A User Model-Based Resource Scheduling Framework

In traditional resource scheduling algorithms, hardware condition and network performance, such as CPU, storage space, network bandwidth and delay, play important roles. In previous studies, researchers have ignored imbalance of resource requirement of different user who has special interest affecting the quality of service. In this paper, we present a User Model-based Resource Scheduling (UMRS) framework, which allocates storage resources according to both user information and hardware condition. By predicting the interest of user and pushing possible desired resource to user node or nodes nearby actively, the overall schedule performance can be improved.