Exploring the use of a generic spatial access method for caching and efficient retrieval of vario-scale data in a client-server architecture

Master thesis (2016)

Authors

Contributors

M. Meijers (mentor)
P. Van Oosterom (mentor)
Programme
Geomatics for the Built Environment (Architecture and the Built Environment) (TU Delft)
Copyright: © 2016 Rovers, A.
Vario-scale data Vario-scale maps Spatial access methods Client-server architecture Caching
More Info
expand_more

Published Date

11-11-2016

Reuse Rights

Other than for strictly personal use, it is not permitted to download, forward or distribute the text or part of it, without the consent of the author(s) and/or copyright holder(s), unless the work is under an open content license such as Creative Commons.

Programme

Geomatics for the Built Environment

Abstract

Vario-scale data structures make it possible to derive maps at arbitrary scale. When requesting a map, data has to be transferred from server to client. This takes time, which affects the responsiveness of the system, and sometimes costs can be involved for every byte that is send over the network. Redundant data transfers should thus be avoided. Having vario-scale data, the opportunity arises to reuse data already present on the client and retrieve only the missing data from the server when requesting a new map. This thesis explores if a generic data driven spatial access method can be used for this purpose, while keeping the service scalable and responsive. The research shows that there is no ready-made solution available that can facilitate in reaching this goal, but that the Hilbert R-tree is suited to be used as a starting point. Like the generic Hilbert R-tree, an index is build using a space filling curve. However, instead of indexing individual objects, an index is made based on a set of objects. The objects are grouped together into packages based on spatial proximity and level of detail. Using the index, a client can request these packages over the network. The newly suggested method makes it possible to utilize client side processing for constructing vario-scale maps and thereby use the client cache to reuse already received responses. The results show that the efficiency of data communication and the responsiveness of the client can improve for an average user scenario. This is shown with the use of a prototype. Furthermore, the method is generic for the way in which data is retrieved by a client from the server. Therefore, the hypothesis is that the method can also facilitate communication for other use cases with higher dimensional data such as 4D point clouds. It is expected that communication can be similar as with vario-scale data, and that only the filter and visualization steps are different.