28-09-2012, 12:35 PM
Two topics that best fit my research interests in cloud computing
cloud research issues.pdf (Size: 79.03 KB / Downloads: 143)
Current Research Activities:
Currently, we are studying some key issues related to fault-tolerance in cloud computing
environment. Although many mature technologies from other domains can be used as
components in cloud computing, but there are still many unresolved and open problems due to its
unique characteristics which are different from distributed computing, cluster computing, grid
computing, utility computing and service computing. One key issue is Fault Detection in cloud
computing. Fault Detection is an essential building block for reliability, performance, loadbalancing
and scheduling where mobile users engage in computing using data-rich environments
and multi-cloud solutions. Fault-tolerant network systems are designed to provide reliable and
continuous services for wired and wireless distributed networks despite the failures of some of
their components. Although there many FD algorithms available in other domains, they cannot
be directly applied to cloud computing. As far as we know, most existing schemes cannot
automatically adjust their parameters for dynamic or mobile conditions (including mobile users)
which are common in cloud computing environments. We are developing FD schemes for cloud
computing considering its unique characteristics and implement them on a cloud testbed.
Performance evaluation will also be done to validate our schemes. Finally, we are developing
practical Fault Detection systems in multi-cloud computing environment besides proposing
theoretical models and schemes for FD in cloud computing, which we hope can also be used in
other domains such as mobile computing and grid computing environments.
Future Important Research Problems I Plan to Pursue:
Think of a future: A user starts out in the morning with coffee at Starbuck's - his rich data stream
drawn from his location, the surrounding stores, services and sensors, and he begins a compute
processing that lasts all day long - wherever he moves. Indeed as he moves he may be changing
the compute resources used. As he travels from Atlanta to Los Angeles the computing is
continuously going on, being passed securely, reliably and efficiently from cloud to cloud. And
the Clouds are data clouds, compute clouds and perhaps even user clouds. Of course, many
mature technologies are used as components in cloud computing, but there are still many
unresolved and open problems due to its unique characteristics which are different from
distributed computing, cluster computing, grid computing, utility computing and service
computing. The virtual tools for doing this may be software, firmware or hardware components.
We seek to explore these building block components. Fault Detection is an essential building
block for reliability, performance, load-balancing and scheduling where mobile users engage in
computing using data-rich environments and multi-cloud solutions. Fault-tolerant network
systems are designed to provide reliable and continuous services for wired and wireless
distributed networks despite the failures of some of their components. As an essential building
block for fault-tolerant network systems, Failure Detector (FD) plays a central role in the
engineering of such dependable network systems. Fault-tolerant control schemes are designed to
provide reliable and continuous services despite the failures of some resources. Effective failure
detection is essential to ensure acceptable QoS for virtual network applications. For example,
PlanetLab is a global cloud network that supports the development of new network services and
currently consists of 1076 nodes at 494 sites. While many nodes at any time are inactive, and we
may not know the exact status (active, slow, or dead), it makes services variable.