Contemporary Computing Technologies for Processing Big Spatiotemporal Data - Space-Time Integration in Geography and GIScience - page 323

18.2 Computing Challenges

The growth of big data has outpaced the increasing of computing capability and the
advancement of data processing technologies and tools in recent years, and poses
the challenges of heterogeneity, performance, visualization, on-demand processing,
privacy, and social collaborations.

18.2.1 Heterogeneity

Due to the diverse application usages and technologies for data collecting, pro-
cessing, storage and access, the heterogeneities include the following aspects:
(1) Diversified content organization and representation for data models, formats,
specifications and encoding methods. For example, the data can be either structured
or unstructured, such as text, sensor data, images, audio, video, clickstreams, and
log files. (2) The application specific technologies and strategies on storage/access
methods including data access and retrieval protocols, authorizations and priorities.
(3) The geographically dispersed storage and management of big data that is due
to the globalization calling for the integrating of data collected, hosted (stored) and
shared by dispersed providers, organizations and governmental agencies all over
the world. The heterogeneity has posed the challenges on data discovery (search,
identification and indexing) and integration management (Li et al. 2011b ;Gui
et al. 2012 ). The major problem is that current data structures/models, algorithms
and systems are usually designed to support limited criteria based on particular
application requirements. Detailed challenges (Agrawel et al. 2012 ) include how to
rapidly store and retrieve a large volume of data on storage? How to build efficient
index and data structure to organize them in file systems? How to automatically
generate the right metadata to describe what data are recorded and how they are
recorded and evaluated?

18.2.2 Performance

Timeliness is a vital factor in many time-critical applications. For example, we need
real-time data analysis in e-business, emergence response and decision-making for
natural and man-made disasters. Hence, in the big data era, the capacity of the rapid
data acquisition, discovering, integration, processing, modeling and visualization
has become an important measurement of system maturity.

18.2.3 Computing Intensity

Big data processing involves executing complex algorithms to extract informa-
tion from massive raw data. These complex algorithmic processes, such as data