Scheduling with deployment
Grid computing provides computing resources to users with high requirement. The hardware is setup relatively easily. Nevertheless the software stack is still ongoing work. Grid5000 is a platform dedicated to computer science. The goal is to provide a large scale testbed for researchers. This work describes a new challenge for scheduling experiments on such platforms. A new goal is to minimize the number of deployment of software stacks. We propose to use well-known multi-processor list scheduling. We present the performance evaluation of our bicriteria algorithm.