Fuzzy Decision Load-balancing Algorithm for Cloud-based Terminal Services

Abstract

In this paper, we propose a desktop-as-a-service system that can be used on a campus to enable full utilization of powerful hardware and software resources. Our proposed system comprises Microsoft’s terminal services and a dispatcher. The dispatcher provides a single portal and coordinates with different terminal services as part of the overall desktop-as-a-service system. When a client issues a request, the dispatcher uses a fuzzy decision algorithm to balance the load between terminal services. The fuzzy decision algorithm comprises three steps, namely, fuzzification, rule evaluation, and defuzzification. Based on these steps, the final crisp values can be calculated. As such, the terminal service with the highest crisp value is the most appropriate for serving the given client request. Our experimental results show that our fuzzy decision algorithm successfully achieves the highest possible performance.