Oh, dear. This seems… ill-timed.
Some technical experts are perplexed at the U.S. government’s plan to switch web hosts for its new health insurance portal, HealthCare.gov, in the midst of an expected last-minute rush to beat a March 31 enrollment deadline for 2014 coverage.
Switching hosts is not in and of itself a huge risk if it is done carefully and with lots of preparation, according to technical experts interviewed by Reuters. It is the timing of the highly complex maneuver that is risky. If there are problems, the website could become sluggish or even unusable for anyone trying to enroll.
The government is tempting fate, they said.
The good news in all of this, for a given value of ‘good:’ it’s not happening this week (that would be terrifying). It’s scheduled for the end of March… when the government absolutely needs the system up and running, period. Which suggests that the aforementioned given value of ‘good’ may not actually be. And maybe it’s terrifying anyway, come to think of it.
Does… does the government have any IT people? The “nonsensical in all dimensions” alarmed quote is from a top-level IT guy at Health Evolution Partners, which is a health care company investor – which means that it’s an informed alarmed quote. Easy to understand why, too: switching hosts is, ah, uncontrollable. Sometimes it works… fairly well. Sometimes it does not. And when it does not work, it tends to not work in a rather spectacular fashion. I’ve been through a few switchovers myself – heck, probably most of my readers have, too. So we all know that switching hosts is a great way to discover all of those patches, kludges, workarounds, and perennially slow-burning problems that had been conveniently forgotten or tacitly ignored up to that point; largely because those patches, kludges, etc. will take the move as an invitation to explode.
Now, of course, sometimes you have to do a switchover anyway. But is this really a good time for healthcare.gov to do the switch? The site is not actually operating at an acceptably functional level (they’ve had to delay yet another deadline) and the original implementation timeline was taken out back and shot more than a month ago. The Obama administration has pretty much no goodwill left with which to offset any further disruptions of what I will charitably call ‘service,’ particularly if that disruption can be seen as being actually unnecessary.
Or, put another way: Obamacare already has a horrible reputation. If the site blows up just before the March 31st deadline, that reputation will not improve.
Via Jim Geraghty.
Moe Lane (crosspost)
I have only managed comparatively simple web sites and yet have discovered that, for example, changing from a Linux host to a Microsoft host at the same provider can smash your site in the MOST amusing ways.
Should be fun to watch.
With only the best intentions, I cheerfully wish them all the — um — educational events in the world about their beloved program. I have long since given up any hopes — actually, I never had any — that they will learn one blasted thing about actual common sense when it comes to supplying a needed service, which Obamacare is only because they carefully wiped out people’s health plans first.
This is what happens when corrupt government and corrupt government sycophants attempt to run something top down. I firmly believe that they hired the cheapest web designers they could find (and therefore, the most incompetent) and then cut every possible corner in developing the site in order to divert as much money to themselves as they could.
Result?
The fiasco we know as the Obamacare website.
Will they be implementing Disqus?
.
Mew
Ha!
A well developed theory of Magic is great, to bad it tends to fall apart in the real world.
Not to mention all the Happy Fun Problems that only show up when you put a load onto a large system.
So the thing is, if they did their jobs correctly, this really should be as simple as flipping a switch. Web hosts go up and down all the time and you as a user never notice, because bringing up new ones either happens automagically or or is very easy. As for being in a different DC, the infrastructure is already spread out over many different datacenters, so that really shouldn’t be an issue.
So what are the odds they did their jobs correctly?