Research Of Replica Consistency For Spatial Data In Virtual Computing Environment

Virtual computing environment is built on top of wide-area Internet infrastructure, to complete the large-scale application requirements, aggregate kinds of network resources,offer the users with credible, transparent, integration services. Running geographic information systems (GIS) in virtual computing environment, it can complete large-scale spatial data processing, the spatial information querying for large number of users, and distributed resource sharing and other services.The technology of data replica is very important in huge network data systems, to solve the network delay in remote host caused by frequent access, and reduce resource consumption. Data replica is proposed to create the corresponding data copy in local machine. Using replicas can improve the delay of data access, save network bandwidth, achieve load balance of servers, so that the entire system can run stably and have a good performance. However, the replicas stored in multiple sites, there are many data inconsistencies between them.This research aimed at resolving the spatial data inconsistencies in virtual computing environment, offering reliable data sources for GIS, and the users can access to the latest data, get the high quality geographic information services. According to the feature of spatial data, like large-scale, heterogeneous, distributed, through data classification, then research in-depth in replica consistency update of spatial data. This paper includes:In virtual computing environment, study the theory of spatial data and GIS, understand their characteristics, through data classification in different types of spatial data, we proposed a feature-based replica consistency model for spatial data. According to the update frequency of spatial data, use different replica update policies to manage geometric data and attributes data. In the technology of replica update, using a Gossip-based message propagation algorithm for instant update of attribute data. Finally, through the contrastive experiment, using hybrid strategy to manage spatial data replicas, it may be helpful in reducing the network communication between data sites, increasing the performance of whole systems, and providing users with real-time, rapid and stable geographic information services.