Skip to content

The BBQ – Disaster Recovery

Business Continuity/Disaster Recovery.  The four words every finance head hates.  You can’t make money with DR.  Let’s use the capital we would tie up there for things that can make us money.  As usual, you talk about risk, they go glazed over.  The thing is – BC/DR is big business.  Real big business.  Everybody – and I mean everyone – in IT is or should be about BC/DR. Half of IT is BC/DR – staying in business when the building burns down, and you’re out front with your sticks and hot dogs (in SMB the CFO is crying because he knows he’s done – you’re the only one with backups showing he shot down your plan to buy a couple thousand dollars in backup capacities at an offsite location that could have saved the business so he could fly to some useless conference in Hawaii).

Now I just said half of IT is BC/DR.  That may be a old way of looking at it from a cost perspective. IaaS/SaaS/ITaaS’s core value to me – and I believe lots of IT personnel – is BC/DR.  For a shop with less than 50 servers, it’s cheap to ship data into the cloud, let it set there, and when the building burns down restore it (somewhat) quickly.  It moves cost from CAPEX into OPEX.  It’s easy to go to management and say “This costs you x per month”.  X is usually a small number that doesn’t require significant amounts of financial analysis to show them the value.  Over 48 months or however long you’d use that EMC or NetApp box at some colocation center, you may be paying maybe more.  But it’s a small number, and you’re managers agree.  You set it up.  You move your data there.  It’s slow, but only you notice – but you can recover from the disgruntled guy in the basement burning the building down.   For most businesses post-1995, loss of computing capacity means more than loss of revenue – it is loss of confidence in our services, and that is often more expensive than 2 or 3 hours of lost sales in any business.

Nearly any backup system is cheap compared to the end of the world scenarios that many SMB organizations face – mostly because the someone with the budget is revenue oriented, not risk or cost prevention oriented.  That revenue disappears quick when you’re down for a week.  You’re goal in IT is to not prevent the BBQ that will result when the place catches fire – that’s physical plant/landlord/someone else’s job usually.  You care only about recovering when it does happen.  The hardest part is explaining value to management – sprinklers save lots of buildings but not all.  Floods happen.  Weird things happen.  Complex RAID groups fail.  You have to have backup, and you have to test it.  It’s not a maybe, it’s a yes.  Go out and find a solution:  enough cheap/free solutions exist today – thanks to the cloud.  You don’t have to have EMC RecoverPoint or Symantec NetBackup.  But you have to have something.


Skewers and kabobs

For those who don’t know, i’m the IT guy at a small firm in Austin, TX.  For those who don’t know where Austin is, it thinks it is at the center of the Universe, and it’s probably pretty close.  Closer than the Valley, which is only the center of it’s own universe.

As of late, I’ve been running lots of vendor evaluation in storage.  Now I love storage: my previous employer may have been a storage firm.  The reason I love storage?  If something goes really wrong, we’re all up the creek without a paddle.  The skewers will come out, and IT (me) will be the kabobs.  It’s expensive, it’s tedious to manage, and it’s never fun to consider what happens when one, if not the single most expensive piece of your pie (thanks to virtualization turning off lots of servers) goes bad wrong.  I don’t want to consider that meeting with management, nor the SAN/NAS hating software developer who’ll just laugh and blame me as the root cause of all of his problems.  So I have to put lots of love around vendor selection.  The problem is:  storage companies hate vendor selection.  The emperor is left with no clothes, and they all have problems that are unique to their situation:  EMC has theirs.  NetApp has theirs.  Everyone does.  With EMC I’m losing disks to the OS.  With NetApp I’m losing disks for either extra parity or extra hot spares.  No one says “you buy X storage array with Y number of Z drives and you’ll get X space”.  Not even Dell’s EqualLogic, which they brand as idiot proof (or non-storage administrator friendly) tells you the capacity.  Marketing is a PITA by design.

Tonight, when I go to bed, I won’t lose any sleep over my Dell servers, or my Cisco switches.  No, I’ll lose sleep over what storage vendor is providing the least amount of bullshit, the most value, and the least risk.  Because if one provides too much crap, too little value, or too much risk, I’m skewered.