Cloud computing can provide a great capacity for massive computing, storage as well as processing. The capacity comes from the cloud computing system itself, which can be likened to a virtualized resource pool that supports virtualization applications as well as load migration. Based on the existing technologies, the paper proposes a resource virtualization model （RVM） utilizing a hybrid-graph structure. The hybrid-graph structure can formally represent the critical entities such as private clouds, nodes within the private clouds, and resource including its type and quantity. It also provides a clear description of the logical relationship and the dynamic expansion among them as well. Moreover, based on the RVM, a resource converging algorithm and a maintaining algorithm of the resource pool which can timely reflect the dynamic variation of the private cloud and resource are presented. The algorithms collect resources and put them into the private cloud resource pools and global resource pools, and enable a real-time maintenance for the dynamic variation of resource to ensure the continuity and reliability. Both of the algorithms use a queue structure to accomplish functions of resource converging. Finally, a simulation platform of cloud computing is designed to test the algorithms proposed in the paper. The results show the correctness and the reliability of the algorithms.
机会移动传感网中数据收集策略既要保证传输成功率、减小网络开销,也要尽量降低传感器的能量消耗,从而延长网络生命期。遵循简单实用的原则,提出了基于方向感知的数据收集策略(Data Gathering based on Perceptive Direction,DGPD)。当两个传感器相遇时,以距离它们最近的Sink节点为参照点,分别计算各自的感知方向。把感知方向作为一个重要参数来确定两个相遇传感器的消息转发路由,把消息转发给更有利于接近Sink节点的传感器,从而提高数据收集成功率,减少过多的消息转发。模拟实验结果表明,这种策略可以有效地完成数据收集,并获得较高的网络性能。
