Abstract
The progress of Cloud computing as a new model of service delivery in distributed systems encourages researchers to study the benefits and drawbacks about scheduling scientific applications such as workflows. Cloud computing allows users to scale their workflow applications dynamically according to negotiated service level agreements (SLA). However, the resources available in one cloud data center are limited, e.g. in Amazon 20 VMs is available. So if a high demand for a workflow Application is observed, a cloud provider will not be able to deliver consistent quality of service to process the application and the SLA can be violated. Our approach to avoid such a scenario is to allow the growing of resource requests by scaling Workflows applications on multiple independent data centers. Our approach is achieved by the installation of agents called CloudCoordinators and a special agent called CloudExchange. These agents follow economic market policies to get virtual machines across multiple data centers. In our approach, a user can offer his resources already obtained while waiting the end of an input/output operation in a Workflow. The discovery of offers and requests is done via the specific services offered by the CloudExchange agent. Clouds providers and users access the CloudExchange via there CloudCoordinator.
Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
