Live Webcast 15th Annual Charm++ Workshop

The idea of a ``Computational Grid'' suggests that high end computational power can be thought of as a utility, similar to electricity or water. Making this metaphor work requires a sophisticated ``power distribution'' infrastructure. In this paper, we present the Faucets framework that aims at providing (a) user-friendly compute power distribution across the grid, (b) market-driven selection of clusters for each job, resulting in effective utilization of resources across the grid, and (c) improved utilization within individual clusters. To function efficiently, the producers of compute power , the parallel servers , must be able to maximize their profit, or another metric of system utility. Utilization of individual clusters is improved by the notions of adaptive jobs and smarter job schedulers. Server selection is facilitated by quality-of-service (QoS) contracts for parallel jobs. Market efficiencies are then attained by a bidding and evaluation system that makes the clusters compete for every job by submitting bids, thus transforming the computational grid into a free market. Job submission and monitoring is simplified by several tools and databases within the Faucets system. (a) allows clients to submit jobs with QoS requirements, (b) makes the servers compete for every job by submitting bids, and (c) selects the cheapest bid on behalf of the client in scalable manner. A distributed software framework, with associated databases and redirection servers simplifies job submission, monitoring and interaction for the users. We describe the overall architecture of the system. All the essential components of the system have been implemented, which are described in the paper. We also discuss ongoing work and future research issues. scalable resource identification mechanisms, bid-generation strategies In this paper we present the Faucets framework which provides a scalable bidding and evaluation system and that clients select the cheapest bid and make the servers compete for every job making the computational grid a free market. It also provides smart job schedulers which try to maximize the profit metric of each parallel system. These features of the faucets system will be described in the next two sections. So a bidding framework is necessary to make servers compete for the jobs and let the clients find the most appropriate servers which in this case would be the server which has the least bid. In this paper, we present the Faucets framework that aims at providing both compute power distribution across the grid and improved utilization among parallel servers. The performance of the parallel servers is improved by smarter job management and adaptive jobs.