Distributed databases synchronization in named data delay tolerant networks

Abstract

Delay Tolerant Network (DTN) is a small regional network designed to provide better communications when the end-to-end connection is not always possible. DTN is well known for intermittent connections and long delays. Nodes store data packets in the buffers and forward later when the connection is restored. Recently, Named Data Networking (NDN) has been drawing wide attention as a Future Internet architecture. This architecture shifts the emphasis from host to content and pays little attention to where is the content. Routing in NDN is based on the name of the content. Named Data-Delay Tolerant (ND-DT) network is an integration of DTN and NDN. It takes the advantages of both architectures by applying named data approach in DTN scenarios. In ND-DT network, distributed databases are maintained by a group of fixed or moving nodes. Data inconsistency always exists because of the intermittent connections and long delays. However, data synchronization solutions can minimize this inconsistency, helping to reduce the data access delay. ChronoSync is a well-known NDN state synchronization protocol. Data synchronization in ND-DT networks are challenging because of the intermittent connections and the nodes’ mobility. Moreover, the connection between nodes is not assured, which may make synchronization to fail. In this work, it is assumed that there is at least one path between each pair of database nodes. The aim of this work is to improve the recovery process of ChronoSync in order to enhance its adaptability to ND-DT network scenarios. For this goal, ChronoSync and our improved solution were implemented and tested on an ND-DT network simulator. The results show that our improved ChronoSync is more adaptable to ND-DT networks. The improved ChronoSync consumes less time to finish synchronization tasks in all the scenarios. What’s more, in three database scenarios, IChronoSync decreasing about 83% of the synchronization time while Chronosync decreases 62% when changed from sparse network to dense network. What’s more, improved ChronoSync generates 27% fewer data packets, which can increase the probability of other network nodes getting connected.