We've recently moved the last servers in a nearby datacenter. It has been the end of a project we've decided and initiated almost 3 years ago. Like many smaller IT companies, we began to set-up a High-Availability Infrastructure in our backyard and I suppose like many IT companies, we've been very lucky that the disaster did not happen.
I'm not talking about a power outage or similar - compared to the claim - fairly trivial problems. Todays infrastructure gets more and more optimized and the problem is that with highly optimized systems you introduce more and more fragility. A trivial problem can cause a disaster easily. Welcome to non-linear systems!
Over the past decades we've engineered beautiful and complex systems with a high customer expectation when it comes to reliability and performance. You might have battery backup but this will only buy you 30 minutes then you are stuck.
I recently got invited to a datacenter opening nearby and I can tell you - having seen this impressive state-of-the-art set-up - we were not even close to high-availability.
I even felt abit ashamed having stressed this term in presentations so frequently but haven't thought through the whole system chain. A single router not connected to the battery backup can cause you significant headache - I can tell you.
Delivering ITIL Operational Support - applying the latest best practice - is now a 3rd Party Service I'm happy to pay for.
