Welcome Brett Glass

The following post is from our new co-blogger, Brett Glass. Brett and I first crossed paths when we were working on the “Skywalker” token-ring project at Texas Instruments in the early 80s. Brett was part of the team in Houston doing the chipset, and I worked on a team on Austin doing a terminal server application for it. We both spoke at an ITIF event in Washington, DC, last spring on network management. He’s been a valuable commenter here for a while, and I’m very happy to have him contributing posts as well. Here’s his bio: Continue reading “Welcome Brett Glass”

Professional Complainers Blast Cox

Cox Cable announced plans to test a new traffic management system intended to improve the Internet experience of most of their customers yesterday, and the reaction from the network neutrality lobby came fast and furious. The system will separate latency-sensitive traffic from bulk data transfers and adjust priorities appropriately, which is the sort of thing that Internet fans should cheer. In its essence, the Internet is a resource contention system that should, in most cases, resolve competing demands for bandwidth in favor of customer perception and experience. When I testified at the FCC’s first hearing on network management practices last February, I spent half my time on this point and all other witnesses agreed with me: applications have diverse needs, and the network should do its best to meet all of them. That’s what we expect from a “multi-purpose network”, after all.

So now that Cox wants to raise the priority of VoIP and gaming traffic over background file transfers, everybody should be happy. The neutralists have always said in public fora that they support boosting VoIP’s priority over P2P, and Kevin Martin’s press release about the Comcast order said he was OK with special treatment for VoIP. And in fact the failure of the new Comcast system to provide such special treatment is at the root of the FCC’s recent investigation of Comcast, which was praised by the neuts.

So how is it that the very people who complain about Comcast’s failure to boost VoIP priority are now complaining about Cox? Free Press’s general-purpose gadfly Ben Scott is practically jumping up and down pounding the table over it:

Consumer advocates certainly aren’t impressed. “The information provided by Cox gives little indication about how its new practices will impact Internet users, or if they comply with the FCC’s Internet Policy Statement,” says consumer advocacy firm Free Press in a statement. “As a general rule, we’re concerned about any cable or phone company picking winners and losers online.”

“Picking winners and losers” is bad, and failing to pick winners and losers is also bad. The only thread of consistency in the complaints against cable, DSL, and FTTH providers is a lack of consistency.

Make up your mind, Ben Scott, do you want an Internet in which Vuze can step all over Skype or don’t you?

UPDATE: For a little back-and-forth, see Cade Metz’ article on this The Register, the world’s finest tech site. Cade quotes EFF’s Peter Eckersley to the effect that Cox is “presuming to know what users want.” They are, but it’s not that hard to figure out that VoIP users want good-quality phone calls: a three-year-old knows that much.

Technorati Tags:

What recession?

So here’s your recession-proof business, ladies and gentlemen:

Netflix, the company which mails out DVD rentals and also offers streamed programming via the internet, saw a 45% jump in profits and 26% rise in consumers to 9.4 million in the fourth quarter.

This was the quarter in which Netflix released Watch Instantly on non-PC platforms. It’s so ubiquitous now I have it on three platforms: a home theater PC, TivoHD, and a Samsung BD-P2500 Blu-Ray player. It looks best on the Samsung, thanks to its HQV video enhancement chip.

Internet Myths

Among my missions in this life is the chore of explaining networking in general and the Internet in particular to policy makers and other citizens who don’t build network technology for a living. This is enjoyable because it combines so many of the things that make me feel good: gadgetry, technology, public policy, writing, talking, and education. It’s not easy, of course, because there are a lot of things to know and many ways to frame the issues. But it’s possible to simplify the subject matter in a way that doesn’t do too much violence to the truth.

