Fault tolerant scheduling and dynamic data replication framework with cost efficient data storage for cloud computing environment

Abstract

As the number of users of Cloud computing systems increased, the tasks to be scheduled in cloud increased proportionally. Hence, there is a need for better algorithms to schedule tasks on these systems. Scheduling can be static or dynamic. Static scheduling system already has the information about complete structure of tasks and mapping of resources before execution, estimates of task execution/running time.Dynamic scheduling depends on the submitted tasks to cloud environment and the current states of system. A good scheduling strategy should allow shortest access to the required data, (i.e.) to reduce the data access time. Data replication strategy allows to access data in a wisely manner and offers a faster access to the files required by cloud jobs. It also results in reduced bandwidth consumption. newlineIn addition to real-time requirements, fault tolerance is also a significant property of cloud computing. Fault tolerance provides guarantee of timeliness and safety for real-time systems in case of failure. Most of the existing works consider the system as non faulty, but in real-time, various faults may occur at every data center. Hence, there is a need to design a scheduling algorithm which should be fault tolerant and also to integrate this algorithm with dynamic data replication strategies to form one framework to reduce the data access time and bandwidth consumption. newline newline