Optimization & performance analysis of grid protocol for scalable location services

Grid presents a distributed location service, which allows the construction of ad hoc mobile networks that scale to a large number of nodes as compared to previous work. Grid protocol tracks the location of mobile nodes with its scalable location service architecture and forwards the packets through geographic forwarding. Each mobile node is assumed to be GPS capable and periodically updates a small set of other nodes (its location servers) with its current position. We experimented and analyzed the grid behavior in detail building on original code from grid developers. Our experiment uses NS-2 for up to 600 mobile nodes and 2900/spl times/2900 m as maximum simulation area. Our results show that the storage and bandwidth requirements of grid protocol grow slowly with the size of the network. Grid tolerates node failures well and the query performance of Grid is almost insensitive to node speeds and number of nodes. We analyzed grid behavior over variants of "random mobility models", variations in transmission ranges and finally different node distributions. We believe that our experiment help analyze grid behavior in greater depth.