Hybrid cloud infrastructure to handle large scale data for bangladesh people search (BDPS)

The objective of our paper is to propose and experiment a hybrid Cloud computing framework which is able to handle huge data. Previously we proposed this architecture with adequate feasibility analysis. Here we discuss a prototype system using national ID database structure of Bangladesh which is prepared. We implemented and experimented each of the modules of the system prototype. The infrastructure of our proposed system divided into two parts: locally hosted cloud and remote cloud. Remote cloud is experimented with Elastic MapReduce (EMR) of Amazon Web Service (AWS) and “Eucalyptus” is justified for local cloud. This structure of cloud computing is handled by apache Hadoop map/reduce framework where database is implemented and experimented by HiveQL. An experimental analysis is presented in this paper. User information visibility, authorization, authentication and integrity issues are also discussed in our literature review.