Some truth about Comcast - WikiLeaks style

Rich Kulawiec rsk at gsp.org
Fri Dec 17 13:26:46 UTC 2010


On Wed, Dec 15, 2010 at 07:14:01PM -0500, Jon Lewis wrote:
> On Wed, 15 Dec 2010, Rich Kulawiec wrote:
> >That's rich, given the enormous quantity of spam sourced from Comcast's
> >network over the last decade.   (And yes, it's ongoing: 162 unique sources
> >in the last hour noted at one small observation point.)
> 
> Spam is irrelevant.  In this context, abuse = sending large amounts
> of data to Comcast customers (at their request) without paying at
> the Comcast toll booth.

Yes, I know; I did read that in context and understand the point
the original author was making.  I probably should have made that clear.

> >Now I realize that SMTP abuse isn't exactly the most bandwidth-chewing
> >problem.  However, it's a surface indicator of underlying security issues,
> >which in this particular case can be summarized as "one heck of a lot
> >of zombies".  Given that those systems are known-hostile and under the
> >control of adversaries, it's certain that they're doing all kinds of
> >other things that chew up a lot more bandwidth than the spam does.
> 
> It might even "improve" their ratios if they stopped those zombies
> from sendig spam, participating in DDoS's, etc.  After all, that's
> outgoing traffic, and the less they send, the worse the ratio gets
> for networks sending data to Comcast.

True enough.  But its continued presence, *seven years* after it was
well-known to be a serious problem, tells us that Comcast either (a) can't
or (b) won't run its network properly.  So given this prima facie evidence
of either (a) systemic, chronic incompetence or (b) systemic, chronic
negligence, I think it's reasonable to wonder how many other aspects of
their operation are just as horribly broken, and what the impact of
that on their ability to carry steadily-increasing traffic might be.

---rsk




More information about the NANOG mailing list