Track 2: Network Computing
Grids to maintain the competitive edge
Organizations are facing needs for increased computing resources
and cost efficiency. Grid computing and blade servers can help easen the way
INFORMATION needs are increasing at an unimaginable rate every
year. The biggest contributor to this has been the networked nature of information
exchange. Adding to this is the growing dependence on
network services and rich content such as streaming media
applications. The second conference track on Network Computing will examine
how to combat the increasing demand for network computing resources.
Fuelling networked growth
Regulations are another factor driving the growth of network computing. Many
regulations like the Sarbanes-Oxley Act call for an increase in infrastructure
along with strict IT policies. Since conforming
to regulations is mandatory, it has spurred network computing requirements to
The next network computing trend is that CIOs are accountable
for an IT investment the way business heads are responsible for business ventures.
The prime focus now is on the lower acquisition costs and operation.
While it is inevitable in such a scenario that there is a
premium on how fast hardware and software network components can evolve, there
is hope that technologies like grid computing and blade servers will come to
the rescue offering higher performance levels at lower costs.
Grid computing takes off
Grid computing has finally managed to shake off its earlier
mantle of being a concept endorsed only by the scientific community. It has
now evolved to become a practical technology that can provide phenomenal and
cost-effective computing capability.
The main benefit that grid computing has over traditional
server-based computing is that it results in the optimal utilization of computing
resources. In a typical data centre environment, it is not unusual to see certain
servers performing at full load while many others lie underutilised. Grid-based
computing does away with this since there is uniform load distribution. The
optimized load distribution ensures increased performance boosts at a lesser
cost. Grid technology can provide a pool of shared server and storage resources
for enterprise applications.
Grids are also inherently more resilient and flexible than standalone systems.
A properly configured grid has no single point of failure. On the flexibility
side, grids are not limited to single applications but can run many applications
simultaneouslywithout any of the resource constraints that may arise from
running multiple applications on a single server.
Increased computing requirements have placed a heavy premium
on the available silicon real estate. Servers have to provide higher performance
without taking up too much of rack space. Blade servers have an advantage here
due to the higher server densities that they provide at comparable costs.
Blade servers were earlier dismissed as non-feasible due
to their lower server densities and higher costs. With increasing processing
requirements and space constraints, server manufacturers are concentrating on
making their high volume systems available in lower cost blade forms. With more
organisations going for Linux on Intel platform (Lintel) due to its better performance-to-cost
ratio, blade servers will become widely adopted in the enterprise space.