【摘要】 在大规模存储中,经常面临组件的增加和失败。为了提高可用性和可管理性,研究和比较了不同的数据定位机制,设计了一个离散的、自适应的算法。这个算法保证任何一个特定的数据对象的副本不会被放置在同一个存储节点上。并且根据存储节点的能力公平地分布数据对象到存储节点上。当新的存储节点加入系统或已存在的节点退出系统时,需重新分布的数据对象被尽可能地减少。算法中没有集中控制点,保证了系统的可扩展性。更多还原

【Abstract】 In a mass storage system, it is common where components are added and failure. For the advancement of the usability and manageability, this paper explores and compares many different schemes of data location, and designs a discrete, adaptive algorithm. It guarantees that replicas of a particular object are not placed on the same node. It can distribute the data objects to the storage nodes evenly and minimizes data movement when a new node is added to or existing node is removed from the system. There isn’t a central point, so the algorithm can guarantee the scalability.更多还原