Mobile Intelligence

24.3 DATA-REPLICA MECHANISMS

The basic data-centric mechanism aims to develop data-centric storage architecture and reduce the query overheads. However, the data stored in exactly one data-centric node would cause the failure access when the data-centric node is failure. Another consideration is the query efficiency. In the data-replica mechanisms, some backup nodes surrounding the data-centric node store the same data. In this way, the sink node may access one of the backup nodes and thus has more opportunities to improve the query efficiency. To achieve this, the data-replica mechanisms provide some architectures for storing and accessing the backup data, and hence increases the fault-tolerance capability. Most of the existing data-replica mechanisms backup the same data based on a particular topology, making ease of the data accessing. According to their topologies, these mechanisms can be further classified into ring-based, tree-based, and zone-based classes.

24.3.1 Ring-Based Mechanisms

The basic idea of the ring-based approach is to backup the same data on those sensor nodes that are located on the ring. Zhang et al. [3] proposed a ring-based mechanism to implement the data-replica concept. The application scenario is that the sensor network is used for detecting targets within a sensing field. Each sensor detects nearby target and periodically generates reports. When the operator intends to know the status of target, he sends a query packet to the network for the purpose ...