Improved Fault Tolerant Elastic Scheduling Algorithm for Cloud Computing

The paper focus on Fault Tolerance, a long standing problem in cloud computing by extending Primary Backup model to include cloud features such as virtualization and elasticity. Fault tolerance is a challenging work in Cloud Computing as virtual machines are the basic computing instances rather than hosts that enable virtual machines to migrate to other hosts. The on demand provisioning of resources makes the cloud elastic thus resources can be added on demand and when no longer requirement exists, the resources can be released. Prevailing fault tolerant scheduling algorithms which studied PB model approach in cloud computing investigated single Host crashes. The proposed algorithm, improved Fault Tolerant Elastic Scheduling Algorithm identified as IFTESA, is able to handle two host crashes concurrently. IFTESA attempts to boost performance by involving virtual machine migration and overlapping technique. The constraints have been investigated to full fill fault tolerance needs.