The following text is copyright 2003 by Network World, permission is hearby given for reproduction, as long as attribution is given and this notice is included.

 

In the end, it's the end that counts

 

By Scott Bradner

 

From the beginning the end was the important thing.  One of the seminal  papers in the history of the Internet is "End to End Arguments in System Design" by Saltzer, Reed and Clark.  This paper laid out the basic idea that the Internet should support applications running over it by not caring what applications are, in fact, running over it.  It is this idea that created the incredible diversity of services and applications  reachable through the net, it is also the idea that, over 20 years after it was first published, still scares the begezus out of governments, carriers and some standards organizations.

 

The paper, (http://www.reed.com/Papers/EndtoEnd.html) often referred to as the "e2e argument," is more than a factoid of history.  It, along with David Isenberg's "The Rise of the Stupid Network" (http://www.isen.com/stupid.html), describes the key feature of the Internet that must be preserved in order for the Internet to continue to live up to its potential.  That feature is that the Internet does not get in the way when a device at one "end" of the net wants to talk to a device at another "end".

 

In too many places this feature has been badly distorted by the installation of network address translators (NATs) and firewalls.  In an ideal e2e world a user who wanted to create a new application or service would not have to get permission from the managers of the networks between his machine and the machines of people who wanted to use his services.  But it is not an ideal world, people put in NATs to deal with real or imagined difficulties in getting address space and put in firewalls in a band-aid approach to security (see Crustacean security (http://www.nwfusion.com/columnists/2002/0211bradner.html).

 

Open networks scare a lot of people, not least of whom are the carriers and governments.  Carriers, because they think they need to control what their customers can get to so that the carrier can charge extra to get "special" services.  Governments, because they want to protect the poor users by controlling what the users can do and see on the net and because they want to watch, just in case the user does something naughty.

 

Some people are trying to at least get the carriers to provide an Internet service that includes all of the Internet and not just a controlled corner.  (See http://lists.elistx.com/archives/interesting-people/200212/msg00053.html).

 

But one of the things that has been lacking is a good, clear, well thought out on-line resource that people like me can point people to so that they can find out just what the importance of the e2e argument is and what will happen if the e2e Internet is lost.  There is now such a site that I recommend most highly.  It is called "World of Ends." (http://worldofends.com/)  Please do not just take a look, take the time to read it and you will see what the importance is of an end in the Internet world and what is at risk if the ends get captured.

 

disclaimer: Some of Harvard's schools understand clarity, some teach how not to be clear, but none cooperated in this opinion.