[c-nsp] Something I was thinking about whilst idle the other day.

[Gert Doering]

Hi,

On Thu, Mar 20, 2008 at 08:22:23AM -0400, Drew Weaver wrote:
>         As someone who works for a company that has potential
> customers that use trace route to determine whether or not to use
> one company's service over another [yes, I realize that is a very
> poor metric but it is a very competitive market] I would like to
> interject here for a moment. If you sign someone up for a service
> (say internet transit) where does it say that you can choose which
> packets to pass and which packets to drop, and where does it say
> that you can (now a days) not block RFC 1918 [bogon] traffic and
> pass it along to your customers? 

People sign up to connect to the Internet.  So that's what we do: *forward*
their packets at full speed.

Our customers have not signed a contract that specifies their right to
send unlimited numbers of packets *destined to our routers*.

Since the second sort of packets can harm the first sort of packets (if
the router is overloaded, transit packets might be dropped or delayed),
we reserve the right to prioritize transit packets.

Which is exactly what is happening: transit packets are handled full speed,
while pings to the router, or traceroutes (that need to get handled by
the router CPU) are processed with lower priority, and with rate limits
in place.

This is a trade-off to be able to *better* fulfill the service that the
customer is paying for.

> I think people expect traceroutes
> to work because they use them for diagnostic tools, when they don't
> work they assume something is wrong with your network.

I like traceroute, and really get annoyed when networks hide their
topology and/or break traceroute right away.

But I do know how to *read* a traceroute, and I know that the occasional
hop with a higher RTT doesn't mean "the network is broken" if the hops
after that are fast again.

gert
-- 
USENET is *not* the non-clickable part of WWW!
                                                           //www.muc.de/~gert/
Gert Doering - Munich, Germany                             gert at greenie.muc.de
fax: +49-89-35655025                        gert at net.informatik.tu-muenchen.de
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 304 bytes
Desc: not available
Url : https://puck.nether.net/pipermail/cisco-nsp/attachments/20080320/72b73981/attachment.bin