Improved QoS Using Novel Fault Tolerant Shortest Path Algorithm in Virtual Software Defined Network (VSDN)
Amandeep Kaur Sandhu, Ranbir Singh Batth, Amandeep Nagpal · 2019 International Conference on Automation, Computational and Technology Management (ICACTM) · 2019
In this Era of technology, the Internet has become a primary means of communication and information exchange. According to the world statics (2018) 55.1% of world population use internet. The major drawback in today's Internet is that it there is no guarantee of Quality of Service (QoS) to users. To fulfill the required services Internet Service Providers (ISP) establish a Service Level Agreement (SLA) with the users. This SLA includes different parameters that provide QoS. These QoS parameters are bandwidth usage, reliability, energy consumption, failure detection, elastic resource utilization, cost and many more. Either to maximize the profit or due to Capital Expenditure (CAPEX) and Operational Expenditure (OPEX), ISP tries to establish maximum SLA's which result in degradation of requests services by users. Connection failures occur frequently, which is due to node failure in the network. When a node fails re communication from the source is done, which leads to wastage of energy and time. One of the solutions is mixing of SDN and VN to form VSDN. This paper presents a fault tolerant shortest path algorithm which takes into account matrices like energy consumption, bandwidth usage, saturated links and successful allocation of nodes and examines node risk failures in the allocation process. When a failure in network occurs, an alternate path is used which is unprofitable for ISPs as it leads to less resource wastage and increase in traffic tolerance. The proposed Fault Tolerant Shortest Path (FTSP) algorithm examines the node risk failure and find the shortest path to having higher bandwidth availability and focus on energy saving. Results shows 35.8 % increase in performance of the proposed algorithm w.r.t Bandwidth energy efficient algorithm with comparatively lesser overhead during node failure.