Heuristic virtual machine allocation for multi-tier Ambient Assisted Living applications in a cloud data center

Abstract

Cloud computing provides the essential infrastructure for multi-tier Ambient Assisted Living (AAL) applications that facilitate people's lives. Resource provisioning is a critically important problem for AAL applications in cloud data centers (CDCs). This paper focuses on modeling and analysis of multi-tier AAL applications, and aims to optimize resource provisioning while meeting requests' response time constraint. This paper models a multi-tier AAL application as a hybrid multi-tier queueing model consisting of an M/M/c queueing model and multiple M/M/1 queueing models. Then, virtual machine (VM) allocation is formulated as a constrained optimization problem in a CDC, and is further solved with the proposed heuristic VM allocation algorithm (HVMA). The results demonstrate that the proposed model and algorithm can effectively achieve dynamic resource provisioning while meeting the performance constraint.