The Internet routes around problems
People often say “the Internet routes around problems”. This harks back to the days when the precursor to the Internet was designed to be resilient against attacks through the use of redundancy.
It’s still true, to an extent, though depending upon where in the cycle we are, you might notice the strain being put on particular pipes because two routes from A to B have been reduced to one.
When people say it nowadays, though, they’re not really talking about actual routing protocols, they’re talking about the architecture of applications which use the Internet.
Most applications use TCP, which itself sits atop IP. There’s usually some DNS in there, and quite a few applications run on top of well-known high-level protocols like HTTP or SSH. All of these things can change. Trivially.
HTTP doesn’t much care whether you’re using TCP or SCTP, for example. You could, if you really wanted to, do HTTP over SSH without too much difficulty. CVS used to run atop RSH. Then SSH. Then CVS was replaced with Subversion, which can use HTTP (or HTTPS) or SSH. Then Subversion was replaced by git, Hg, bzr, and so on. HTTP itself replaced Gopher. SFTP replaced FTP. RSS (which uses HTTP) mostly replaced NTTP. IPv6 will (most people think, given enough time) replace IPv4, still with TCP, UDP, HTTP, SSH, NNTP and so on all running on top of it.
This is why the people who keep our Internet connections fed and watered just sigh a little when they hear talk of efforts to “regulate” the Internet. When such things are spoken of, they rarely mean “ensure proper training such that law enforcement can apply existing laws properly to things which have made use of the Internet as a communications tool”; instead they often mean “identify and prevent some application”.
This rarely comes to anything, though, and there’s a good reason for it: were you to draft a law, for example, to ban some particular application, how would you make it broad enough that a change in protocol or technique didn’t render the law worthless, and narrow enough that you don’t implicate just about every technology in use on the ’net today? You soon realise that in order to do that, you have to deal with the underlying activity, and most of the time, the things they want to ban are already illegal.
The Internet has become so important to the economy and our day-to-day lives that were a blunt instrument applied that massively restricted individual’s use, either the instrument would be ignored by pretty much everybody, or there’d be a revolt.
So when there are mumblings about “mandatory content labelling”, for example (something which could only ever be either useless, or so inflexible as to render most e-commerce illegal), it’s worth bearing in mind that they make no sense: just because it’s possible to “label content” in an HTML document or in HTTP headers or somesuch, what happens when content isn’t in an HTML document, or we’ve stopped using HTTP altogether? The Web has been in widespread use by ordinary people for about a decade, now. That’s nothing. No time at all. If anybody in a position of authority believes that the architecture of mainstream Internet applications will stay the same for the next ten years, they deserve all the ridicule they get.
