2017

Conférence

ACM Symposium on Applied Computing

This paper explores one of the critical issues, SaaS placement in cloud data centers, for reducing execution time of composite SaaS applications. We adopt a multi-swarm variant of Particle Swarm Optimization (PSO) to propose a service placement method. Also, a cooperative learning strategy is hybridized to the placement algorithm, which makes information of best candidate servers be used more effectively to generate better placement plan. In the proposed method, for each sub-swarm of servers, the worst placement learns from the best servers, so that worst servers can have more excellent exemplars to learn and can find the optimal placement for SaaS components more easily. Experiments show that our solution is efficient in comparison with existing SaaS placement approaches.

-