Automatic Data Replication Using Distributed Datacenters in Online Social Network

Online Social Network (OSN) is a network which consists of real users that interact with each other in various ways. Nowadays the number of users in an OSN increases diversely with respect to time. In this paper we are introducing the new OSN model, which distributes datacenters worldwide, to help decrease service latency which leads higher inter-datacenter communication load. For example in Facebook, recently it becomes very famous and most visited online network site. Each datacenter of Facebook has all the data, which are updated by the master datacenter, leading to remarkable load in the new model. This model uses datacenters to store data at their geographically nearest datacenters. The regular interactions in online network between the various users can generate long service latencies. In this paper, we proposed location based load balancing concept using which user will be redirected to the geographically nearest server and achieving low service latency. For this we proposed the concept of Automatic Data replication concept in Distributed Datacenters (AD3). In this model, AD3 jointly considers update rates and visit rates to create replicas, these replicas are created using selected data for replication. AD3 also includes users various types of data such as friends request, status update, friend post, music or videos update, etc for replication, while considering all these updates AD3 can lower the service latency. AD3 also includes replica updates, replica deactivation, and document security, availability.