Distributed transaction processing system

Abstract

The invention provides a distributed transaction processing system, which comprises a transaction configuration module used for configuring a business service name, a business service method name anda transaction mode type of a business service according to an application scene of the business service; a server which comprises a first distributed transaction control component used for obtaining aglobal transaction identifier of a distributed transaction after receiving a distributed transaction starting request of service orchestration and according to the global transaction identifier of the distributed transaction and the type of the transaction mode of the business service in the service orchestration, controlling the configured business service in the distributed transaction in combination with a second distributed transaction control component; a global unified signal sending component which is used for generating a global transaction identifier of the distributed transaction;and a client which comprises a second distributed transaction control component which is used for controlling the configured business service in the distributed transaction in combination with the first distributed transaction control component. According to the invention, the data consistency problem of distributed transactions can be solved.

Description

technical field [0001] The invention relates to the field of computer technology, in particular to a distributed transaction processing system. Background technique [0002] In the process of transforming the traditional large-scale monolithic enterprise architecture to the microservice architecture, the monolithic system is divided into microservices that can be deployed independently according to the business value, so as to achieve the purpose of high availability, flexible scaling, and agile development. However, due to the independent deployment of service databases under the microservice architecture, how to maintain data consistency between services has become a major problem in the transformation process. [0003] Most of the existing distributed transaction solutions are implemented according to specific scenarios, and the interfaces between the various solutions are not uniform. The application system needs to be customized and secondary developed according to diff...

AI Technical Summary

Problems solved by technology

The main disadvantages of the 2PC scheme: 1) Synchronous blocking: Since both stages use synchronous blocking, it will cause performance loss

2) The transaction manager has a single point of failure problem. If the second stage fails, other participants will always be locked

The main disadvantages of the TCC solution: 1) It is relatively intrusive to business services, and business services need to be transformed to provide Try, Confirm, and Cancel interfaces; 2) It is difficult to implement, and the business service layer needs to provide different rollbacks for failure reasons in different situations Strategy; 3) When the business service implements the Confirm and Cancel interfaces, it needs to ensure idempotence

The main disadvantage of the Saga scheme: it does not provide transaction isolation guarantees: each sub-transaction is submitted in turn, and at this time another transaction will read the data of the intermediate process when the resource is queried, so if the same resource is operated at the same time, it may cause operation coverage. Dirty data and other issues

The main disadvantages of eventual consistency based on message queues: 1) The downstream system needs to negotiate with the upstream system on how to achieve idempotence

[0010] The above existing technical solutions are relatively independent, and each technical solution has multiple implementations. For the above business scenarios, it is necessary for each business service to implement the corresponding functions or refer to the implementation of the third party, which will cause interfaces and standards between services. Inconsistent, repeated development, the scale of the system is difficult to expand, thus restricting the continuous evolution of the system architecture, and it is difficult to form a unified enterprise-level microservice business middle platform architecture

[0011] To sum up, there is currently a lack of a unified distributed transaction processing system to solve the problem of data inconsistency in distributed transaction processing

Images