Database Reference
In-Depth Information
(a) 2nd Query Performance
(b) 2nd Query CPU Load
Fig. 6.
2nd Experiment results for the proposed system's LB configuration
7 Conclusions and Future Work
This article has presented a cloud-based geospatial LD Management System
which is scalable, cost-effective and sustains good levels of (geospatial) query
performance. Apart from its performance and cost-effective capabilities, the sys-
tem offers a particular Service called Linked Data Management Service (LMS)
which exposes added-value functionality in terms of different ways of publishing,
exporting and querying geospatial LD. LMS relieves the application developer
from the peculiarities of the underlying triple store and as it is REST-based
enables the developer to use any programming language of his/her choice. More-
over, LMS caters for different LD publishing scenarios which vary in terms of
provider data format and publishing control and which require from the provider
the provisioning of the least possible amount of information. LMS also supports
standards, such as GeoSPARQL, SPARQL and INSPIRE. In fact, the support
of INSPIRE can be considered as a fundamental feature of LMS as it enables the
potential data providers exploiting it to follow the strict forthcoming INSPIRE
directive which dictates an INSPIRE-complianet form of data published and
exported by them. LMS finally supports simple and more advanced forms of
geospatial queries which require the use of sophisticated geospatial operators
and functions.
The evaluation conducted on the proposed system reveals its benefits in terms
of increased performance and scalability as well as justified the content of the
particular scalability policies followed. What remains to be performed is a more
thorough evaluation which includes highly demanding scenarios involving both
LD updating and querying. In addition, we will investigate additional scaling
factors which could be exploited towards performing the scaling activities of our
system. The VM's main memory seems a very good candidate as Virtuoso tends
Search WWH ::
Custom Search