Distributed system for time-sensitive applications with multiple execution options (MEO)

Abstract

Cloud Computing is enabling the consolidation of millions of applications on shared infrastructures due to its wide application. So many applications share common resources, making it increasingly difficult to meet their quality of service (QoS) needs. Aside from that, the characteristics and workload of different applications change over time, which further complicates the system. As part of this study, two online QoS aware adaptive task allocation schemes are developed and compared to demonstrate that the experimental system can exploit a variety of online QoS aware adaptive task allocation schemes. They are mainly, ‘Opportunistic Load Balancing’ and ‘Shortest Job First’ Virtual Machine (VM) Provisioning Scheme. They are allocation-driven algorithms that send jobs to subsystems that provide lower response times. After that, the algorithm divides the stream of job arrivals into sub-streams. The aim of these schemes is to achieve higher resource utilization and system benefits by carefully balancing resource usage efficiency, input workloads, & request deadlines.