Exabytes? Now, That's A Lot Of Data!

An estimated 1-2 exabytes of unique, new information is produced (and stored) each year worldwide, according to researchers at the University of California (Berkeley), counting all media including books, magazines, documents, the Internet, PCs, photographs, x-rays, TV, radio, music, CDs, DVDs, etc. That translates to roughly 250 megabytes for every person on the planet, and most of it's digital (93%, to be precise).

Because its so huge, exabytes are the units needed to measure the amount of global information generated every year. [For the neophyte, one exabyte equals a million terabytes or a billion gigabytes. See the box below for more fun with the data powers of ten and their media equivalents.]

Although, printed materials (240 terabytes) make up less than a fraction of 1% of the total storage of new information, there are about 7.5 billion office documents printed each year, as well as almost one million books, 40,000 journals, and 80,000 magazines.

Magnetic storage costs are falling rapidly, which may make it possible for the average person to access virtually all recorded information, say report authors Hal Varian and Peter Lyman of UC Berkeley's School of Information Management and Systems.  Today, a gigabyte of storage costs less than $10, and that may drop to $1 by 2005.  EMC Corp., the world's largest data storage systems company, sponsored the study.

For further details on "How Much Information", including the complete report or an executive summary, go to: http://www.sims.berkeley.edu/how-much-info/index.html

Although this study was primarily concerned with stored content, its authors did look at other means of information flow, such as telephone calls, e-mail, and even snail mail.  About 11,285 terabytes (if stored) of e-mail zip by each year globally; but it is dwarfed by phone calls and delivered mail. Just in the U.S. (the only region with"good" data), telephone calls would account for 576,000 terabytes and the U.S. mail 150,000 terabytes of storage requirements, if archived.

Courtesy of Roy Williams Clickery of Cal Tech (California Institute of Technology, Burbank, CA)at  http://www.ccsf.caltech.edu/~roy/dataquan/