As I see it, the Internet is different from the other networks that we’re accustomed to in a couple of important ways: for one, it allows a machine to connect simultaneously to a number of other machines. This is useful for web surfing, because it makes it possible to build a web page that draws information from other sources. So a blog can reference pictures, video streams, and even text from around the Internet and put it in one place where it can be updated in more-or-less real time. It enables aggregation, in other words. Another thing that’s unique about the Internet is that the underlying transport system can deliver information at very high speed for short periods of time. The connection between a machine and the Internet’s infrastructure is idle most of the time, but when it’s active it can get its information transferred very, very quickly. This is a big contrast to the telephone network, where information is constrained by call setup delays and a very narrow pipe.
Continue reading “Internet Myths”

Damned if you do, screwed if you don’t

The FCC has finally noticed that reducing the Quality of Service of an Internet access service affects all the applications that use it, including VoIP. They’ve sent a harsh letter to Comcast seeking ammunition with which to pillory the cable giant, in one of Kevin Martin’s parting shots:

Does Comcast give its own Internet phone service special treatment compared to VoIP competitors who use the ISP’s network? That’s basically the question that the Federal Communications Commission posed in a letter sent to the cable giant on Sunday. The agency has asked Comcast to provide “a detailed justification for Comcast’s disparate treatment of its own VoIP service as compared to that offered by other VoIP providers on its network.” The latest knock on the door comes from FCC Wireline Bureau Chief Dana Shaffer and agency General Counsel Matthew Berry.

Readers of this blog will remember that I raised this issue with the “protocol-agnostic” management scheme Comcast adopted in order to comply with the FCC’s over-reaction to the former application-aware scheme, which prevented P2P from over-consuming bandwidth needed by more latency-sensitive applications. My argument is that network management needs to operate in two stages, one that allocates bandwidth fairly among users, and a second that allocates it sensibly among the applications in use by each user. The old Comcast scheme did one part of this, and the new scheme does the other part. I’d like to see both at the same time, but it’s not at all clear that the FCC will allow that. So we’re left with various forms of compromise.

The fundamental error that the FCC is making in this instance is incorrectly identifying the “service” that it seeks to regulate according to a new attempt to regulate services (skip to 13:30) rather than technologies.

Comcast sells Internet service, telephone service, and TV service. It doesn’t sell “VoIP service” so there’s no basis to this complaint. The Commission has made it very difficult for Comcast to even identify applications running over the Internet service, and the Net Neuts have typically insisted it refrain from even trying to do so; recall David Reed’s fanatical envelope-waving exercise at the Harvard hearing last year.

The telephone service that Comcast and the telephone companies sell uses dedicated bandwidth, while the over-the-top VoIP service that Vonage and Skype offer uses shared bandwidth. I certainly hope that native phone service outperforms ad hoc VoIP; I pay good money to ensure that it does.

This action says a lot about what’s wrong with the FCC. Regardless of the regulatory model it brings to broadband, it lacks the technical expertise to apply it correctly. The result is “damned if you do, damned if you don’t” enforcement actions.

This is just plain silly. The only party the FCC has any right to take to task in this matter is itself.

The pirates who congregate at DSL Reports are in a big tizzy over this, naturally.

Keeping the Issue Alive

Friends of Broadband should be pleased with President-elect Obama’s proposed broadband stimulus program, which proposes $6 billion in grants for wireless and other forms of broadband infrastructure. Granted, the package isn’t as large as many had wished; Educause had asked for $32 billion, ITIF wanted $10 billion, and Free Press wanted $40 billion, but this is a good start. Harold Feld puts the size of the grant package in perspective and praises it on his Tales of the Sausage Factory blog.

But there’s no pleasing some people. Free Press has mounted an Action Alert, asking its friends to oppose the stimulus package as it currently stands. The Freeps, who run the “Save the Internet” campaign, want strings attached to the money, insisting it only be given to projects that meet their requirements:

1. Universal: focused on connecting the nearly half of the country stuck on the wrong side of the digital divide.
2. Open: committed to free speech and without corporate gatekeepers, filters or discrimination.
3. Affordable: providing faster speeds at lower prices.
4. Innovative: dedicated to new projects only and available to new competitors, including municipalities and nonprofits.
5. Accountable: open to public scrutiny so we can ensure that our money isn’t being spent to prop up stock prices and support market monopolies.

