Working on a refresh of my IBM internal SAN problem determination course from last year, I stumbled over the first couple of slides again. They are a little bit of a "raison d'être" for the course - the answer to the question "Why should I learn that?". And they got me pondering again. How long will this still be relevant? Here's what I think:
Every once in a while someone restarts the discussion if tape will die soon. The same discussion comes and goes for Fibre Channel in total. There are lots of people predicting each year that SAN is a dying concept at all. The cloud is here today. Not some spooky future concept but deployed in many forms and flavors. And there are stacks, too. A whole data center in a rack. Pre-integrated, pre-configured, pre-optimized, pre-fueled with software and "expertise".
So does it still make sense to build up SAN skills?
If all the expertise is already in the product, why spend time to become an expert? If management UIs become childishly simple, why should a company pay certified specialists? So why to learn all the stuff and get certified in a world where storage comes as a commodity? Any why are there still people out there saying everything becomes more and more complex if it appears that everything gets so easy and simple now?
A look back...
Unfortunately in many regions in the world water is not a commodity. It has to be brought from remote sites over long distances. Often there are people whose only task is to fetch the water. And if a drought lasts longer than the water in the reservoirs their special skills to find alternative sources of water are in demand. It's undoubted that such skills must be maintained, transfered and extended to ensure the survive of the family or community. In return we - the people in the industrialized countries - take water for granted. It comes out of the tap. My 1.5-year-old understands that concept. Need water? Open the tap.
But does that mean "No experts needed anymore"?
Certainly not, quite the opposite! The preparation of drinkable water and its distribution as well as the handling of the sewage is a complex process chain today. It involves infrastructure specialists, biologists, chemists, process technicans, civil engineers and many more highly skilled persons working together. And given the challenges of the future it will become even more complex most probably.
The same is true for SAN skills
As long as we don't have a worldwide grid of quantum-entanglemented RAM-locality-based servers, I predict that there will still be something like a SAN. There will still be need for architects, for specialists implementing it and for sure for well-skilled people troubleshooting it if problems arise. Storage - as well as Computing in general - might become a commodity like water and we'll definitely won't need everybody to know how it works. But the ones remaining in that area will need to be the real experts. Skilled, trained, experienced and motivated.
You might say: "That's the case with virtually everything!" and you are right. So why not SAN?