A Distributed Transaction Execution Method Based on Dynamic Fragmentation

Abstract

The invention discloses a distributed transaction execution method based on dynamic fragmentation. A dynamic transaction fragmentation unit used for controlling and managing transactions is configured in a database management server in each data center; the dynamic transaction fragmentation unit comprises a dynamic fragmentation module and a dynamic conflict analysis module; the dynamic transaction fragmentation unit is used for processing transaction requests from an application server, performing dynamic fragmentation and dynamic conflict analysis on the transaction requests, executing the transactions subjected to fragmentation by adopting a fragmentation execution module, and executing specific data access operations on a database memory. According to the method disclosed by the invention, the dynamic fragmentation module and the dynamic conflict analysis module are introduced into the data center based on a transaction fragmentation technology, and a programmed automatic transaction fragmentation technology is realized, so that the transactions are submitted to the local data center with more opportunities, the frequency of communication across the data center in a distributed storage system is effectively reduced, and time delay of transaction management is reduced. Therefore, good response performance is achieved.

Description

technical field [0001] The present invention relates to the field of data access, and more specifically, to a distributed transaction execution method based on dynamic sharding. Background technique [0002] A large-scale distributed storage system will place data blocks in different data centers, that is, data partitions. For applications deployed on multiple data centers, transaction operations usually need to access data in different partitions, that is, to access different data centers. According to the transaction sharding technology, a transaction is divided into multiple shards according to the different access data partitions. Each shard will only access the data partition of a single data center, and execute and commit in the form of a single ACID transaction, while the first shard is executed in the local data center where the client is located. After the first shard is executed and submitted, the transaction can return the completed status to the client, and the...

AI Technical Summary

Problems solved by technology

These will bring additional workload

[0006] 2) Lack of flexibility

[0007] 3) Only static data partition schemes can be used

The communication delay across data centers is very high, causing very high time overhead for transaction management

Images