osewalrus (osewalrus) wrote,

When Broadband Providers Fight, You Win? Maybe . . . .

Posting here whilst I resolve some technical difficulties over at PK and wanted to proof this.

Throughout 2006, while cable cos and telcos bitterly fought each other for regulatory advantage, they remained united on one thing: DEATH TO NETWORK NEUTRALITY. So I’m a little surprised that this letter from telco sock puppet Hands Off The Internet (HOTI) saying that the FCC has authority to act on our complaint and declaratory ruling against Comcast and urging the FCC to do so expeditiously got so little attention. Frankly, this should draw the same kind of attention as Rudy Guliani saying he would root for the Red Sox in the World Series.

But while my friends at Moveon have reacted to HOTI’s letter much the same way Red Sox fans in NH reacted to Guliani (i.e., “Who the Hell do you think YOU ARE to root for the Red Sox? Go back to NYC!”), this letter deserves somewhat more attention than a casual dismissal accompanied by the international gesture of respect (aka in geekland “count to four on your fingers in binary”). In fact, given the way Scott Cleland keeps blasting the complaint against Comcast, it would appear even the telco’s most shrill shill failed to get the memo.

So what’s going on here? As usual, I have but uninformed speculation and wild-ass guesses based on my limited observations. But here it goes.

First, lets review how we got here, with an emphasis on where telcos and cablecos actually differ. Because the technical differences and their accompanying economic ramifications drive policy, just as surely as policy drives economic and technical decisionmaking. (You can see a somewhat longer discussion of this, and why cable operators made their build out decision, on my Tales of the Sausage Factory blog.)

Back in the mid-1990s, when cable decided to get into the broadband biz, they made a number of decisions about how to build out their network. Critical to this, they opted to go for hybrid-fiber-coax and an ethernet-like system developed by CableLabs called “Docsis.” Building the network this way was much cheaper than moving to pure fiber, worked better with the delivery of video (which was still cable’s main business, after all), and allowed cable operators to maintain a much greater level of control (should they so wish) over subscribers.

But these decisions had a serious drawback. Everyone sharing the same cable head end shares the same capacity. If I am using all of it, then my neighbors get none of it. Of course, this doesn’t usually happen. When Comcast and the other cable companies opted for this architecture, people tended to use applications that connected to the internet, sent a burst of traffic, waited (so no traffic), downloaded a burst of traffic, then did not use the internet connection until the next burst of data. With this kind of use patterns, and particularly with applications tolerant of minor delays (like downloading static web pages), it was easy for cable operators to share capacity among their subscribers while claiming a very fast always on speed based on the average user load for the system.

Even in the old days, however, you could still have the system slow to a crawl at peak use times when the entire neighborhood got online. Cable companies kept making marginal upgrades to their systems to try to keep up with demand, but changing user patterns make this increasingly harder to do without making major expenditures. In the last several years, users have increasingly shifted to applications that expect that a 5 mbps “always on” pipe gives you 5 mbps all the time – like downloading huge files with BitTorrent and other p2p applications.

Telcos generally do not have the same kind of constraints. Yes, they have network capacity issues and at some point, a handful of users sucking down lots of data can degrade traffic for others on the network. But the tolerance of DSL for this sort of use is much higher than the tolerance of cable broadband. So already, telcos are in a somewhat better position with regard to the actual capacity of their networks (as opposed to the advertised capacity of their networks) to handle the changing user patterns.

But Verizon, which geographically overlaps heavily with Comcast, went one better. Back in 2003 or so, Verizon decided that users wanted really fast networks with oodles and squindoodles of capacity. So they bit the multibillion dollar bullet, took a beating on their stock for a few years, and started to build out actual fiber to the home in the form of FIOS. Lo and behold, the doubters were wrong. Fiber is better than hybrid-fiber-coax. Yes, even FIOS can get overwhelmed eventually. But whereas cable is chugging along desperately trying to keep up, huffing and puffing like some overweight coach potato forced to run a marathon, Verizon’s FIOS network is jogging along at a fine clip without breaking a sweat (except for the occasional problem with setting people’s houses on fire and their new annoying habit of monetizing domain name misspellings).

