Geography Reference
In-Depth Information
18.6
Computing Technologies
To address the computing requirements from model simulations, data discovery
by spiking numbers of users, and visualization pipelines, different computing
accelerating technologies can be utilized. This section demonstrates how grid
computing, cloud computing, and GPU accelerating technologies are utilized to
solve the computing challenges.
18.6.1
Cloud Computing for Spike User Access
of Clearinghouse
To tackle the data intensity challenges, a series of techniques and services are
developed to enable the operations of GEOSS clearinghouse (Huang et al. 2010 ).
Currently, the GEOSS clearinghouse is hosted by the Amazon EC2 to handle
potential massive users and data requests. For example, given the elasticity of
the computing platform for the clearinghouse, we can autoscale one, two, and
five instances to respond to different spike number of user access (Yang et al.
2011a ). Figure 18.11 illustrates how the cloud responds to massive concurrent user
requests by spinning off new EC2 cloud instances. The example illustrates a varying
number of requests to the GEOSS clearinghouse. All instances are running from
beginning except the autoscaling case, which has one instance running at beginning
and elastically adds instances when needed by concurrent requests. It is observed
that when more computing instances are utilized, higher gains in performance can
be obtained. The elastic automated provision and release of computing resources
greatly prepared us to respond to concurrent access spikes while at the same time
reducing cost by sharing computing resources with other applications when there
are no concurrent access spikes.
Fig. 18.11 Server response performance comparison by single, two, five, and five autoscaling
instances (Yang et al. 2011a )
Search WWH ::




Custom Search