Grid Computing - Guide to Cloud Computing for Business and Technology Managers - page 246

• Grid computing results in improved efficiency of computing, data,
and storage resources due to parallel CPU capacity, load balancing,
and access to additional resources. As computing and resources can
be balanced on demand, grid computing results also in increased
robustness and reliability—failing resources can be replaced easier
and faster with other resources available in the grid.
• Grid computing furthermore enables a more efficient management
of distributed IT resources of companies. With the help of virtual-
ization, physically distributed and heterogeneous resources can be
better and uniformly managed. This makes possible to centrally set
priorities and assign distributed resources to tasks.
• Grid computing, in combination with utility computing, enables
the transformation of capital expenditure for IT infrastructure into
operational expenditure and provides the opportunity for increased
scalability and flexibility. However, the usage of utility computing
results in comparatively higher security and privacy risks.
• Grid computing enables cost savings in the IT departments of com-
panies due to reduced total cost of ownership (TCO). Instead of
investing in new resources, greater demand can be met by higher
utilization of existing resources or by taking advantage of utility
computing.
12.2.1 Virtualization
Grid computing also differs from virtualization. Resource virtualization is
the abstraction of server, storage, and network resources in order to make
them available dynamically for sharing, both inside and outside an organi-
zation. The universal problem that virtualization is solving in a data center is
that of dedicated resources. While this approach does address performance,
this method lacks fine granularity. Typically, IT managers take an educated
guess as to how many dedicated servers they will need to handle peaks, pur-
chase extra servers, and then later find out that a significant number of these
servers are significantly underutilized. A typical data center has a large
amount of idle infrastructure, bought and set up online to handle peak traf-
fic for different applications; virtualization offers a way of moving resources
from one application to another dynamically. Three representative products
are HP's Utility Data Center, EMC's VMware, and Platform Computing's
Platform LFS. With virtualization, the logical functions of the server, stor-
age, and network elements are separated from their physical functions (e.g.,
processor, memory, I/O, controllers, disks and switches). In other words, all
servers, storage, and network devices can be aggregated into independent
pools of resources. Some elements may even be further subdivided (server
partitions, storage LUNs) to provide an even more granular level of control.
Elements from these pools can then be allocated, provisioned, and managed,