Task Ranking Based Allocation of Scientific Workflows in Multiple Clouds with Deadline Constraint

Abstract

The advent of Cloud computing as a new model of service provisioning in distributed systems, progress researchers to investigate
its benefits and drawbacks in executing scientific applications such as workflows. One of the effective problem in Clouds is workflow
scheduling, the problem of satisfied the QoS of the user like deadline as well as minimizing the cost of workflow execution. Existing
work QoS based workflow scheduling algorithm based on a novel concept called Partial Critical Paths, which tries to minimize the cost of
workflow execution while meeting a user defined deadline. Today cloud provider’s mainly concentrate about the increasing their
revenue. This will lead to the selfish behavior which may cause the QoS violation of cloud users. In the existing work, workflow scheduling
is done in only single cloud where there may be the situation occurs in which enough resources are present to satisfy the user demand. And
also the priority of tasks is not considered in the scheduling of tasks. In that case the existing work will still continue to process the user
demands in order to increase their revenue. To address this problem, a novel replication aware dynamic workflow scheduling is introduced
with the consideration of ranking of tasks for multi cloud. The main objective of this algorithm is to dynamically allocate the
workflow across multiple cloud domains with the consideration of reduction of cost for processing those workflows as well as satisfying the
QoS requirement of user. This is achieved by ranking the tasks based on their load level and its successor tasks load level. The
experimental results prove that the proposed methodology can provide the better result than the existing methodology.