Web data management. [0003] Cloud storage refers to a system that integrates a large number of different types of storage devices in the network through application software to work together to provide data storage and business access functions through cluster applications, grid technology, ...

AI Technical Summary

Problems solved by technology

Existing cloud storage systems such as Bigtable, Hbase, PNUTS, etc., all have some deficiencies in the indexing of massive data, mainly including: 1) The existing systems only have indexes on the primary key of the data, and lack of data indexing. Effective indexing methods for other attributes; if the client application server needs to query based on the retrieval conditions on some other attributes of the data, the system can only implement the query by traversing all the data, which makes the query efficiency very low

2) The existing system lacks an effective concurrency control transaction mechanism in index operations; such defects directly lead to the fact that the client application server must use its own code to ensure transaction consistency in a concurrent environment when using the index interface, which makes the client The development of the application server becomes complicated, and there is a lack of universal support for different index operations

Images