If Microsoft is trying to lead by example, it is having a very, very bad week.
First, there was The Lone Engineer Theory. Microsoft said a technician made some changes to its Domain Name System routing configuration at 6:30 P.M. Pacific on Wednesday that effectively cut off access to many of its Web sites. Microsoft still hasn't fully explained why it took almost 24 hours to correct the error.
Then Thursday, there was The Evil Hacker Theory. Apparently sensing an opportunity to kick Microsoft while it was down, someone or several people instigated a denial-of-service attack against Microsoft's DNS routers - according to Microsoft - that lasted for at least two hours. Microsoft said it has notified the Federal Bureau of Investigation of the attack, and said the Thursday DNS problems had nothing to do with the Wednesday DNS problems.
Regardless of the source of the infrastructure problems, it's almost inconceivable that Microsoft would leave itself vulnerable to such outages. Microsoft, which fancies itself the industry's leading Internet software and services company, evidently did not have a fail-safe system to protect its DNS routers.
Consider what would happen if other industries had a similar lack of safeguards. Take FedEx, for example: Uh, we accidentally forgot to fill up our single New York City truck with gas so we'll have to get you that package tomorrow.
The problem, according to DNS experts, could have been prevented - or at least fixed more quickly - if Microsoft had set up its servers differently. Microsoft was running all of its DNS servers on the same network segment, a setup that created a single point of failure for its DNS infrastructure, according to Men & Mice, a DNS diagnostic software company based in Iceland.
And yet, many sites are similarly at risk. Men & Mice said it conducted a random sampling of 4,910 dot-com domains and found that 38 percent of those sites had the same DNS configuration problem that left Microsoft susceptible to its prolonged outage.
The only compensation Microsoft is able to offer its millions of users - many of whom probably depend on its free Hotmail service for business correspondence - is an apology: Sorry, we screwed up. A notice posted on its Web site late on Wednesday said, "Microsoft regrets any inconvenience caused to customers due to this issue."
Of course, paying customers routinely get shafted by service outages. Internet providers and phone companies - not to mention cable operators and, in California these days, electric utilities - experience service outages with frustrating regularity.
So is it unfair to single out Microsoft as the poster child of flaky Internet service? Granted, everybody experiences technical difficulties sometimes. And yes, the Internet is based on relatively immature technologies. But if Microsoft would like us to believe that it can sell the most reliable computer software in the world, it should act the part. I'll bet Microsoft will be better prepared next time.
Todd Spangler is Matrix Editor at Interactive Week.