So now we come to today, where we have the consequences of these choices. Comcast, which directly competes with Verizon, would like to have a network as stable and fast as Verizon’s FIOS. But they don’t. Instead, they have a network that has serious problems keeping up with the way customers want to use it. Comcast could spend the money to upgrade to fiber the way Verizon did and actually have a network that works like Verizon’s FIOS system. But Comcast doesn’t want to do that either.

So Comcast has opted for the much cheaper solution. They will advertise that they have a network as good as FIOS, but they really don’t. This strategy is very cost efficient. Except that users trying to use their 5 or 10 mbps connection all the time, at maximum capacity, just like in the ads, find out real quick that Comcast’s network is not as good as FIOS. (Not that Verizon is trying to attract the 24-hr BitTorrent crowd, mind, but at least they handle the load better.)

So again, Comcast has a choice on how to manage its very real network constraint. It can come clean with subscribers and start setting realistic terms of service and advertising realistic speeds/capacity (for example by going back to metered pricing). Except that if they do, people will use FIOS instead of Comcast. It can spend billions of dollars to upgrade all the way to fiber. But that will cut into profit margins, take lots of time, and Comcast will still lose customers to FIOS. But those should be Comcast’s choices, because Verizon guessed right, Comcast guessed wrong, and Verizon therefore reaps the free market rewards of having the guts and the foresight to anticipate customer need, and forces its rivals to upgrade to fiber, and a happy ending for everyone just like the neo-cons promised, yes?

Well no, because Comcast went and invented a third option. Instead of spending lots of money to build a network that matches what they advertise, they spent a lot less money to go to a company called Sandvine, bought equipment to designed to “shape” user traffic without users finding out, and continued to lie through their teeth to their customers, even when directly asked. In fact, as this leaked memo shows, Comcast continued to instruct its line staff on how to lie, obfuscate and generally mislead customers who called to ask point blank whether the reports in the AP about shaping BitTorrent traffic were true.

Looked at this way, you can see why the telcos (and therefore their sock-puppet HOTI) would be a shade peeved about Comcast’s decision to “manage” their network in such a “cost effective” but deceptive manner. Because if Comcast can keep pretending its network is just as good as FIOS when it isn’t, and can even lie when asked about it directly by customers, then Verizon just wasted a couple of bazillion dollars and took a two-year stock beating for NOTHING. Because while everyone reading this blog knows the score on Comcast these days, the vast majority of average customers who spend much more time on blogs relating to Paris Hilton or the latest appearance of TomKat in public than they do at blogs relating to traffic shapping have no clue what is going on. And if they did get some vague notion that something was happening from the stories in the newspapers when the AP broke the story, they called Comcast and got told there s nothing to it. Because it can’t be legal for Comcast to just lie to their customers about this, right?

[At this point, I usually get interrupted by some defender of Comcast repeating that this is all really technical and that because Comcast has a real bandwidth constraint, that makes everything OK. Sadly, even the usually savvy Ed Felten seems to have fallen into this trap, for whom a network with constraints is apparently like a kitten in a tree. The usual NN “dittoheads” then start the “you guys don’t know technology” chorus. None of which, of course, actually addresses the minor point that Comcast is telling their customers big fat lies to keep them from finding out how badly Comcast’s network actually sucks. And given that Comcast increasingly charges early termination fees (which amusingly enough, they also lied about – at least they’re consistent), I don’t think they should lie about their terms of service and thus sucker punch folks into paying an early termination fee to get out. And all the fancy technical explanation in the world cannot change that.]

Anyway, back to the HOTI letter. So while the telcos still want to shape and manage traffic, they are probably quite happy to have Comcast slapped around for telling lies to their customers that would keep those customers from switching to FIOS or even DSL. Hence the HOTI letter. Because the FCC can focus on the bad part of what Comcast did (lying to their customers) without touching the question of what constitutes “reasonable network management.”

[At this point, the Comcast defenders again repeat their technical explanations – kind of like those people who assume that if you speak English slowly enough and loudly enough non-English speaking people can understand you. They also throw in at this point that Comcast needs to keep this a secret, because otherwise their customers will circumvent the traffic shaping. They then usually accompany this with some more name calling about how ignorant I am about network management.

