Wed 2 Jan 2008
ARPANET, the “1.0″ version of what we now comfortably call the Internet, has a history going back forty-some years and roots in a defense initiative to create a durable communications network — one that could withstand damage and still function. History lesson adjourned. Now:
At which point — at what practical limits of direct and indirect disruption — would the modern Internet fail? And what levels of failure are we talking about?
Is the Arpanet initiative’s goal of a flexible network even relevant with the rise of ubiquitous public and private networking?
Posted by Josh Millard6 answers so far!
I don’t quite know enough technologically to answer the question, but this article seems relevant, in terms of the the available technology and infrastructure we have simply failing to be able to continue supporting the amount of information that we keep pumping into it, in the near future.
Judging by the traffic in here, it’s already happened.
Since the internet is a composite of peers, it’s unlikely for the whole internet to be killed by a singular event or even a strategic takedown. What tends to happen is chunks fall off. A savvy saboteur will calculate which chunk to take down and work to that end.
Ideally, enough of the internet will continue that the chunks will reattach or be replaced by other chunks. Any chunk may be able to survive on its own — for example, whenever Japan falls off the global internet (it used to happen somewhat often) its residents still have their domestic internet.
In a large-scale cataclysm, the politics of IP number allocation is probably going to be a greater hindrance than the technology. Assuming something happens that permanently wipes MIT off the map, who has the authority to reallocate MIT’s /8 block? That’ll be a lot harder to resolve than the labor of re-provisioning the survivors who got their network access through MIT.
That said, there may be a calculable minimum effort required to pull down the U.S. internet. Not so much knock it out completely but make it unusable for our current expectations, both in terms of bandwidth available and territory accessed. And it might not be noticed by anybody in China. Massive failure rarely looks like failure to the end user — instead, traffic gets arbitrary and slow.
I am not a whiz either, but I found this article pertaining to something called “denial of service” attacks. I guess these have been around a while but I’ve never heard of them before.
My first initial thought was a virus or a worm. I remember when the Melissa virus struck and I worked for a big company with fancy firewalls and topnotch network admins. They caught it right away and put out warnings, but it had still spread.
Can they write worms that attack routers? Because that would be the most efficient way to go in my mind. Or anything that prevents a hardware handshake anywhere along the line.
Unless you get a bunch of rebel commandos planting bombs at key major backbone providers around the world.
Judging by the traffic in here, it’s already happened.
The BBQ needs to be such that any random moron can BS about it. This particular moron looked at today’s offering and realized it was impossible to pretend he had anything valuable to say.
Hah, Go Meatbomb.
Yeah, half the questions that pop up in here I look at and think “Weeeeell if I say anything I’ll destroy that carefully-crafted image of a semi-educated, semi-literate middle-class citizen I’ve built up, SO… *twiddles thumb*”