Information Technology Reference
In-Depth Information
region is typically represented by a significant polygon with thousands of
points to be verified.
An increase in data size requires not only more storage but also other
computational resources. For example, more data require more processing.
As the area covered by the system and the number of people using the system
increase, the capacity of the whole infrastructure must increase due to the
inherent computational complexity. By providing elasticity and on-demand
consumption, any feasible approach should allow the system to scale down
when the user count is low and, conversely, should cope well with any usage
peaks. Cloud computing ought to help overcome this scalability problem
efficiently, with a service model that enables on-demand resource access by
aggregating configurable computational resources that can be rapidly provi-
sioned and released.
Although such a model seems to be tailored for data-driven environments
such as a GIS, programming, manipulating, and processing geospatial data
typically requires the inclusion of complex data structures and demand-
ing mathematical transformations. Even though standard web program-
ming techniques have evolved to be applied in cloud environments, only a
few proven pattern-based programming paradigms have been successfully
applied in the Cloud. So, it should be clear that seamless cloud deployment
entails a substantial amount of work, and current GIS tools are typically
associated with a GIS software developer while cloud ones can be locked to
a given cloud provider. Arguably, to have different GISs in the Cloud there
has to be an orchestration of infrastructure and applications that can show
tangible financial and computing benefits.
Different authors have started to evaluate the distinct possibilities in this
area. Some argued for the need for the next generation of cloud infrastruc-
ture to be supported through traditional multitier architectures [2], while
others have pursued the provision of innovative object-oriented data models
and algorithms to retrieve data in a distributed environment [19]. But, the
vast majority of the relevant approaches encourage the creation of generic
GIS web services on top of map image files and geographic information
in general, accessed as a web service through the Google App Engine [3],
MapReduce-BigTable [20], a private cloud [4], or simply the Internet [23].
Having reported initial performance figures on a par with similar
server-based web services, the last two approaches are clearly not associated
with public Cloud deployments, but all four are definitely representative of
the growing trend for the provision of GISs as a service.
From a more general GIS perspective, recent works have advocated the
creation of
spatial cloud computing
—a subarea in which the spatiotemporal
principles of geoassembling cloud-based GIS spatial sciences, and by exten-
sion a well-designed GIS, can be effectively represented in the Cloud given
the continuous nature of GIS constraints [11, 15, 21, 22]. As part of this growing
trend, there have been few comparative analyses of GISs in grids and clouds
[14], and on different public infrastructures [24], using ad hoc GIS deployments
Search WWH ::
Custom Search