No I'm not talking about a poker hand or the 49ers. The concept of four nines or 99.99% is a performance statistic for service providers that considers a 0.01% downtime acceptable. In other words if you are a hosting provider your servers can only be offline for 52 minutes a year. This is practically impossible given the unpredictable nature of hardware but more importantly because of necessary upgrades, which require the entire system to be taken down.

Vinnie Mirchandani recently featured an interesting post called The mythical 99.99% uptime SLA that illustrates the absurdity of four nines reliability. He notes that companies like IBM and EDS often schedule more than 52 minutes of downtime per month, and in some cases outages can last several hours. But as he notes most companies serve fewer than seven times zones and as such have a window for scheduled outages during off peak hours

Still it's hard to convince clients that your service will "be there" when they need it if you don't retain some semblance of reliability. At best what you can hope for is no downtime during the busiest hours and a limited amount of time during the graveyard shift. If you find yourself in a position of downtime during heavy traffic or critical customer periods then you definitely have a problem that needs to be corrected immediately.

In times of tech trouble, sometimes laughter is the best medicine

Everyone hates it when the servers are down, none more than the lowly techs that are supposed to keep said servers up and running. IT folks endure a lot of abuse from angry users who don't understand that equipment failure is part of life and that things often fail without warning or reason. Since techs often have a twisted sense of humor and obvious affinity for cats it's not surprising that many humorous "system unavailable" images have cropped up over the years. The following montage was inspired by one in particular that Twitter allegedly has used.

The strange thing about technology is that after all the hype and build up inevitably it seems to break all at once.

I'd like to think that tech desires to break on Friday afternoon at 3pm. I'd like to think that there is something malicious in the way that a piece of hardware or software acts when the pressure is on, or more importantly I'm out to lunch. I'd like to believe all of these things but deep down inside I know that hardware and software are just things. Deep down inside I know that Murphy was a realist.

Some have said that to err is human but to really foul things up requires a computer. Whoever utter this bit of dime store wisdom obviously never met the idiot users who repeatedly click a non functional button in the hopes that one of these tries I'll start working. I suppose they sit there, patiently waiting for their Pavlovian crack tablet, pondering whether they are pressing the button or if the button is forcing them to press it. Regardless they continue to press the button until the entire world ends or some bald headed guy smashes the computer and then truly all hell breaks loose.

Buttons and users aside have you ever wondered why Murphy or anyone else for that matter never came up with a more true statement about machines like: Computers fault, users are at fault. That is to say that a computer or any other machine can suffer from a defect or mistake in construction, but where blame is required truly there is a user. In other words: Maybe it was a mistake to put that button there, but on the other hand you're the one that keeps pressing the damned thing.