Design And Implement Of A Distributed Architecture That Supports Batch-computing And Realtime-computing

As the rapid development of computer technique, now we are in age of big data. This paper designs and implements a distributed application architecture——DABR that merges the advantage of batch system and realtime system according to two typical types of application requirements:high concurrency low latency system such as online application and complex calculating on constant data set such as compliance system.This paper elaborates how can this application architecture get the advantage of both batch parallel computing and flow parallel computing, how to design this application architecture to get this feature. At the same time, this paper elaborate how to implement this application architecture, how to use interface mode to get good scalability so user can easily expanding any module of this architecture.Then this paper introduces how to implement DABR architecture in detail, introduces all APIs that DABR architecture applies and introduces this architecture’s usage and functionality by implementing SNSTester application.Finally, this paper introduces finance compliance system and usage of DABR architecture in it. Then elaborates how to merge batch parallel computing and flow parallel computing in different scenario.