This Week’s Wake-Up Call

When it comes to the web, hosted solutions, and (dare I say it) “the Cloud”, I’m all in.  Unlike some of my friends and colleagues who are suspicious of moving their data offsite, I take the opposite stance.  I use Gmail for my mail, Dropbox for shared files, and even a web based password store.  Using these tools, I can move seamlessly from device to device without having to synchronize anything.  I can even lose a hard drive and not miss a beat.

I’ve staked my career on hosted applications.  BioIT Solutions is a hosted application provider.  When I meet with a prospective client, part of our conversation is about the relative merits of hosting the application at their site or in our offsite data center.  Since more and more businesses are working across the traditional Local Area Network (LAN) boundary, the merits of a hosted solution are compelling.  No servers to buy and configure; no VPNs; no poking holes in Firewalls.  The list of advantages is lengthy.  But there is always that nagging fear of being removed from one’s data.

When Hurricane Sandy arrived last week, it offered up an example of the downside of hosted applications.  After the power was restored, and courier service was back up and running, businesses were ready to restart operations.  The internet was functioning, but Sandy had damaged the Wide Area Network (WAN) and it was misbehaving in subtle ways.

One of our customers was experiencing very slow performance when accessing their site.  They rely on this site for most of their back-office operations, so it was really hampering productivity.  But when I visited the site, it was performing as usual.  We suspected the customer’s LAN, so we dispatched our field engineering team to find and fix the issue.  After performing a variety of tests, it was determined that the problem was not the LAN, but instead it was their ISP; their “on-ramp” to the internet.

It turned out that all our customers using that particular ISP were having similar performance issues.  With this knowledge in hand, we put a contingency solution in place.  We outfitted our customer(s) with a Wireless Access Point so they could gain access to the internet from a different ISP.  Site performance returned to normal and so did their productivity.  By the end of the day, the ISP in question had repaired the damage to their network and the contingency could be retired.

So what steps should hosted providers and hosted customers take to lessen the impact of inevitable turbulence on the internet? First, the providers can provide redundancy, so that a problem at the data center will not take sites down.  And as this episode reminded us, it is prudent to have multiple ways of getting to the internet in case there is an ISP specific issue.  And customers should think about Business Continuity Planning, so that vital functions can be performed off-system.

There is a lot of hype around the Cloud.  But the benefits are real.  And I do believe we’ll increasingly see businesses choosing to host applications offsite.  But the same kind of contingency planning you would perform for an onsite system should be applied to your hosted applications.  And if you use this episode to help plan for future disturbances, perhaps this week’s wake up call; Hurricane Sandy, will turn out to benefit your company in the future.