10 Laws of Information Technology

1. Never lose data
2. Centralize when you can, distribute when you have to
3. Standardize when you can, customize when you have to
4. Simple is better than complex
5. Spend as little as possible to solve problem or achieve a goal
6. 100% utilization of IT assets always yield the lowest cost
7. Always have a “plan B”
8. If it CAN break, it WILL break
9. Volume Changes Everything
10. Do Not Be Married To Your Technology

Always have a “Plan B” and if you are really good and/or have deep pockets, have a “Plan C” and a “Plan D”.  Here at Little, we’ve even turned it into a verb: Looks like we’re going to have to “plan B” it.

Simply put, expect that technology will not work all the time.  When it fails you, you won’t be disappointed or surprised. Rather when it works, you’ll be pleasantly surprised.  An IT infrastructure for a typical firm is nothing more than a mine-field of things that are going to blow up.  It’s just a matter of time.  You get that new patch and the system that had been running great is now tanking.  Or a piece of hardware just fails—we didn’t expect that you’d be running that 24x7 is common answer.  Going into a technology deployment, I like to ask the question and have a plan B.

Find the single points of failure and then eliminate them.

I’d rather have two cheap items than one expensive time, e.g. two refurb’d Dell servers with the free version of vmware may be better than one big, bad server.  If I lose one server, I can quickly fail over to the other one.  Cisco has said to me many times, buy smartnet and you’ll have a spare-in-the-air.  I said forget that, I’d rather have a spare-over-there (on the shelf).  I can find a backup router on Ebay dirt cheap and it provides better real-time redundancy.

Break your infrastructure down into the following categories and analyze how to cost-effectively provide more and better redundancy. And remember, this isn’t a one-time event.  Build this into your annual planning and budgeting.  Over the years, systematically work to eliminate failure points.  In a few years, you’ll realize how few meltdowns and fire drills you have.

People

• People redundancy can get expensive, particularly in a small company.  What can you do if your key engineer gets hit by a bus or wins the lottery?

• Make sure they document key configurations so that another IT employee (or consultant) can pick up and do what they do.

• Many private cloud providers are better able to offer people redundancy.  They have a standard, large enterprise IT infrastructure where they can leverage multiple people over a larger customer base.  They can provide people backup much more cost-effectively than an SMB (small/medium business)

Technology

• Microsoft – Windows server and desktop are complex animals and there are lots of hackers out there trying to exploit them.  By far the most common disruption of service comes from a Microsoft-related event.  It could be a security patch, a virus, etc.  The best defense is to have more than one.  We have spare laptops and all our servers are virtualized.  We put our servers on a different network that are harder to get to from the internet.  We alternate patches on our servers so that they would not all get knocked out at once with a bad patch or update.

·         Power – if you can’t afford a local generator, think about how you might be able to leverage the infrastructure from a private cloud provider.  Data centers used to be out of reach for most SMB’s-now they are not.

·         Servers – hands down—virtualize your servers.  It is now too easy and so cheap to virtualize your servers.  The disaster recovery and business continuity advantages are huge.

·         Storage – SAN/NAS technology has come a long way and you can do to storage what you did to servers—virtualize it, carve it up logically, and spread it over multiple pieces of hardware.

·         Circuits – it would be nice to be able to afford two (or more) MPLS networks from difference carriers coming in over different fiber paths into all your offices, but many times that is cost-prohibitive.  Pick a good, reliable carrier, try to get dual-fiber (SONET) connections into your building.  But if you can’t, build a private cloud.  That way even if you lose connectivity, all you need to do is migrate to an internet connection (home, Starbucks, hotel, another office, etc.) and you can hook into your computing infrastructure.

·         Telephones – have spare handsets and a robust PBX that has multiple components.  Shoretel has a good N+1 strategy and the “big boys” have duplicate components in their chassis.  Many businesses are also moving to managed/hosted/cloud provided telephony.  If your PBX is not on generator, it should be.  Even if your handsets are dead, the PBX can still take calls from your customers and route to your cell phones (backup phone system)