If a customer is using yoru product more that you initially thought, great, sell them more of it, they are hooked. Don’t limit what they want. I find Comcast’s practice to be both immoral and bad for business.

First, he mentioned that Comcast sets traffic limits - this is true. However they do NOT publish these limits. There have been many reports of Comcast’s heavy users getting warnings and terminations if they violate these bandwidth caps, but I have yet to see Comcast PUBLISH the caps. Right now best estimates are on the order of 100GB/month. But that’s an estimate based upon reports of users who got cut off. This is yet another area where Comcast is hiding behind FUD, instead of being up-front with their customers. And there is no “traffic management” reason not to tell us the caps - unlike traffic obfuscation, there’s no gaming a byte-count. This is solely a marketing move, so they can tell everyone about their great “unlimited internet” without disclosing the gotcha’s

Second, the issue of “firewalls use TCP resets” is a misdirection. Firewalls don’t use resets to “shape traffic”, they use it to block access by bad actors and unauthorized protocols. A TCP reset isn’t a valid traffic shaping tool, because the real effect is to cause either a total termination of the communication or a storm of retries if the software has any recovery logic.

“Traffic Shaping” is usually used to limit the bandwidth used by a protocol, and while completely blocking the traffic can count as “limiting the bandwidth” (to zero), the term is usually used to describe rate limiting, not blocking access. If Comcast doesn’t want it’s users to use more than 10K of their upstream for BitTorrent, then there are tools that can do that. But TCP Resets don’t limit rate, they limit access. And delaying a connection for 10 minutes is pretty much the equivalent of a total block on access, especially in the instant gratification world of “internet time”.

In one of his comment responses, Mr. Bennett mentions that he has no trouble downloading torrents via Comcast, but he totally ignores the fact that BitTorrent is a distributed peer-to-peer protocol. If the other ISP’s followed Comcast’s example, he would NOT be able to download.

If the people who have already downloaded the torrent cannot also host the file, then the system breaks down. The whole point is to distribute the upstream load so that no one host delivers the bulk of the data, and thus the load on any one server/network/ISP isn’t as heavy as if only one system hosted the file.

The amazing fact is that he doesn’t seem to recognize that this system is beneficial to the ISP, compared to the single-host solution (FTP, etc.). This is the kind of misdirection that needed to be addressed in the podcast, and I’m amazed that the other participants chose not to do so.

When the AP story broke, I was having major surgery. I’m still recovering. I’m amazed at how big this story has become.

I found myself agreeing with Ed Felton the most. With respect, I felt that Richard Bennett was the most off-course on his opinions and his facts.

As I repeatedly say in the DSLReports message, I personally believe that reasonably managing the passive activity on the network during periods of congestion is okay.

One thing I’d like everyone to know, a development discovered after my first message, which is being missed: My testing shows Comcast is interfering with 40% of my completed connections — regardless of the time of day, day of week, and presumably the corresponding differences in network load. Furthermore, because these Sandvine devices are installed at each CMTS, they even interfere with connections that never route outside of the Comcast network!

Richard: the SlashDot uber-geeks are incorrect. Most firewalls never use RST on an established connection. Sally Floyd wrote in the fw-wiz mailing list on 09 May 2001 that only 300 sites out of 24000 tested seemingly enforced policy using RST. Most use some other kind of control, several of which she describes in RFC 3360.

As employed, Comcast’s P2P management is both damaging and ineffective. It’s a perfect example of why their silence on this issue is a Bad Thing(tm).

The question, as I argued at the FTC last February, is what do we do in an uncertain world? Given the known variables, including the natural incentive for profit maximizing firms to maximize profit at the expense of users (if not disciplined by the market), I come down on the need for prophylactic regulation.

Reasonable minds can differ, in the face of the unknown, on the proper course. But the notion that doing nothing is the preference in the absence of certainty is, to me, a bias unsupported by historical evidence. Failure to take action can be as determinative and as much of a disaster as taking action when patience is required.

If I am right, then waiting for disaster to become clear means it will be too late to salvage or communications platform which is critical to our economy and our ability to govern ourselves in a democratic society. Of course, the counter argument is that, if I am wrong, then regulation may throttle the development of the internet and bring the disaster I fear. Ain’t public policy a lark! But it should, I would hope, instill in all of us a bit of humility and a willingeness to debate the issue.

Like you, I’ve been applauding Verizon for their investment in fiber, and encouraging people to reward them by subscribing (and doing it on my blog since 2003.) I want to see them succeed, and have no love or commercial interest in Comcast. I do have a basic sense of fairness, however, so I’m annoyed when advocates of regulation for their own sake misrepresent engineering practices. As a Comcast customer, I want them to engage in traffic shaping that improves my interactive experience, and I’ve had no problems downloading files with BitTorrent. My experience mirrors that of millions of others who haven’t noticed any ill effects from traffic management that’s been in effect since March.

Are you sure there’s a problem that needs to be addressed with net neutrality regulations? It appears like a “tree falls in the forest but nobody hears it” situation to me.

Granted I’m rooting for Verizon, because only if the market rewards Verizon will we ever see fiber to the home deployed. But it won’t happen if customers can’t adequate judge the package.

Incidentally, I Skyped over a Comcast link into the podcast today. It seemed to work OK, but I wasn’t running BitTorrent at the time.

However, Richard Bennett made a good counter-argument on the show: The more specific a broadband operator gets about what service is being “managed” or how it is being managed, the more likely it is that aggressive users will seek to evade those management efforts.

I think that’s a fair point, although I still think more could be done to clarify TOS agreements regarding traffic management techniques.

Comcast added fewer Internet and video customers than expected in its third quarter.

The rub: The 450,000 Internet access subscribers were down from 538,000 a year ago. Morgan Stanley was projecting Internet additions of 533,000 in the quarter. Comcast’s video additions were roughly in line with expectations. Comcast voice subscribers were 662,000 in the quarter, below Merrill Lynch’s estimate of 725,000.

Read the whole thing. By any account, the market is working:

On the Verizon front, the company is likely to note big growth in FiOS TV and Internet subscribers. Verizon reports its third quarter results on Monday. Morgan Stanley is expecting Verizon to post solid gains on FiOS and Verizon Wireless.

Oops, no need for regulators.

Even under the best of explanations, Comcast is managing its traffic in a way that renders a particular application (BitTorrent) far less usable than Comcast has actively led customers to believe. Verizon has invested billions to upgrade to fiber on a theory that the majority of customers want access to high-bandwidth services like BitTorrent. Comcast has bet that customers will tolerate certain forms of traffic management. In an efficient market, the provider that bet right gets rewarded with more customers. Either Verizon starts attracting enough business away from Comcast that Comcast changes its architecture, or Verizon stops its expensive upgrades to fiber. Or the market provides enough business for both.

But allowing Comcast to use “security” as an excuse to actively mislead customers by refusing — even when explicitly asked — to state clearly that an application such as BitTorrent might be targeted effectively nullifies Verizon’s multi-billion dollar investment. How will the market ever achieve an efficient price, or the other pro-consumer benefits of an unregulated market, if consumers cannot properly reward or punish providers?