Information Technology Reference
In-Depth Information
devices to store and input reusable programmatic code. Paper punch
cards gave way to paper tape, magnetic media, optical storage, and newer
mechanisms with increasing storage density and greater access speeds.
Gordon Moore is credited with the famous Moore's law, which identifies
an effective doubling of computing power every 18 months. Mark Kryder
is credited with an extension to Moore's law explaining the doubling of
storage on roughly the same 18-month cycle.
This chapter focuses on data storage solutions and techniques for deal-
ing with Kryder's law in the modern enterprise network, where terabytes
of data can be found on individual desktop computers and multigigabyte
flash drives are embedded in phones, watches, pocket knives, and even
coffee mugs. Rich content media and expanding databases mandate care-
ful planning so that capacity remains ahead of demand, with a watchful
eye toward security and recoverability.
Everything in Its Place
The earliest personal computers had 4 kilobytes (4k) of RAM and were
loaded using audio tape recorders. A 4k program such as
Haunted House
took only 15 minutes to load—assuming everything worked just right.
Hard drives have existed for some time, storing electronic data in large
disk packs that were once mounted manually into drive systems the size
of small refrigerators. The times before that are dark indeed, when data
was stored on magnetic tape reels that required large, carefully controlled
storage areas and huge reel-to-reel systems for media access and storage.
Looking farther back, paper tape reels and punch cards were the
medium of programmatic data input. At 40 bytes per card, a single
3-minute MP3 file would have taken up 75,000 punch cards; and if a
card jammed halfway through the file input, all cards to that point would
have to be collected and resorted before replacing the damaged punch
card and starting the input process again. I can well remember the excite-
ment of completing an entire program and data set on punch cards, sub-
mitting the run to computer technologists who then went into sealed,
chilled areas to feed my hard work into hoppers to processed data into
early mainframe systems. This was typically followed by the disappoint-
ment of being told that card #4031 had jammed in the reader and needed
to be re-punched (and re-sorted along with cards #0000 through #4030
that were jumbled together in the output hopper).