These goals are not even consistent with each other. Half of America uses broadband today, and half doesn’t. Most of the unconnected half have chosen not to subscribe to services that reach their homes already, opting to remain outside the broadband revolution for their own reasons. So we can’t very well pursue numbers 1 and 4 at the same time. Most of this money will be spent in rural areas that are currently served by Wireless ISPs like Lariat. Rural population isn’t as large as urban population, so going into unserved or underserved areas isn’t going to do much for the digital divide-by-choice that plagues America’s inner cities.

I suspect there’s some self-interest involved here, such that Free Press wants to keep the issue of America’s place in the global ranking of broadband penetration about where it is (between 7th and 15th, depending on whose numbers you like) in order to raise money, have a soapbox, and keep on complaining.

I don’t see any other way to explain this.

UPDATE: Freep has sent letters to the committee chairs with much less incendiary language, but arguing the same line: the Internet is a telecom network and has to be regulated the way that telecom networks have always been regulated. This angle is clearly good if you’re a career telecom regulator, but it’s blind to the technical realities of IP network management. Making an IP network fair and functional requires “discrimination”, and the Freep doesn’t get that. Not even a little bit.

This organization has established an amazing ability to confuse its self-interest with the public interest in the short time that it’s been around. Freep’s first issue, after all, was a series of regulations designed to prevent the rapacious newspaper industry from taking over the television industry. They still push for limits on TV and newspaper cross-ownership, and only got into the Internet-as-telephone fight to advance their initial cause. The number of people who think free societies need to be protected from “powerful newspapers” is vanishingly small, or course, around the same size as the flat-earther demographic.

UPDATE 2: It gets even stranger. Open access provisions are already in the bill, as Matthew Lasar points out on the Ars Technica blog:

As for the net neutrality and open access ideas; well, they’re already in the bill (PDF; see p. 53). NTIA, the executive branch agency tasked with disbursing the broadband money, is required to ensure that all grant recipients operate both wired and wireless services on an “open access basis,” though it’s left up to NTIA to define what this means and how it works.

In addition, anyone taking grant money must “adhere to the principles contained in the Federal Communications Commission’s broadband policy statement,” which lays out four basic neutrality provisions for Internet companies. In other words, although “network neutrality” isn’t mentioned, it’s already in the bill in a basic way. (Note that the FCC policy statement only protects “legal content,” however; it’s not a pure “end-to-end” packet delivery guarantee.)

Here’s a suggestion for the Freep: before issuing your next mouth-breathing Action Alert about a pending bill, read the damn thing. You won’t look like such a bunch of knee-jerk alarmists if you do.

Technorati Tags:

Briscoe explains Re-ECN in plain English

See the current issue of IEEE Spectrum for a nice description of Bob Briscoe’s Re-ECN, A Fairer, Faster Internet Protocol:

Refeedback introduces a second type of packet marking—think of these as credits and the original [ECN] congestion markings as debits. The sender must add sufficient credits to packets entering the network to cover the debit marks that are introduced as packets squeeze through congested Internet pipes. If any subsequent network node detects insufficient credits relative to debits, it can discard packets from the offending stream.

To keep out of such trouble, every time the receiver gets a congestion (debit) mark, it returns feedback to the sender. Then the sender marks the next packet with a credit. This reinserted feedback, or refeedback, can then be used at the entrance to the Internet to limit congestion—you do have to reveal everything that may be used as evidence against you.

Refeedback sticks to the Internet principle that the computers on the edge of the network detect and manage congestion. But it enables the middle of the network to punish them for providing misinformation.

The limits and checks on congestion at the borders of the Internet are trivial for a network operator to add. Otherwise, the refeedback scheme does not require that any new code be added to the network’s equipment; all it needs is that standard congestion notification be turned on. But packets need somewhere to carry the second mark in the “IP” part of the TCP/IP formula. Fortuitously, this mark can be made, because there is one last unused bit in the header of every IP packet.

This is a plan that will allow interactive uses of the Internet to co-exist happily with bulk data transfer. It’s quite brilliant and I recommend it as an alternative to a lot of nonsense that’s been floated around this space.

Technorati Tags: ,