Creating a crystal clear and pure Internet

Deepak Jain deepak at ai.net
Wed Nov 28 00:14:11 UTC 2007



Valdis.Kletnieks at vt.edu wrote:
> On Tue, 27 Nov 2007 09:38:40 EST, Sean Donelan said:
>> Some people have compared unwanted Internet traffic to water pollution, 
>> and proposed that ISPs should be required to be like water utilities and
>> be responsible for keeping the Internet water crystal clear and pure.
> 
> What's the networking equivalent of "remember to build your water intake
> *upstream* of your sewage plant"?
> 
> Or, more accurately - "how do you get all those people with private
> wells^Wcomputers to *not* insist on building their leach fields uphill
> of their wells?".
> 
> There's a limit to what an ISP can do to make it "crustal clear and pure"
> without an incredibly intrusive presence.  The technically easy way is what
> many corporations do - Borg the boxes into an Active Directory domain, and
> impose fascist controls via Group Policy (for all of my anti-MS ranting, I'll
> grant the AD/GP stuff *is* pretty slick ideas for corporate PC lockdown).
> 
> But how do you sell that idea to the consumer user?

I think we'd have to standardize on what our networking equivalent of 
"water" is. Are we talking about just port 80 traffic? Just email?

A water utility is not an open network, nor is it bidirectional. If I 
want to start transporting Koolaid, I need to take the water from the 
utility, and create my own distribution network outside of the water 
utility; I can't lease capacity from the water utility.

If all we were carrying were unidirectional traffic, "crystal clean" 
would be very easy. If all we were carrying was port 80 traffic, "pure" 
would be easy.

A better analogy would be make sure all the roads and vehicles on it 
were crystal clean and "pure"/efficient -- hell, I'd settle for "insured".

IMO, an intrusive Internet (ignoring the political talk about MSFT or 
not) would not be the Internet, but just a proprietary network [pick 
your flavor].

But rather than debate technology, I think regulators/operators/etc 
would need to settle on an unambiguous definition of "what" is carried 
and "what" isn't to be carried, at what quality level for a given level 
of service.

Deepak Jain
AiNET



More information about the NANOG mailing list