Join Query Processing Over Delay Tolerant Network

Increasing requirement of different applications results in diversification of networks. In somenetworks, mobile nodes are moving constantly with message transmissions depend on chances oftheir meeting. These networks are called Delay Tolerant Network (DTN). Different to othernetworks, the characteristic of DTN are long delay, intermittent connectivity, resources-restricted,etc.Nowadays, DTNs are used in many areas, such as data collection over battlefield network,vehicle network or environmental monitoring network. These applications are usually used by datacollection rather than data query. Actually, query processing, especially join query processing, in aDTN is very useful. For example, in a vehicle network, a user may want to query traffic and roadcondition in a given area by joining data among several nodes. However, the research is not fullyexplored..The article introduces mobile models of nodes, mobile model and relationship betweennetwork scale and node performance. Then we define the node model and network model. Basedthese, this article has the following contributions:(1) The article detailedly analyzes routing policies and proposes a routing policy based on theBloom Filter. We also analyze query processing schemes and study multi-join query processingalgorithms in distributed network, P2P or wireless sensor network. We then propose a whole multi-join query processing scheme for DTN.(2) Focusing on the problem of speed of join in mobile nodes, a co-processing method byhardware and software is presented to accelerate the window join operation over data streams. Inthe CUDA (Compute Unified Device Architecture) architecture, CPU (Central Processing Unit)transfers data to memory of GPU to be processed in parallel. Experiment results show that thismethod can greatly improve the processing speed with about50times faster than softwareimplementation.