A Dynamic Resource Balance Algorithm for Multi-tenant Placement Problem in SaaS

In this paper, we consider the Online Tenant Placement Problem(OTPP) Software-as-a-Service level of cloud computing systems. Online boarded tenants are optimally places on limited nodes (or servers) to maximize the total supported number of tenants without violating their Service-Level Agreements (SLAs). The application combination based dynamic tenant dispatch heuristic (ATH for short) is proposed for OTPP, which is a problem feature related strategy combining with a greedy one. To shorten the tenant dispatch time, a new method is introduced for computing the distance between resource vectors. As compared with the best existing method so far for OTPP, simulated experimental results show that ATH disposes more tenants using less dispatch time within limited resources, especially on large instance.