Critical Path(s)
You’d think that having two backup copies of critical computer data and program files would be a good plan, right? Well, of course, it is a good plan. And I do, in fact, have two backup tapes for my data and programs (patting myself proudly on the back).
But plans and reality don’t always coincide.
Even though the data on them is up-to-date, both tapes are more than two years old. And — you guessed it — both of them broke when I needed to restore a corrupted program file.
It’s a real kick in the butt when you do everything right and it still messes up.
The point of this personal "learning experience" is that even duplicate tapes left me with only one path for recovery: my tape drive. However, until recently, a tape drive was the best choice for economically backing up large amounts of data.
But times (and technology) have changed. As of now, I have data stored on the Internet on various free storage sites and on CD/RW disks. Why? Because it gives me two entirely different and independent recovery paths.
So the question is, how old is your thinking and how adequate are your systems for backup and recovery in case something goes wrong? And how many recovery paths do you have for your critical processes?
Are you totally dependent on the host or can you tolerate some downtime and work locally? If one node of your WLAN goes down, do you have enough overlapping coverage or another way to get onto the network? How flexible is your LAN? Can you route around problems or are you out of luck until someone fixes whatever’s crashed? Do you have live spares (e.g., scanners, printers) that you can swap out to keep production running? When was the last time you checked your spares to make sure they were still in good working order?
Let’s take a look at another critical path: people.
Is there enough redundancy or overlap in your people’s knowledge so that operations, maintenance or repair won’t suffer when a couple of people are out with the flu at the same time? And, if not, are equipment manuals and procedures readily available (and up-to-date) for those who have to fill in?
Times, and technology, have changed. We now have more tools available and more options than ever before. It may well be time to update both management and business transactions and move toward a distributed, Internet model. Perhaps we should begin thinking more along the lines of a Napster model (without the intellectual property theft). The Napster peer-to-peer model gives us multiple, redundant paths to the same goal, allowing us to tolerate more types of problems and still keep working. Some paths may not be as efficient as others, but slow is better than stopped.
I’m not actually suggesting that anyone try to apply the Napster model to all aspects of management and operations. I’m suggesting you see what new technologies can offer and begin developing more ways to route around local problems or equipment failures.
At the very least, adopting the Napster model means re-evaluating existing procedures and systems to make sure you have a backup plan for your backup plan.
Bert Moore
contributing editor
