To differentiate between different
types of internet traffic, as described in the article (eg. distinguishing between web surfing and P2P filesharing) doesn't require any "deep" inspection of the traffic. Each packet sent by an IP application (like a web server or filesharing program) has what's called a source or destination "port", in the same layer of data as the source and destination IP addresses. The ports are just numbers that can be used to determine what the type of traffic is. Ever seen a URL like "http://somesite.com:8080/"? The "8080" is the port address (normally web requests go to port 80). There's nothing deep about it - in the sense that you don't need to inspect the
contents of the packet.
So the article is completely incorrect where it says
"What we're really talking about," says IDC Canada consultant Lawrence Surtees, "is packet-sniffing, like <U.S. President George W.> Bush is doing.
To do packet-sniffing a la NSA, you need to look at a different layer of the information in the packet - that which contains the actual content. To do bandwidth shaping, you just need to look at the address data. It's the difference between looking at an envelope and looking at the letter inside it. Not that it's very much harder to look at the information content (unless it's encrypted) of course.
The technology to allocate bandwidth on the basis of address data has been out there for years - it's been a standard part of the Linux kernel since about 2001, I think.
I used to admin a neighbourhood LAN in an inner-city area in the UK, and when P2P applications started to get popular we'd find the whole network suffered - web access would become painfully slow because the bandwidth was clogged up by P2P programs. I had no hesitation in using Linux's bandwidth shaping capabilities to preserve a small dedicated chunk of bandwidth for web access, which let web pages load at normal pace again, while giving the P2P programs free rein in the remaining bandwidth. I think a lot of campus network admins will be doing the same.
I'm not saying ISP's
should be doing this (though having been a network admin for an ISP, I would have some sympathy with a little light bandwidth shaping on their part, if it made their service more usable for the majority of their customers).
But I think bandwidth shaping based on port addresses has little to do with the actual threat of removing net neutrality.
That has more to do with prioritizing packets over others
of the same type - making web pages from google slower than pages from microsoft.com, for example, unless google cough up money to the ISP or network carrier. That is bandwidth shaping based not on the traffic type (or port number) but based on the ownership of the source and destination addresses (ie. is it www.google.com or www.microsoft.com), and
that is what all the fuss over net neutrality is about.