MapReduce framework based distributed NMS architecture

In this paper we describe an architecture for a Distributed Network Management System (NMS) for large, geographically spread, heterogeneous networks, integrating Map Reduce framework along with the prevalent Service Oriented Architecture (SOA). MapReduce is the software framework popularized by Google to support distributed computing on large data sets on clusters of computers. In a nutshell, it is a way to take a big task and divide it into discrete tasks that can be done in parallel. This new architecture enhances the robustness, scalability and reliability of NMS. We also discuss the implementation of this architecture using Hadoop, the open source implementation of MapReduce Framework