During times of recession, you may have to pinch pennies to make the balance sheets work -- we get it. But for this strategy to succeed, there needs to be, you know, actual balance. My plea from IT to finance gurus is to get your head out of la-la land and come back to earth -- to make decisions based in reality instead of looking solely at a magical number at the bottom of the ledger.
I'm thinking of one CFO in particular who took this thriftiness to the extreme, kicking the proverbial can down the road until it cost more in the long run. This CFO, "Bob," has good -- if misguided -- intentions.
[ Your next great IT hire already works for your organization. Here's how to discover hidden IT gems. | Get your weekly dose of workplace shenanigans by following Off the Record on Twitter and subscribing to the Off the Record newsletter. ]
Servers are unnecessary
Bob entered the company at a time when the network systems administrator, who reported directly to Bob, was putting in a request for new servers to replace the old, less efficient hardware. Bob countered, saying we needed to reduce our setup to use as few servers as possible and to get back to him with a plan on how to do so.
We went to the drawing board and came up with a few proposals. We had about 15 servers at the time, and we came up with the "absolute minimum" scenario in which four new high-end servers would replace all the old ones. We were also ready with an "optimal" scenario that used a mix of new servers and old until we could raise enough money for more.
Bob thought we could go even lower. He told us to make use of two new servers and virtualization to replace all the existing servers, and he gave us enough money to buy two new midlevel servers instead of high-end units. We were aghast, telling him that if it came down to two servers, at the very least we needed high-end models. He insisted we couldn't afford it.
With little choice, we implemented his plans. It took several months to complete the project, which turned out to be somewhat of a disaster.
Users started complaining about slowness on the network. The complaints escalated, and Bob called us in to ask what had happened. The servers were new, so what was the problem?
Yes, we explained, the new servers were faster than the old servers. But the aggregate computing power -- including memory, network I/O, and storage I/O -- was slower than the aggregate of all the older servers, not to mention the overhead of virtualization even with hypervisors. Bob didn't seem convinced.
After several more weeks of user complaints and his own firsthand experience, Bob finally budgeted more money for servers. At least it was easy to migrate the virtual machines.