Vonage Hits ISP Resistance
Owen DeLong
owen at delong.com
Thu Mar 31 22:41:32 UTC 2005
> Heard of a little thing called a 'rhetorical question'?
>
> Who decides that it is okay for ISPs to block SMTP and not okay for them
> to block VoIP? If it is okay to block SMTP because "people do bad
[snip]
Well... Here's how I define things:
1. Blocking ports is bad.
2. Certain chronic abuses are worse.
3. When something is blocked because of chronic abuse, the following
conditions should be met:
+ Any customer should have the option of having said block
removed from their traffic upon request.
+ All customers should be informed of the block either at
the time it is instituted or as an addendum to the contract
for service when they sign up.
+ The block should be the least intrusive most selective block
possible to reduce the abuse to a tolerable level (i.e. a
level at which it is not preventing legitimate use of the
network).
+ The block should be regarded as a temporary solution until
a better way to resolve the abuse can be found.
+ The block should be removed at the earlist opportunity
once the previous item has been accomplished.
In this case, I can see reasons for blocking client<-->relay and/or
relay<-->relay SMTP access as default under current circumstances.
I can't see any such reason for SMTP. If you're running a network
for <$40/mo flat rate subscribers, then, I believe your cost model
may require you to block certain broadband isochronous
services (internet radio, voip, etc.) or at least QOS them to
the point where they lose if others want the bandwidth in order to
provide reasonable service to all of your customers without your costs
exceeding your revenue. Customers who want these isochronous services
have the option of paying more for the bandwidth they need, or, they
can go to a provider that provisions for this (and will likely cost more).
I don't see a case for blocking VOIP from competitors if you are selling
VOIP. I see that as likely an antitrust issue. I don't see a case for
blocking NNTP currently.
Anticompetitive blocking is bad. Anti-abuse blocking is bad, but, not as
bad as allowing the abuse to prevent the normal function of the network.
Blocking of isochronous high-bandwidth services to support higher
levels of oversubscription for lower-priced service is not unreasonable.
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/20050331/5ed7006c/attachment.sig>
More information about the NANOG
mailing list