MDLB: A Matching based Dynamic Load Balancing Algorithm for Road Side Units

The growing interest in Vehicular Ad hoc NETworks (VANETs) has introduced a large number of vehicular applications. Many such applications generate data that needs to be processed. Virtual machines (VMs) created in Road Side Units (RSUs) with some storage and computing resources can be used to serve the application requests for data processing at the edge. However, the resources available at RSUs are limited and proper load balancing is needed to avoid overloaded RSUs and satisfy application requirements. Migrating VMs between RSUs can be a useful technique for achieving this. However, VM migrations can be costly. In this work, we propose a matching-based dynamic load balancing algorithm which aims to serve maximum number of requests from vehicles while incurring minimum migration cost. Detailed simulation results obtained using real traffic traces show that the proposed technique improves the number of applications served while reducing the migration cost.