Points on your Internet driver's license (was RE: Even you can be

Sun Jun 13 17:02:59 UTC 2004

> And that is a problem. Unlike your electricity, where the supplier has an
> obligation to provide a certain level of clean energy, there is nothing
> like it with internet bandwidth. All the crud and exploits are dutyfully
> forwarded to the customer.
>
Clean internet service is internet service that delivers only valid IP
datagrams.  Most internet service is clean internet service.  Any internet
service that looks above layer 3 to make forwarding decisions is not clean
internet service.

> I argue that this is way overboard. I don't believe anyone should require
> any particular knowledge to obtain an internet connection and use the
> internet. Instead internet needs to be available as a clean conditioned
> service for consumption by the clueless.
>
I agree that the IDL is overboard.  I even agree with your second sentence.
Consumers need to demand software which does not support these exploits from
their software vendors.  That is the real solution.  The internet is a
transport, just like the phone line coming into your home.  Nothing prevents
someone from making an obscene phone call to your house.  The most common
problem software today is like having a telephone that won't let you hang
up on the prank caller, then, demanding that the phone company prevent those
calls from coming in the first place.

Problem is that people understand that TPC can't tell a prank call from a
legitimate one, but, for some reason, they expect ISPs to be able to 
magically
tell whether this HTTP session is an exploit while this other one isn't.

> The reason this isn't economical today is because ISP lack any
> responsibility. It is cheaper for an ISP to buy more bandwidth and pass
> the  worms and viruses customers PCs spew to the internet than it is to
> deal  with the problem. Seriously, if I send an ISP reasonable proof that
> a  broadband customer hits my mailserver with thousands of emails an hour
> I  should be able to expect an immediate response. Not hours, days or
> weeks,  minutes and the originating account should be shut down. If this
> doesn't  happen I should be able to go to the upstream of the ISP,
> present my  case, and have connectivity to the ISP suspended.
>
The reason is that the ISPs can't tell the exploits from the legitimate
traffic in most cases, and, even if they did, do you really want ISPs making
value judgement about content on behalf of their users?  That's a really
bad model.  It's just not good for innovation, free speech, mom, or apple 
pie.
Yes, ISPs should investigate abuse complaints and immediately disconnect
users that are spewing abuse.  Yes, this needs to happen more consistently
and more rapidly.  However, content filtration at the ISP level is not a
solution, it's just a different problem.

Owen

-- 
If it wasn't crypto-signed, it probably didn't come from me.
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 186 bytes
Desc: not available
URL: <http://mailman.nanog.org/pipermail/nanog/attachments/20040613/808fc4e4/attachment.sig>