Abstract
Here, we present the design and architecture of an Agent-based Manager for Grid Cloud Systems (AMGCS) using software agents to ensure independency and scalability when the number of resources and jobs increase. AMGCS handles IaaS resources (Infrastructure-as-a-Service - compute, storage and physical resources), and schedules compute-intensive jobs for execution over available resources based on QoS criteria, with optimized task-execution and high resource-utilization, through the capabilities of grid clouds. This prototypal design and implementation has been tested and shown a proven ability to increase the reliability and performance of cloud application by distributing its tasks to more than one cloud system, hence increase the reliability of user jobs and complex tasks submitted from regular machines.