A Robust Algorithm for Multi-tenant Server Consolidation

We study server consolidation problems in the cloud under conditions of simultaneous failure of multiple servers, where consolidation means that cloud providers put tenants on shared servers to improve resource utilization and thus reduce operation and maintenance costs. With replicas of each tenant on multiple servers, our objective is to minimize the total number of open servers and ensure that a particular failure will not result in overload of any remaining server. In this paper, we propose the Adaptive Tenant Placement Considering Fault Tolerance algorithm. Unlike existing consolidation algorithms, our algorithm can tolerate multiple server failures while ensuring that no server becomes overloaded. Furthermore, it does not classify replicas into different types. Through experimental evaluations, we show that the proposed algorithm performs better than existing solutions and achieves near-optimal replication allocation.