There is a bit of a conundrum with the Internet. The more popular a site is, the more likely it is to fall over. This is especially true when you get a sudden spike of popularity.
I can remember on September 11th, hitting the news sites to be met with a spectacular array of error numbers but mainly telling me “server too busy”, at the very point the world really needed the Internet, it killed it in the process. The only way I could find out anything on the Internet was too seek out a less popular news site and try there. I eventually got some news from the then fairly new, Ananova. Truth was, the only real way to get any news that day was to phone someone who could see the infinite power and bandwidth of a TV.
( This sort of thing happened again recently after the sad death of Steve Irwin. )
Things haven’t changed much over the intervening years. A number of times I have seen the BBC website choose to feature a new site or service, only to temporarily kill it in doing so. You can imagine the chat in the offices of the site:
“Hurrah, we’re featured on the BBC website…”
“Only, no one can see what our site does…coz its now dead.”

The basic rule is, if a high performance/high hit rate site features a site of less power and bandwidth, there is a good chance it will kill it with the flood of hits. The site will only gain the visitors that remember to go back and try later, which is far less than they would hope for from such exposure.

I saw this again this week, when my favourite photo site MorgueFile suddenly got a massive peak in traffic because of an article on Digg (well done to Michael and the gang for keeping it just about alive).

This is clearly something that the next generation of hosting has to address. There is no point in having a service that is almost guaranteed to fail at the point of maximum demand. Neither does it make sense, economic or otherwise, for sites to plan for the peaks and have a server farm sitting and waiting on the off chance. The ability to turn up the power and bandwidth virtually instantaneously (on-demand or automatically) will be a killer feature of hosting. This has already started to emerge, particularly in a lot of the hosting that BT is starting to offer (although with an ironic quirk of fate, when I went looking for a link to see what hosting they are providing now, their site was down!).

If you are unable to read post its most likely because it has just become massively popular

Technology moves very fast these days (Cybil Fawlty, specialised subject ‘the bleeding obvious’).
Not so many years ago, well, 26 years actually, I had a ZX81 and it was the pinnacle of home computing. And, believe me, it was fab. Thing is, it has no particular legacy now, only a museum like curiosity that was once kept for a difference engine or Colossus.
The technologies that I want to take about are those that leave a more permanent echo. Two candidates are mobile phones and laser eye surgery. No one knows what 20, 30 years of mobile phone use can do to you. Similarly with 20 year old laser’d eyes. “It’ll be fine” they say. Supposedly the same people that used to sell asbestos to builders. I’m not a scaremonger, I love technology, advocate it and use it widely (although you can stick your laser eye surgery where the sun don’t shine). I’m merely using these examples to illustrate the fact that we tend to get so caught up in the new and the now that we rarely look at the long term impact.
One of the potentially most interesting forms of this is the internet itself. We are creating content now that could still be around in 20, 30, 50, 100 years time. My grandchildren might read this very blog post and think “what an old twat” and click on the link to a now neurally implanted version of Wikipedia to find out just exactly how little power a ZX81 had. When my daughter is older, her teenage friends might track me down on the Internet and tease her at school because “her Dad is such a dull square and there’s no way to you were on a billboard in Las Vegas”.
The thing is, new technology today isn’t all disposable future museum pieces, it has the ability (and perhaps role) to persist, to form part of a personal, cultural historical record.
We can already go to the WayBack Machine to see how gloriously crap the original BBC website was. Who knows where this post, along with the rest of the internet content, will end up in the future. Will there only ever be one transitory internet, with sites living and dying with us? Or will we devise a mechanism to maintain an on-going record (the WayBack Machine being a laudable but limited early attempt) so that this will post and many millions more like it will remain beyond my/our lifetime as a record of this time?

Website of the day:
Well, how apropos (was it deliberate?), have a look at this. People are already seeing that the passage of time has a place in applications of today.

Track of the day:
I was taken with Panther Dash by The Go Team! tonight. You can never have too many harmonicas.
( all together now grandkids of the future, “Granddad that is soooo 2004…”)

My reader(sic) may notice that I have added Snap to my site and blog (if you have no idea what this is, hover over an external link and you’ll see). I suppose there are many arguments with regard to its usefulness but for now the novelty of it is tipping towards the ‘quite cool’ end of the scale.
I have to say it is a really good implementation, a doddle to integrate and it seems to render the page previews very quickly. Obviously, with my penchant for the funky client-side implementation, I’m always going to be a fan.
Does make me think that it could change the surfing dynamic a little if it takes off, people could perhaps see a lot less spurious hits with low wait time.
I first saw it being used on JP Rangaswami‘s excellent blog Confused of Calcutta which I would encourage you to read.