An MMAS-GA for resource allocation in multi-cloud systems

Abstract

Delivering computing resources from multiple cloud providers to satisfy the users' needs, known as multi-cloud, gives the possibility for cloud consumers to get access to a wide variety of performing resources with competing costs. The main challenge of this infrastructure is the providers' systems interoperability. In this paper, cloud broker architecture is proposed, which aims on finding a providers coalition to satisfy a user's request with the minimal execution time and cost. The request is a set of independent tasks where no communication is considered. To solve this problem, a Max-Min Ant system (MMAS) and a Genetic Algorithm (GA) were developed as well as their hybrid, and a comparative study was established.