Resource management techniques for handling requests with service level agreements

The prominence of cloud computing that provides resources on demand to various types of users including enterprises as well as engineering and scientific institutions is growing rapidly. An effective resource management middleware is necessary to harness the power of the underlying distributed hardware in a cloud. The resource manager needs to be able to effectively perform mapping (matchmaking and scheduling) of user requests (jobs) on to resources to satisfy desired system objectives as well as user's requirements for a quality of service that is often captured in a service level agreement (SLA). This paper concerns the problem of meeting an end-to-end SLA (characterized by an earliest start time, an execution time, and a deadline) for applications that require service from multiple resources (referred to as multi-stage applications) on a system subjected to an open stream of request arrivals. A new budget-based algorithm and a resource manager called MapReduce Budget-based Resource Manager (MRBB-RM) are devised for effectively performing matchmaking and scheduling of an open stream of MapReduce jobs (a popular multi-stage application) with SLAs on a distributed environment such as a cloud or a cluster. A detailed description of the algorithm and its performance analysis are presented.