Some Numbers That Blew My Mind

Submitted by Tom Huntington on Wed, 12/01/2010 - 6:03pm

During my last two business trips, I had the opportunity to tour a very large IBM i-centric retail data center and spend some time with the casino industry. During these visits I heard some staggering numbers.

For example, in the IBM i-centric data center, my customer told me that they process $28,000,000 worth of business each hour. That’s an awful a lot of business. And, this business is primarily being processed by our old friend, the AS/400. Not too bad for a “legacy system.”

Another staggering number I encountered was the amount of storage a casino gobbles up.

One casino I visited keeps 14 days of surveillance video online and has over 10,000 cameras in their infrastructure. To put it in perspective, 100 high-definition cameras can produce between 250 and 300 terabytes of data per week. For this casino, that means more than 3 petabytes—3 quadrillion bytes of data to store and that’s just their video data. (For those who have trouble keeping these kinds of numbers in perspective, one million seconds is 11.5 days; one billion seconds is 32 years; one trillion seconds is 32,000 years; and one quadrillion seconds is 32,000,000 years.)

This was not one of the larger casinos, either. Just imagine the amount of storage bigger sites might need. And, once the cameras are in operation, they must remain continuously in operation. By the way, the casinos don’t keep this data just for the fun of it—they’re required to by law.

If you have some impressive Power Systems (AS/400) IBM i numbers, I would love to hear them. I won’t share your name or anything else—I just think it’s important that others know how productive, reliable, and secure the IBM i environment is.

P.S. The tour also reminded me of how much security, power, and air conditioning a larger data center needs. The good news is that no one “lives” in the data center like they did in the old days!

Comments

Submitted by Rohit (not verified) on Mon, 01/03/2011 - 6:59pm.

Reliability! I have a relatively easy life as an iSeries admin compared to my colleagues in Windows, Linux/UNIX administration. We have our jobs scheduled across different platforms using Robot/Schedule Enterprise and one of the things I've noticed is how often we are asked to place a set of jobs on hold as the remote server will be down.. of course, many times when the don't tell us about a bounce we get to know with the "lost connection message". The way I see it, our iSeries is always on the job and telling others what to do.. but sometimes the others are asleep!