Immune Network Optimization of Composite SaaS for Cloud Computing

Abstract

Serving the needs efficiently for a wide gamut of cloud users is a challenge. One way to address this challenge is to decompose SaaS (Software as a Service) into application components and then consider them as loosely coupled processes that achieve higher functionality. Optimization occurs in efficiently pairing virtual machines to application components in order to lower operating costs for cloud service providers and to lower subscription costs for customers. This thesis explores utilizing an immune network algorithm that mimics antibody activation and antigen and antibody suppression for resource optimization. Experiments are conducted with a series of SaaS configurations, application components placed with virtual machines. Results generated by the proposed algorithm are compared with a previously proposed grouping genetic algorithm. This data reveals that the immune network algorithm outperforms the grouping genetic algorithm in time taken to calculate a resource distribution strategy.