We propose a localized address autoconfiguration （LaConf） scheme for wireless ad hoc networks. Address allocation information is maintained on the network border nodes, called addressing agents （AAs）, which are lo...We propose a localized address autoconfiguration （LaConf） scheme for wireless ad hoc networks. Address allocation information is maintained on the network border nodes, called addressing agents （AAs）, which are locally identified by a geographic routing protocol GFG （Greedy-FACE-Greedy）. When a node joins the network, it acquires an address from a neighboring AA （if any exists） by local communication or from the head AA （a geographic extreme AA） by GFG-based multi-hop communication. A Geographic Hash Table （GHT） is adopted for duplicate address detection. Each address is hashed to a unique location in the network field, and the associated assignment information is stored along the face perimeter enclosing that location （in the planar graph）. When a node receives an address assignment, it consults with the perimeter nodes around the hash location of the assigned address about any conflicts. AAs detect network partitions and merger locally according to neighborhood change and trigger AA re-selection and network re-configuration （if necessary）. We propose to apply a Connected Dominating Set （CDS） to improve the performance. We also evaluate LaConf through simulation using different planar graphs.展开更多
基金Supported by the NSERC Collaborative Research & Development Project (CRDPJ 386874-09)
文摘We propose a localized address autoconfiguration （LaConf） scheme for wireless ad hoc networks. Address allocation information is maintained on the network border nodes, called addressing agents （AAs）, which are locally identified by a geographic routing protocol GFG （Greedy-FACE-Greedy）. When a node joins the network, it acquires an address from a neighboring AA （if any exists） by local communication or from the head AA （a geographic extreme AA） by GFG-based multi-hop communication. A Geographic Hash Table （GHT） is adopted for duplicate address detection. Each address is hashed to a unique location in the network field, and the associated assignment information is stored along the face perimeter enclosing that location （in the planar graph）. When a node receives an address assignment, it consults with the perimeter nodes around the hash location of the assigned address about any conflicts. AAs detect network partitions and merger locally according to neighborhood change and trigger AA re-selection and network re-configuration （if necessary）. We propose to apply a Connected Dominating Set （CDS） to improve the performance. We also evaluate LaConf through simulation using different planar graphs.
