Distributed transaction processing method based on connection pool management

Abstract

The invention relates to a distributed transaction processing method based on connection pool management. The distributed transaction processing method based on connection pool management comprises the steps that a connection pool is established on a proxy server; when the proxy server receives a first service request, sent by an application server, for transaction processing, persistent connection is established between the proxy server and a relevant database server, the service request is sent by the proxy server to each database server to be processed through the persistent connection, and all following service requests belonging to the transaction processing are sent to each database server to be processed through the established persistent connection; when the service request sent by the application server is the request for ending of the transaction processing or is not a transaction request, the proxy server withdraws the persistent connection established between the application server and each database server. By the adoption of the distributed transaction processing method based on connection pool management, the consistency and the integrity of transaction processing data are guaranteed, the isolation level of transactions is guaranteed, the promotion function on a high-concurrency data service is achieved, and the data consistency of high-concurrency access is guaranteed.

Description

technical field [0001] The invention relates to the field of distributed database and cloud computing, in particular to a distributed transaction processing method based on connection pool management. Background technique [0002] In many key fields such as the Internet and telecommunications, with the continuous development of enterprise business, the I / T environment of enterprises is also constantly evolving. It is difficult for a single database to meet the requirements of massive database storage and high concurrent data access. The emergence of distributed databases has become Inevitably, distributed databases can effectively solve massive data storage and concurrency problems. Through horizontal alignment of data and vertical segmentation of data tables, massive data storage is effectively alleviated, and high Concurrency issues, but how to ensure the consistency, data integrity, and data isolation of each node's data has become a problem that distributed databases nee...

AI Technical Summary

Problems solved by technology

[0006] 1) If there is no persistent connection between the database node and the proxy server, the rollback operation of the data cannot be guaranteed;

[0007] 2) Data consistency cannot be guaranteed in the case of simultaneous operation of multiple nodes;

[0008] 3) The connection between the application server and the proxy server cannot be effectively inherited, and the database node and the proxy server are in a disordered state. The disordered state of the connection makes the database unable to be highly available, and the distributed database is greatly restricted;

[0009] 4) It cannot support services with high data requirements;

[0010] 5) The database transaction isolation level cannot be supported, and some data cannot be read, resulting in the inability to perform row locks and dirty reads

Images