Concurrent access control method in distributed database system

Abstract

The invention discloses a concurrent access control method in a distributed database system. According to the invention, the transaction number is generated by using the timestamp of the client and the hardware unique identifier. And when the read-only transaction accesses a certain row of data for the first time, reading the latest submitted version of the row of data, and recording the transaction number of the version in the client to provide evidences for repeated access. And when the read-write transaction reads a certain row of data, reading the latest submitted version of the row of data, and adding a read lock of the row in a memory of a node where the row is located. And when the read-write transaction updates a certain row of data, caching the transaction number of the current version of the row and the updated content in the client. When the affairs are submitted, the updated content is sent to the corresponding node, whether updating is feasible or not is determined according to the current state of the line, and when waiting or rollback is needed, one affair rolls back according to the sequence relation of the affair numbers. According to the invention, deadlock detection with extremely high cost in a distributed environment is avoided, and high concurrency and high consistency are obtained.

Description

technical field [0001] The invention belongs to the field of information, in particular to a concurrent access control method in a distributed database system. Background technique [0002] Since the chip manufacturing process is close to the limit, the main frequency of the central processing unit of the computer has not been increased in recent years, and the rapid development of technologies such as the Internet and the Internet of Things has created a business demand for massive data processing. Therefore, data processing technology has begun to scale up. Development in the direction of distributed processing. The current distributed database software is often built on hundreds or thousands of computers, which cooperate to process data through LAN or WAN communication, and provide external services as a whole, allowing upper-level applications to access the system like accessing a stand-alone database. [0003] In a distributed database, the most difficult problem to de...

AI Technical Summary

Problems solved by technology

But the problem is: 1) an incremental timestamp will also cause a bottleneck in system scalability, because all transactions must access the timestamp generator synchronously when they start and commit; 2) when two transactions update each other, dead

Images