Chapter 5, Section 5.2 Grid computing{0}
The article from Businessweek magazine talks about a still premature but cost saving technology called grid computing. “Grids allow processing jobs to be split up and farmed out over a network to many computers so the work can be done fast on any machine that’s available”. Companies like IBM, HP and Sun Microsystems offer computing power in an on-demand basis. They provide computing power when their customers need it and if not needed other users will use it. This computing power can be less expensive than the million dollar super computers used by some companies today. Sun Microsystems charges $1 dollar per processor per hour. IBM offers its utility like computing grids that includes grids from 11 IBM data centers all over the world. Organizations like the US Tennis Association used IBM computing grids to maintain sports fans informed of every detail in every game during the US Open. Usopen.org was plugged into a powerful computer grid that could handle the surges of demand from sports fans. This dozens of servers at IBM’s data centers behaved like one big super computer that can serve up the US Open website one minute and do banks credit analysis the next. Other companies like DreamWorks are plugged into a 1,000 processor compute farm at HP labs to render the complicated animations of its movies after its artist go home. Another example is Axiom corp. “It analyses customer interactions for banks, retail stores, and credit card issuers”. By using these grids companies get the computing power and resources with out having to invest in hardware. Companies save millions of dollars and hundreds of hours of time by switching from expensive main frame computers to grids. Axiom cut its annual expenses in half by switching from a 2 million dollar mainframe computer to a grid of cheap servers. Computing time is also cut down in work that usually took one month to compute now it can be done in four days. Grid technology is still immature, some security issues need to improve and some hardware and software compatibility issues resolved. The goal is to standardize the grids and be able to link the different hardware and software that is made by different companies.
The article is related to the topic of the week in chapter 5 section 2 on grid computing. It is interesting how companies are looking into more efficient and money saving ways to operate and using resources available over a grid server thousands of miles away.
Work Cited
Hamm, S. (2004, October 13). Getting a grip on grid computing. Businessweek, Retrieved september 28, 2010. From: http://www.businessweek.com/magazine/content/04_42/b3904107