Hybrid cloud load prediction model for LMS applications based on class activity patterns

Abstract

Hybrid cloud is a cloud computing model that combines internal resources of the organization with external resources. One major advantage of hybrid cloud computing is to accommodate short but significant transient loads that enterprise server cannot handle. A challenge of managing load balancing in this environment is usually on the tradeoff between user satisfaction and cost of external resources. In some applications, like learning management system (LMS), it is possible to predict load in advance using existing class activity patterns stored in its own database, and therefore it makes resource provisioning easier. In this paper, we analyze the class activity data of an LMS site and model the self-aware load prediction based on these patterns.