I really enjoyed reading this article by Freeform Dynamic’s Jon Collins. I wish, of course, that he would focus more on ‘continuous’ rather than ‘high’ availability, but that aside, it’s worth a view, though you really need to watch the webcast too.
Jon ‘gets it’. I like his approach in asking “what are we trying to solve” and tackling this from a service delivery perspective.
Great stuff. There really are some good snippets on the webcast – not least of all if your disk drive fails, put it in the fridge, go for a walk, and when you get back it will work and you will be able to recover your data – this is a Mastercard moment, i.e. priceless!
Anyway, I digress. Back to the podcast. I was a bit perturbed that Andrew from Microsoft knew little about geo-clustering but he did give some excellent advice when commercialising the new improved Windows 2008 R2. First of all, he said, setting up a cluster, for high availability of course, was not as onerous as before (I presume he was referring to Windows 2003). There was then a classic line about, and I quote, “It’s (still) not necessarily that easy.”
Also, when asked the question “Did Windows 2008 include availability in the box?” there was a rather long drawn-out dialogue (a bit like you get from a politician) that didn’t answer the question; so I will answer it for you – NO.
The conversation then developed into something we all agree on – too many companies implement systems and then they think about availability and bolt it on afterwards. The strong advice from the panel was “….”this is absolutely the wrong approach. We come full circle back to that fault tolerant platform I keep banging on about.
Looks like the latest disaster avoidance strategy is to invest in a Fault Tolerant Platform and a fridge. Maybe we’ll see some unlikely partnerships emerging soon. Stratus and Smeg… I can see it now…