Two problems with this revised “technical explanation.” First, the NY Attorney General recently told Verizon Wireless that you cannot protect the fact that you have a crappy network by lying to your customers. Verizon Wireless – which operates under some truly Hellacious physical constraints – advertised an “unlimited” wireless broadband access plan. But, because they have Hellacious physical constraints, they actually would kick people off if they exceeded a 5 GB cap. So, as required by the NY AG, Verizon Wireless changed their terms of service to say “We impose a 5 GB/month cap. If you exceed this, we will kick you off.

And ya know what? Verizon Wireless did not suddenly suffer horrible security messiness. Sure, they may have some customers who decided that it wasn’t worth paying for a service with a bandwidth cap. But the fact that Verizon Wireless was forced to ‘fess up when it didn’t want to and may therefore have lost a few customers did not make this a “technical issue” or a “security issue.” It made it a “disclosure issue.” I.e., you have to tell you customers the truth about their terms of service no matter how much you would like to pretend you have a better network. And if Verizon Wireless can survive disclosing their actual terms of service, I don’t see why Comcast can’t.

The other problem with the “security” argument is that it boils down to “we want to lie to our customers about how good our network is. So to protect that, we need to mess with their traffic and lie to them about messing with their traffic. Because if we don’t lie to them about how we mess with their traffic, then they will circumvent our messing with their traffic so that we can no longer pretend we have a better network than we have.” While that may satisfy Messrs Ou, Bennet, & Co., I remain unpersuaded. Either advertise your actual terms of service, or bring back metered pricing or build a goddamn better network like Verizon did. But stop pretending that a technical limitation automatically translates into a right to lie through your teeth to your customers when they ask you about it point blank. You built a crappy network, now deal with the consequences – without lying to your customers.]

But getting back to our friends at HOTI. As I said, their patrons at the phone companies certainly have good reason to be annoyed at Comcast. Because if Comcast can claim to have a network as good as the telco networks without actually spending the money to make that true, then the telcos just dropped a couple of billion in investment for no good reasons. Oh, and for those of us who care about public policy outcomes, that means that no one in this country will ever invest in fiber to the home again, because it will always be cheaper to lie and “manage traffic” than to build fiber. Which, from the perspective of our avowed national broadband policy, is a rather sub-optimal result.

But in addition, their telco backers are also far more politically sophisticated than the cable chorous of Ou & Co. Because while Ou & Co. apparently want this to go away, HOTI would like to prove to the world that there is no need for legislation by having the FCC resolve this. From the telco (and therefore HOTI’s) perspective, the best result is for the FCC to say “we have authority to address the issue, and we hold that lying to your customers is not an acceptable form of network management.” Then the telcos (and HOTI) can run back to the various Presidential candidates that have supported net neutrality legislation and Congress and say “See, the FCC can deal with these problems. You don’t need to pass new legislation....we can go about our business...these aren’t the driods you’re looking for...move along...move along....” And the Presidential candiates, mesmerized by an old PAC mind trick, will repeat “we don’t need to pass legislation....you can go about our business....”

Needless to say, I do not think that the FCC addressing one well publicized case after the fact is proof that we don’t need net neutrality legislation. Especially as I still want to ban the kind of traffic tiering where third parties can pay to override user preferences (what I call “Whitacre Tiering”) for reasons I explain here, here, and here. So while I am hoping the evil PAC mind control trick will not work, I am still glad to have HOTI acknowledge that the FCC has jurisdiction and call for a thorough investigation. Because I find the idea that lying to your customers is an acceptable form of network management fairly unacceptable. Although for Messrs Ou & Co. and others with no such concerns, I am happy to offer a special deal on my 10 Gigabit per second network. Just be aware that it has a $1,000 early termination fee, I can raise monthly charges whenever I want, and I’m not actually saying I’ll give you 10 Gigabits per second. Well, I am saying it. But I don’t mean it. Since I may not be able to give you that speed for technical reasons. But don’t worry, if you discover you’re unhappy, you can always switch. So long as you don’t mind paying the $1000 early termination fee. Which, I’m sure you’ll agree, is also a technical issue, in that I technically don’t want you to switch.
  • Post a new comment


    Anonymous comments are disabled in this journal

    default userpic

    Your IP address will be recorded 

  • 1 comment