If you are using the internet from
your desktop, then your computer is part of a network. This network in turn is
part of another network. In short, the internet is a networking infrastructure.
It was the result of some visionary thinking by the people in the early 1960s
that saw great potential value in allowing computers to share information on
research and development in scientific and military fields. Shrinking computer
sizes and information sharing services made it possible for every computer in
the world to interconnect. The 'World Wide Web' is a service built on top of
the internet which enables millions of computers around the world to share
information.
The internet has served well so far
but has a few shortcomings. Scientists and computing experts are always on the
lookout for the next leap forward. The internet allows you to only share
information and nothing more you can think of apart from that. In order to
address this issue, the experts have devised a new system which is called the
'Grid'. A grid computing system is a type of parallel computing system which
enables us to share computing power, disk storage, databases and software
applications. The term 'Grid' first appeared in Ian Foster's and Carl
Kesselman's seminal work - The Grid: Blueprint for a new computing
infrastructure. Ian Foster is the Associate Division Director in the
Mathematics and Computer Science Division at Argonne National Laboratory
(United States), where he leads the Distributed Systems Laboratory, and he is a
Professor in the Department of Computer Science at the University of Chicago.
In one of his articles, Ian Foster lists these primary attributes of a Grid:
1. Computing resources are not
administered centrally
2. Open standards are used
3. Significant quality of service is achieved
2. Open standards are used
3. Significant quality of service is achieved
A grid computing system, also known
as a distributed computing system, relies on computers connected to a network
by a conventional network interface like the 'Ethernet' (the port into which you
connect your LAN cable). These computers have the capability to combine and
yield better results than a supercomputer. These computers are independently
controlled, and can perform tasks unrelated to the grid at the operator's
concern. This is what Ian Foster indicates in the first point of his checklist.
The striking feature of a grid
computing system is that it enables sharing of computing power. This is also
known as CPU scavenging in which a computer steals the unused cycles of any
other computer. These cycles are nothing but the idle times of a processor in a
computer. Shared computing is the act of sharing tasks over multiple computers.
In other words, it is as good as getting the same job done from a bunch of
people rather than a big strong man. As a result, a task which took days to
complete will be achieved over a relatively small period of time. IBM is
currently working on developing a global scale supercomputer based on the
concept of shared computing. They have named it as 'Project Kittyhawk' and it
will run the entire internet as an application. Shared computing is one of the
services that a grid will provide.
Desktops, laptops, supercomputers
and clusters combine to form a grid. All of these computers can have different
hardware and operating systems. Grids are also usually loosely connected in a
decentralized network, rather than contained in a single location, as computers
in a cluster often are. Hence, a grid's flexibility and additional features
distinguish it from its competitors. Moreover, a grid is built from open
standard protocols and interfaces like TCP/IP protocol suite which is important
to realize internationalization rather than subjugating it to local limits.
In addition to shared computing
power, a grid allows you to share disk storage, databases and software
applications. Altogether, this system will be very helpful to people belonging
to different categories, ranging from scientists to consumers. Most of the
computers in mid size and large size organizations are idle for a higher
percentage of time. These idle processors can be utilized for other important
tasks by the means of CPU scavenging. Different scientists of the world would
like to visualize their applications in real time rather than wait for ages
till their results are shipped, verified, and then sent back. With the help of
grid, people from different fields of expertise will be able to hook up to
remote computers and share their findings. This won't only speed up the process
but will also produce accurate results.
One of the biggest projects on grid
is being carried out in Switzerland by the European Organization for Nuclear
Research (CERN). Thousands of desktops, laptops, mobile phones, data vaults,
meteorological sensors and telescopes will constitute the biggest grid which
will produce an annual database of roughly 15 million gigabytes. This event
will become possible when two beams of subatomic particles called 'Hadrons',
will collide in the 'Large Hadron Collider' - A gigantic scientific instrument
100 meters underground at Geneva, Switzerland. Thousands of scientists around
the world want to access and analyze this data, so CERN is collaborating with
institutions in 33 different countries to operate a distributed computing and
data storage infrastructure: the LHC Computing Grid (LCG).
To constantly reap the benefits of
such a system, it is important to make sure that the computers performing the
calculations are entirely trustworthy. The designers of the system must thus
introduce measures to prevent malfunctions or malicious participants from
producing false, misleading, or incorrect results, and from using the system as
an attack vector. In a grid, computers drop out either voluntarily or due to
their failure.