I’ve been doing some work at a large financial organisation over the last couple of weeks. As always, the result of my analysis is that the technology isn’t the main problem.
Let’s face it, most enterprise storage technology is pretty similar; arrays all have similar features – RAID, Fibre Channel support, massive scalability and so on. Switches are no different; 90% of the functionality required is connecting hosts to storage.
What matters is how the technology is used and that all comes down to how process and procedures are implemented.
ITIL gives you a framework in which to work; it does categorisation for you, but you’re going to have to implement the process yourself. Here’s a few things that seem to always crop up:
- Decommissioning old Technology. Or what we should call Technology Lifecycle Refresh. Many sites have technology stretching back n-1, n-2, n-3, or n-4 generations old. They have multiple vendor types, often running with different levels of code.
- Keeping Host Firmware/Driver Levels Accurate. Unfortunately this piece of work tends to fall between teams – the platform team won’t do it because its storage software; the storage team can’t do it because they aren’t permitted to make host changes. Not keeping host levels up to date is a potential disaster waiting to happen. It introduces risk into an environment and will eventually prevent upgrades as there is a limit to the level of code that vendors will support.
- Demand Management. By this I don’t mean capacity planning, which is a whole subject in its own right. I mean negotiating with the business to understand their requirements 3, 6, 12 or 18 months out. This also means discussing the specifics of their requirements – what tier, what performance requirements, what replication requirements and so on. By understanding customer’s needs, it becomes easier to identify technology which can be used to reduce cost and increase business advantage.
There are lots more; feel free to throw a few out for discussion.