Some of you may have noticed that several parts of the BBC website were down for several hours last night. I don’t know if this had anything to to with 10:10:10 - but it any case it brings to light the continuous fragility of the Internet and web hosting in general.
It is inevitable that everything that is UP must at some time come DOWN, it’s not so much about the frequency of how often this happens, but much rather about how well the Internet stays up and how much we depend on the Internet in daily life now - the actual incredible power and resiliance of the Internet as a whole.
We know from personal experience that with ongoing software development and updates and DOS attacks, viruses and worms - every day’s worth of hosting is a sequence of mini battles to hold off the onslaught that might some day momentarily disable your site. For our high performance Affino hosting, we have a vertiable quadruple fortress of hosting defences - using all manner of clever systems and applications to keep those that might do damage at bay.
Almost regardless of the provisions you have in place you will always encounter something new and somewhat unexpeted - when this happens it’s a case of how quickly you can deal with the problem and restore normal service again. All the big sites suffer from the occasional downtime - whether Apple or Twitter, Amazon or Facebook, or Google.
It’s a little bit like an electricity blackout in some ways - regardless of who’s site goes down, I am always enormously relieved to see it back up again - just as I was when the BBC site magically reappeard once more without its 500 Error screen. Weekends always mean that a reduced skeleton staff is on duty, and because of this - if something does go wrong it often takes longer to solve the problem than on a weekday. Beyond all the built-in hardware and software defences, you also have the posibility of load-balacing across several servers, or even clustering several different server farms for rapid switchover when things go badly wrong.
It’s not so long that the Glastonbury site went down under the clamour of music fans trying to get tickets for next year’s show. Website hosting is really very much like building flood defenses - you put in place a provision for what you believe will be the maximum impact level, yet you often find out that your significant provisions have been rendered inadequate by an even higher level of traffic than what was envisaged. Of course having extraordinary powerful flood defenses makes for an enormous cost - as if putting in place defences to withstand a tsunami - when tsunami’s rarely occur. With most modern hosting, restoration of services can be made very rapidly and seamlessly typically - in 99% of cases you would expect services to be restored within mimutes, an hour at the most. What knocked out the BBC for several hours then is an intriguing mystery - who or what was responsible? No doubt there will be some major investigation to discover what went wrong, for me though - I’m just happy that everything is back up and functioning well again.
We modern consumers have become so used to the good life now that any slight disruption seems to cause quite a disproportionate level of discomfort. It’s really not so so long ago that civilisation used to suffer from extended power outages and blackouts as the electrical grid of that time could not cope with the increased demands placed on it. I am always mindful that every current technology has a finite limit. We are just very lucky that there are always people around who can invent and innovate new technologies that help take us to the next level. As a student of Geography, every 30 years or so you read some paper about how the planet can no longer sustain food production for the current level or population - yet we always somehow manage to endure. The same is the case for the Internet, whose demise has often been predicted - this is for sure a marvellous technology, and the fact that it is up at all and works as well and as consistently as it does - well that is a marvel in itself ...