An Efficient Architecture for Network Information Management

Abstract

Session initiation protocol deployments in VoIP networks primarily emerged with an aim to ensure high scalability, simplicity and flexibility. Despite all the listed advantages, the increasing SIP signaling traffic needs to invoke methods attempting to mitigate overload and fail over that the data centers are usually facing. In this paper, a load balancing model based on cluster based replication architecture is proposed to achieve high capacity and a highly available architecture. We quantitatively measure the improvements by implementing asterisk and openSIP based server setup for voice exchange for the indication of scalable and redundant architecture. In the past years, high availability was achieved by host per host redundancy declaring it as an expensive technique for hardware. Our proposed technique is suitable for a redundant system with mission critical applications and can be offered to extend the service level by sharing the running virtual machines while duplicating the resources. The proposed model implemented in our test-bed configured three asterisk servers and the node statistics are tested via cacti server to guarantee the failover technique. The experimental results establish that the proposed model reduce low server utilization and lack of power proportionality. The proposed solution is unique in its way that it is also recommended for complex and unmanageable environment without relying on add on balancing methods.