Improving VM Placement in fog Center by Multi-objective optimization

Abstract

Fog computing works in conjunction with cloud computing because cloud computing has latency issues for the critical applications. When the client requests reach to the fog server they are allocated to the Virtal machines (VMs) and then the VMs are placed into the physical machines (PMs). During the VM placement the provider tries to allocate all the requests to the VMs and then the VMs has to be placed inside the PMs. Virtual Machine Placement (VMP) problem is a NP-hard problem and the provider always wants to optimize the hosted VMs, minimize the PMs and minimize the resource wastage to optimize power consumption. This paper proposes Multi-objective integer linear programming (MOILP) and solves in two different methods such as lexicographical preference method and weighted sum method and it has been implemented for two categories of fog center such as homogeneous and heterogeneous and it has been found that the weighted sum method outperforms over the lexicographical method and also for heterogeneous fog center.