Software-defined network assisted packet scheduling method for load balancing in mobile user concentrated cloud

Abstract

Cloud Network provides a variety of applications and services to meet the requirements of increasing user demands. With its dedicated components, the cloud enables anywhere anytime access of resources to its associated users. A significant factor that holds the performance of the cloud is its congestion, due to unpredictable traffic and asynchronous user demands. Though load balancers serve the purpose of optimizing network traffic, congestion is unavoidable when the level of user demand increases. In this manuscript, a Resource-aware Packet-level Scheduling with Load Balancing (RPS-LB) algorithm is proposed to minimize congestion. In RPS, the incoming network traffic is disseminated through instantaneous neighbors by analyzing their Store and Forward (SF) factor. SF factor is pre-estimated through Estimated Transmission Count (ETX) metric. With this analysis, the reliability of the instantaneous neighbor for handling the current packet-level transmission is verified. This algorithm addresses the issues in user-level and flow-level to mitigate stagnancy and overflow of network traffic. Besides, the proposed load-balancing algorithm is defined as a linear optimization problem within definite constraints to prevent degradation proportional to user density. The performance of the proposed system is assessed by using the performance evaluation parameters such as Queue Utilization, Request Success Rate, Request Failure Rate, Response Time and Link Utilization. The simulation results show that the proposed system performs well than the existing systems.