Resource allocation and distributed uplink offloading mechanism in fog environment

Abstract

Applying fog computing technology to the shared pattern has two problems to cope with. One is to formulate a rational mechanism for resource allocation, and the other is to design computation offloading strategy of tasks based on resource allocation result. For solving these problems, we construct a three-layer F-RAN architecture first, which consists of terminal layer, access layer and network layer. Second, we adopt differential game and bipartite graph multiple matching algorithm to solve bandwidth resource allocation problem of fog node (FN)-access point (AP) and AP-shared terminal (ST), respectively. Third, Lyapunov theory and proposed deviation update decision algorithm (DUDA) are used to solve computation offloading decision-making and offloading update order-making. At last, simulation results show that our strategy can save 30%–60% system consumption, and the resource demand satisfaction rate can be guaranteed to reach 80% or more.