Method and apparatus for resource management in grid computing systems

An apparatus for resource management in grid computing systems, comprising:

• a service level management unit having a service information storage for defining user's demands on execution of a task as SLA (Service Level Agreements) information and storing the SLA information in the service information storage;
• a resource state management unit having a resource information storage for monitoring states of resources in a grid and storing the states as resource state information in the resource information storage;
• and
• a resource-aware policy management unit including:
• a task processing policy decision unit for determining, based on the SLA information and the resource state information, a task processing policy for executing the task;
• and
• a task management unit for executing the task according to the task processing policy,
• wherein the SLA information includes a deadline of the task, a user's available budget for the task and a type of the task,
• wherein the task processing policy decision unit includes:
• an expected completion time calculation unit for calculating for each resource in the grid, based on the resource state information, an expected completion time of the task;
• an expected profit calculation unit for calculating for each resource in the grid, based on the resource state information, an expected profit to be obtained by completing the task;
• an available resource cluster creation unit for creating an available resource cluster by using the expected execution time and the expected profit;
• and
• a task processing policy creation unit for determining, if the SLA information is satisfied by the available resource cluster, a task processing policy for executing the task by using at least one resource in the available resource cluster,
• wherein the available resource cluster is a set of resources having the expected completion time within the deadline and the expected profit being positive,
• wherein the task processing policy decision unit further includes:
• a task division unit for dividing, if the task is divisible, the task into partial tasks based on the resource state information;
• and
• a task processing policy adjustment unit for adjusting the SLA information via negotiation with the user, if the SLA information is satisfied neither by the available resource cluster nor by the task division by the tack unit, and
• wherein if the SLA information is not satisfied by the available resource cluster but satisfied via the task division by the task division unit, the task processing policy creation unit determines the task processing policy for executing the task on a partial task basis by using one or more resources in the available resource cluster.