An Intelligent Recommender System for Cloud Usage Data Using Predictive Analysis

Abstract

Cloud computing allows tenants to rent resources in a pay as-you-go fashion. It offers the potential for a
more cost effective solution than in-house computing by obviating the need for tenants to maintain complex computing
infrastructures themselves. To achieve this benefit, the right amounts of computing resources need to be given to the
applications running in the cloud. The amount of resources needed is rarely static, varying as a result of changes in
overall workload, the workload mix, and internal application phases and changes. To avoid problems, the amount of
resources allocated to applications should be adjusted dynamically, which brings two main challenges: (1) deciding how
much resource to allocate is non-trivial since application resource needs often change with time and characterizing
runtime application behaviour is difficult; (2) application resource needs must be predicted in advance so that the
management system can adjust resource allocations ahead of the needs. Furthermore, resource-management systems
should not require prior knowledge about applications, historical data such as application behaviour profiles, and
running the resource management system itself (including its prediction algorithms) should not be costly