Stochastic Network Calculus for Network Function Virtualization in Underwater Wireless Sensor Networks

Abstract

Underwater Wireless Sensor Networks is a popular sub domain of WSN. The Number of Ocean monitoring applications is increasing day-by-day. For effective operations of such underwater applications service provisioning plays a major role. Network Function Virtualization (NFV) has been considered as an effective technology to make flexible service provisioning in terrestrial network. Since UWSN also demands such flexible service provisioning capability, NFV can be adapted for UWSN as well. By providing network operations through NFV in UWSN operational consumptions and capital expenses will be reduced drastically. So, this research focuses on modeling end-end performance bounds on NFV based UWSN using Stochastic Network Calculus (SN C). Monitoring applications in UWSN expects on demand service provisioning capability from the underlying network. In order to model such on demand service provisioning features and capability of NFV, we have taken into account both the non-bursty and bursty type of traffic of UWSN. In UWSN for modeling estimation of current resource availability of Virtual Network Function nodes with multi-level traffic and their complicated NFV chain, using leftover service property and convolution associativity property of SNC has been used. The proposed mathematical model of NFV service provisioning in UWSN has been evaluated for its correctness using a simulation model. The results of the simulation model and the proposed analytical model have a very negligible difference. So, the proposed model can be adapted in real time of effective NFV based service provisioning in UWSN.