Information Technology Reference
In-Depth Information
10 times on the same workstation via the accident application, it took variable
times between 83 and 91 seconds, and the average time was 84 seconds. Here,
it is clear that SQL Azure took most of the time, which means the rest of the
processing in the application was fast, taking approximately 7 (84 − 77) seconds.
In both cases, the query returned 6,642 accident records, and the size of
data returned by server was 543 KB. At that time, the application never
crashed. It is important to mention that the Internet connection used was
a 10-Mbps broadband connection with a DSL (digital subscriber line) test
downloading speed of 3.84 Mbps and upload speed of 0.94 Mbps.
7.6 Conclusion and Future Work
This chapter addressed the need for generic design patterns for GIS cloud-
enabled deployment. The proposed architecture utilizes the SQL Azure
geospatial database in conjunction with other Microsoft technologies and
services. This architecture is designed to be open and extensible.
Considering the original nature of the problem, it is noted that GIS deploy-
ment on clouds is per se a complex issue. It requires the seamless integration
of distinct GIS capabilities to search, access, and utilize geospatial data with
the cloud computing capabilities to configure, deploy, and manage computing
infrastructure to permit the computability of intensive models and databases.
The chapter has provided a proposed methodology and architecture to
enable systematic assembly of GISs in the cloud. This approach provides
a viable solution to build stable and complex GISs in the cloud that can
perform under extreme conditions, but there are a few minor issues, such
as SQL server scaling and the need for a more robust and comprehensive
API to convert WKT geospatial data into Bing, Google Maps, and other
base maps objects. It is also interesting that an application built using this
methodology can be implemented on IaaS and PaaS service models because
it is implemented in PaaS and all the resources used in case of PaaS can be
replicated in IaaS.
Further work should be carried out to evolve this paradigm for building
GIS-oriented cost models for cloud computing where resources, computa-
tional and geographical, are correctly represented and priced. Additional
areas to be tackled to further develop our concept are the following:
• Building a more comprehensive WKT geometry to map object
parser APIs for all geometry types for different base maps, such as
Google and Bing Maps.
• Currently, Bing Maps do not provide an object to represent
multipolygons. We propose development of an API that allows
Search WWH ::




Custom Search