MSF: A Session-Oriented Adaptive Load Balancing Algorithm

Abstract

Some content-oriented applications, such as IDS and IPS; require keeping session integrality when scheduling the workload, otherwise, the semantic integrality of the session will be destroyed. The traditional load balancing algorithms focus on parallel processing and have less consideration for session integrality. In this paper, we proposed an adaptive load balancing algorithm named MSF (minimum sessions first). The algorithm integrates the static workload pre-distribution scheme based on flow table with adaptive scheduling scheme, can obtain the session level load balancing by dynamic scheduling the flow-bundles in which the session number are minimum. The simulation results show that the scheduler achieves significant improvement in session integrality disruption and has a fairly good load balance both in packets level and bits level. It's a sample algorithm and appropriate for the application of session-oriented load balancing. .