Remember trays of Hollerith punchcards & clunk clunk clunk clunk card readers? Fortran?
Remember hearing (and saying) “…but I only changed one blessed card!?!”
Remember the lesson that came from that exercise?
You have to test the whole thing together, not just the pieces by themselves... that if the whole thing doesn’t work together, it doesn’t matter that the pieces individually are works of art.
Fast forward 40 years – to BUTT Sets, T-BERDs and unit testing. Then add RFC compliance, protocol testers, sniffers, network assessment, Nessus, nmap, nikto, PerfMon, et al – it is all very important and all very, very necessary. But not sufficient.
Why?
Because it’s all about whether or not your contact center solutions can be used the way you want them to be used under the traffic conditions you expect. Not just whether or not the servers respond to a ping or a traceroute, and your ports are provably secure from attack.
And the Users interact at the top of the stack, not somewhere down in the layers. The layers all have to happily interact with each other but in the end it’s about what happens when someone tries to use it, not just whether or not Tivoli or Performance Center show no red or amber alarms.
You need more than your technology’s inside-out view of the world to make sure your customers are being served. In this situation, the top-down view is the one that matters.
Test and monitor at the top of the stack and you’ll know how your customers are really being treated by your technology.
Mike Burke
