The data centre of the future - part 1 (where have we been)
by
Charlie Bess
EDS is coming up on its 20 year anniversary of the opening of its data centre in Plano, TX, where EDS' corporate headquarters is currently located.
I was asked to pull together a brief presentation as part of the ceremony and thought I'd share some of my findings on this blog as I go through this analysis. This is likely to turn into a series of entries, since I am just starting.
First I thought I'd look at what has happened in those last 20 years. In 1985 the Intel microprocessor speed was 8 MHz. Today, I have a 3.4 GHz machine at my house (about a 400x increase). Most corporate LANs were running Ethernet at 10 Mb/s. Today 10 Gb/s is not uncommon (a 1000x increase). PC Data storage has grown from 20 MB drives to .500 GB (a 25,000x increase).
In 1985, proprietary operating systems running on proprietary computing platforms dominated IT data centers. These rare resources ran at relatively high levels of utilization.
Along the way to today's data centers, many IT organizations have created large configurations of hardware running at low utilization levels with an ever increasing amount of support complexity just to keep the hounds at bay (e.g., viruses, intrusions, hardware/software values). Viruses and similar security concerns were virtually unknown in 1985.
We are currently in the midst of a shift from assembling standard systems into a custom data center out of standard components based on standard designs to assembling a "standard" data center out of virtual components based on standard designs using modeling techniques like the Data Center Markup Language (DCML).
It's been a tremendous 20 years, and I bet it will be looked on as nostalgically naive 10 years from now, let alone 20.
