Job Scheduling for Campus-scale Global Computing with Machine Availability Constraints

Abstract

The rapid improvement of commodity components makes the global computing platform increasingly viable for many applications. This paper studied the problem of job scheduling for campus-scale global computing on theoretical level. Based on system modeling, we first proved the scheduling problem is a NP-hard problem, and then chose the approximation algorithm LS (list scheduling) as a solution. Finally, the theoretical analysis showed that it is a (m+1)-approximation algorithm for our model