大象传媒

I remember living through the Cuba crisis where, through my childish eyes, it seemed the fate of the world depended on the moral actions of 2 powerful men. Now it is my adult blind faith in the view that the scales of dependancy on the internet/www have tipped so far over into need we will find altruism, commercial gain and political power will unite to keep it from destruction.

Sometimes it is difficult to produce hard facts at short notice, but I very much look forward to hearing them from more knowledgeable sources too.

I tend towards scepticism when I hear reports of the web's imminent downfall. These don't seem to come quite as frequently as they once did - remember the succession of aspiring bogeymen that came and went (Sircam, Code Red, etc) in the first few years of the century? I also remember Steve Gibson's fears that Windows XP would create - probably a bit hysterical, in hindsight.

So the boy's cried wolf a few times; but can we count the lupine threat as non-existent? I don't think so.

in February 2008 was an example of the vulnerability of DNS propagation. In this case, a local blacklisting propagated to the world, leading to the temporary removal of one of the world's most popular sites. Could this aspect of the internet's structure be exploited?

Also, botnets have increased in sophistication. If you're unfamiliar with their power and scope, it's worth reading up on the Storm botnet, which reached its peak in September 2007. At its height it was 'intelligent' enough to people who tried to research it. Although Microsoft claim to have crippled Storm last year, more powerful replacements might one day emerge.

Today, I see the main threat to the web coming from something with Storm's power coming under the control of a group motivated by ideology rather than money. However the threat is being diminished by the slow erosion of Windows' monopoly - a more diverse computer ecology will make things harder for botnets of the future.

@wisepacket

One small correction: Pakistan鈥檚 accidental obliteration of YouTube was鈥攁s far as I know鈥攁n issue with poor BGP filtering, rather than DNS. DNS does have its (widely-known) flaws, but a Pakistani ISP would have had to try an awful lot harder than it did if it wanted to block YouTube for everybody by exploiting them ;)

The fact that NANOG folks were able to recover from the YouTube incident quite as rapidly as they did demonstrates quite neatly how the Internet works 鈥渂ehind the scenes鈥�. It's about people working with other people with commercial and national boundaries mattering far less than they ordinarily would.

Take the 大象传媒鈥檚 peering policy for example (and they鈥檙e by no means unique in their approach). They prefer _not_ to have a formal peering agreement, though will draw one up if pressed. Such things tend to get in the way of actually making the network work.


It鈥檚 worth remembering that the web, and e-mail, and even the assorted botnets out there are not themselves the Internet, just applications of it. Sometimes routes fail. Sometimes ISPs have insufficient redundancy in places where they really should. Destroying one particular application is relatively easy. Destroying the Internet less so; it'd require dismantling, piece by piece, and doing it more quickly than dissenters put it back together again. Good luck with that.

And yes, there are locations which鈥攚ere you a dirty terrorist鈥攜ou could plan a coordinated attack against. Certainly, you could make life difficult for one country鈥檚 Internet traffic for a while. Well-timed explosions in the right parts of Docklands, Manchester, Cambridge, Leeds and Edinburgh would have a pretty devastating effect. For a day or so.

I think the main issue here is probably a mental one within the population. People who know what they are doing (and have enough need to) would still be able to route traffic across the place through private networks where necessary. The military and financial systems being examples.

But to the average web user, an attack which stopped them being able to type in "google.co.uk" and get to a search engine would play quite a noticeable mental game - regardless of if it was only that one domain taken off-line by some kind of DOS attack.

Thus there is no single point of failure - every web server becomes a point for a mental attack on any country. And that's an awfully large amount of nodes that could be attacked.

The whole #whentwitterwasdown thing was interesting to watch? I don't condone hacking, but it was funny watching the Twitter equivalent of panicky headless chickens. I didn't even know that Twitter and Facebook had been brought to a halt until much later and I wonder how many other users actually missed the incident too? If Twitter or Facebook weren't as popular a site as they are would people care? Is it a case of "I'm alright Jack as long as it doesn't affect me." Would people really be interested that a Georgian blogger was being targeted if it didn't affect their daily fix of their favourite site?

There is more than one way to skin a cat. The Internet, like all evolutionary emergences is dependent on underlying systems. Electricity for one. Electricity is arguably the largest single point of failure on the Internet. However, should all electricity fail, the Internet would probably be low down on our list of priorities to restore.

Is it possible for all electricity infrastructure to fail.. it may be if you pay any heed to NASA and their report earlier this year on severe space weather.

One thing that is probably true is that over a long enough time line the Internet will fail. I would hazard when it does there is a high probability that it may fail due to some "unaccounted" for factor. Maybe the singularity will actually break it...

This comment was removed because the moderators found it broke the house rules. Explain.