There is work to be done! There's a war to be won!

Monday, 1 November 2010

NetBackup OpenStorage (OST) - Mind the gap.

OpenStorage has been around for awhile and for the most part it works, and works well. The idea is based on interoperability, working together - finding a way to make sure that each vendor delivers on their promise, which at the end of the day gets wrapped-up in a license you pay for. For those not familiar with it, the idea is simply: Your disks or arrays chat to NetBackup media servers via vendor plug-ins that are installed on the media servers themselves. Plug-ins are written by the vendors for their respective appliances and NetBackup gets to understand what the capabilities are - NetBackup then stores images of data on these storage servers (appliances) but doesn't need to know or care more other than, 'My data is there, give it to me when I ask for it ... delete it when I tell you .. thanks, have a nice day'

So everyone in the mix assumes responsibility for their part in delivering/dealing with data. Cool, but how do I deal with these vendors at the other end of the line who've been trying to solve a problem for the last 6 months since the most recent upgrade broke everything, and are nowhere near taking ownership; in-fact, not only are they not taking responsibility or trying to save the planet, they're arguing about where the problem lies and passing the buck. Meanwhile the patient is being plied with sedatives (patches).

How much testing can you ever really do though? You develop something for a year or 18 months, and you can't really test it for 5 years; the market is moving way to quickly for all of that, and by the time you get it onto the operating table, his patient is moaning for something better or faster, more reliable, etc. Cant blame him. So you bang it out there and you generally wait for failures to occur and plug them? The business doesn't help their cause either. Its always something to do with budget, lack of space, not enough justification in trying to emulate a production environment in a testing lab to actually test these things fully before hooking the patient up to the latest technological thrills (inevitably resulting in spills) I think I'd really love this kind of work  - but it needs to be done properly and we're nowhere near slowing anything down enough or caring enough to actually go the distance. We cant eliminate every problem scenario, but we can truly do a whole lot more than we're doing at the moment by letting tech do some of the driving, and getting the business to ride shot-gun for awhile, taking in the breeze, chilling out. We want it installed now, we want it working it now, but we haven't really tested it...Oh dear, I'm in Wonderland, so call me Alice.

So the teenagers (vendors) argue about who's to blame for about 6 months until *the* upgrade, the patient (client) finds himself in a coma getting overly used to a addictive sedative (workaround) which the support guys don't really understand because it doesn't really work, and which at the next implant (version upgrade) will destroy his liver even more. The parents (business itself) cant control the teenagers who they support - wild nights, huge parties, I've spent my pocket money already, dad...not enough responsibility.

Meanwhile I'm riding the London Underground, my head in someone's armpit, waiting for the next signal failure, minding the gap.

No comments:

Post a Comment