Mobility aware load balanced scheduling for mobile grid environment

Abstract

Grid computing is a large-scale virtual computing environment where geographically distributed resources collaboratively provide an exceptional infrastructure for solving complex, scientific and computing tasks. The emerging Grid Community is trying to incorporate mobile devices into the Grid forming mobile Grid. The rationale behind creating the mobile grid is that the number of mobile device users is exploding at a greater pace and the present mobile devices are equipped with high-speed processors and large memory with advanced technology. The mobile devices also have advantages over fixed computing resources such as mobility, portability and pervasiveness. In addition to this, there are lots of highly capable mobile devices, idle most of the time in the same enterprise network. Despite the advantages, the challenging features of the mobile Grid environment are frequent disconnections, battery power, heterogeneity, mobility, volatility, low network bandwidth and security flaws. These inferior characteristics affect the job execution process significantly. In order to overcome the challenges of the mobile Grid, an efficient, load-balanced, fault Tolerant and a secure system is proposed in this thesis. The Load Balanced Scheduling is accomplished by introducing an efficient Mobility Aware Load Balanced algorithm. Accordingly, the scheduler receives the job from the user. It analyses the job and classifies it as a compute-intensive or communication-intensive job. The scheduler calculates response time for all the resources. Based on the response time and the values of the parameters Round-Trip-Time (RTT), CPU speed and CPU capacity, an intellectual criterion is applied for selecting a suitable resource for a compute-intensive job. Similarly, for a communication-intensive job, the response time and the parameters mobility and reliability are considered for selecting a suitable resource. newline