Latency Minimization for Energy Internet Communications with SDN Virtualization Infrastructure

Abstract

Software-defined networking (SDN) technology is expected to be utilized to link energy stakeholders in a way that encourages active participation in building the energy internet (EI) ecosystem. However, EI is a very complex system with various production and non-production business applications that have specific and strict functional requirements. Hence, network service chaining is indispensable to be implemented. Currently, SDN can be combined with network function virtualization (NFV) technology, and they become SDN virtualization infrastructure. NFV has a role in providing service function chaining (SFC) on consolidated middleboxes, and SDN has a position as a glue between those functions. In this paper, we present our work on latency minimalization for EI communications with SDN virtualization infrastructure. We address our problem as an NFV middleboxes placement strategy to minimize the end-to-end flow latency subject to the middleboxes processing power capacity and the SDN-switch resources constraint. We investigate the existing middleboxes placement approaches and propose a network partitioning algorithm as our heuristic solution. The result shows that our approach could improve latency minimization significantly. The average latency can reach 20.19% and 7.10% lower than the baseline approach in two network topologies. We believe that the work presented in this paper will aid in realizing flexible and real-time capable EI communication infrastructure.