Context-Aware Adaptive Packet Scheduling for Real-Time Services in Computer Networks

Abstract

Voice over IP (VoIP) and video streaming are real-time network applications in which delay, jitter and packet loss are hard requirements that a conventional packet scheduling system will not meet in dynamic traffic scenarios. The available scheduling strategies mostly use a static priority which is not effective in reacting to dynamism in the network environment. It is in this paper that we come up with a context-sensitive adaptive packet scheduling scheme which does dynamically adjust the priorities of the packets depending on the prevailing network conditions. The proposed scheduling method determines the scheduling priorities by collectively looking at the delay in real time, jitter, queue occupancy, and the type of traffic and allowing the scheduler to adjust its decisions based on the current state of the network . The computations of the scheduling priorities are compared and contrasted with traditional schemes of scheduling, such as, First-In First-Out (FIFO), Priority Queuing as well as Type of Traffic as being accepted. The simulation performance indicates that proposed adaptive scheduling methodology greatly minimizes end-to-end delay and jitter of real-time services without compromising fairness to the best-effort traffic. The results affirm that context-based scheduling is an efficient and realistic tool of improving Quality of Service (QoS) in the current computer networks.