Cerf on managing networks & the need for industry discussion

by on August 4, 2008 · 25 comments

Google’s Chief Internet Evangelist Vint Cerf, one of the fathers of the Net, has a very thoughtful post up on the Google Public Policy Blog today asking “What’s a Reasonable Approach for Managing Broadband Networks?” He runs through a variety of theoretical approaches to network load management. There’s much there to ponder, but I just wanted to comment briefly on the very last thing he says in the piece:

Over the past few months, I have been talking with engineers at Comcast about some of these network management issues. I’ve been pleased so far with the tone and substance of these conversations, which have helped me to better understand the underlying motivation and rationale for the network management decisions facing Comcast, and the unique characteristics of cable broadband architecture. And as we said a few weeks ago, their commitment to a protocol-agnostic approach to network management is a step in the right direction.

I found this of great interest because for the last few months I have been wondering: (a) why isn’t there more of that sort of inter- and intra-industry dialogue going on, and (b) what could be done to encourage more of it? With the exception of those folks at the extreme fringe of the Net neutrality movement, most rational people involved in this debate accept the fact that there will be legitimate network management issues that industry must deal with from time to time. So, how can we get people in industry — from all quarters of it — to sit down at a negotiating table and hammer things out voluntarily before calling in the regulators to impose ham-handed, inflexible solutions? What we are talking about here is the need for a technical dispute resolution process that doesn’t involve the FCC.

If the anti-Net neutrality regulation crowd (and that includes me!) wants to be taken seriously when they talk about “self-regulatory” solutions, this sort of dispute resolution process becomes essential. And the pro-Net neutrality regulation crowd needs to understand that, even if they ultimately desire some role for the FCC here, regulatory resolutions to technical disputes are notoriously slow and ultimately will always be one step behind the technical dispute du jour.

Therefore, wouldn’t it be nice if, as Cerf suggests above, those parties with a technical dispute about network management had a way of talking things through immediately and before they went to the regulatory equivalent of mutually assured destruction?

All the relevant players in the broadband / Internet sector need to put their heads together and think about how to create a forum or process that can serve as such a technical dispute resolution mechanism. On a smaller scale, Comcast and Bit Torrent did this in a voluntary, bilateral fashion when they sat down to hammer out a collaborative agreement in March. As their press announcement noted:

Comcast Corporation and BitTorrent, Inc. announced today that they will undertake a collaborative effort with one another and with the broader Internet and ISP community to more effectively address issues associated with rich media content and network capacity management. While BitTorrent and Comcast are talking directly, they are also in discussions with other parties to help facilitate a broader dialogue and cooperation across industries.

But we know that countless more technical disputes will arise in the future at every layer of the Internet — not just with Comcast and BitTorrent. Thus, if we are really going to achieve “a broader dialogue and cooperation across industries” then what we really need is the equivalent of a multilateral trade negotiating process or forum to achieve sensible resolutions to complex technical difficulties surround Internet network management.

I am not prepared to say whether a new, formal organization is needed to accomplish this or if existing institutions and individuals (academic, trade associations, etc) might be able to work together to make this happen. For example, and I am just thinking out loud here so don’t quote me on this, what if we had the Internet Society working in conjunction with several major industry trade associations and some respected academic institutions to form some sort of collaborative, dialogue-oriented dispute resolution process? Sort of GATT or WTO for technical Internet dispute resolution.

Certainly that would be preferable to a politicized FCC taking over the show and making all these technical decisions, no? I’d be interested in hearing some input from others.

  • Jamalystic

    This is so true!! People at the extreme end of this net neutrality are seeing everything about the broadband providers as evil and this to me is a fundamental flaw in their argument. WE should know that the carriers have legitimate reasons for doing what they are doing now as Andr Keen pointed out:The Net Neutrality Fearmongers Are At It Again(http://www.internetevolution.com/author.asp?section_id=556&doc_id=160628&F_src=flftwo)

  • Jamalystic

    This is so true!! People at the extreme end of this net neutrality are seeing everything about the broadband providers as evil and this to me is a fundamental flaw in their argument. WE should know that the carriers have legitimate reasons for doing what they are doing now as Andr Keen pointed out:The Net Neutrality Fearmongers Are At It Again(http://www.internetevolution.com/author.asp?section_id=556&doc_id=160628&F_src=flftwo)

  • http://bennett.com/blog Richard Bennett

    Clearly we do need more collaboration on the part of industry, and the FCC needs a technical advisory panel convened by a Chief Technologist. We’re relying on the political process to solve technical problems, and that never works.

    Sadly, the forum that should have dealt with this issue – IETF – was very slow off the mark, only now holding a discussion that should have started about 5 years ago.

  • http://blog.actonline.org Mark Blafkin

    Amen. These things are nearly always better handled within the industry rather than by bureaucrats. I need to give more thought as to what kind of body (ad hoc or institutional) would best be able to manage these debates however. The Internet Society could be a good lead. There are a lot of fabulous people over at the Internet Society, like my friend Frederic Donck, who has a lot of experience as negotiator in technology policy areas.

    On another note: I’m relative outsider to the NN debate, so this may be a dumb question. Has anyone discussed botnets in the context of NN?

    If ISPs are prohibited from throttling any specific application or type of application, would that unnecessarily limit the ability of ISPs to play an important role in responding to DDOS attacks? Do people think that the current legislative proposals have loopholes big enough to allow ISPs to react to bot net attacks when they see some of the infected machines on their own network? Could a black hat use existing P2P or other technologies to launch an attack that an ISP would be legally bound NOT to stop?

  • http://bennett.com/blog Richard Bennett

    Clearly we do need more collaboration on the part of industry, and the FCC needs a technical advisory panel convened by a Chief Technologist. We’re relying on the political process to solve technical problems, and that never works.

    Sadly, the forum that should have dealt with this issue – IETF – was very slow off the mark, only now holding a discussion that should have started about 5 years ago.

  • http://blog.actonline.org Mark Blafkin

    Amen. These things are nearly always better handled within the industry rather than by bureaucrats. I need to give more thought as to what kind of body (ad hoc or institutional) would best be able to manage these debates however. The Internet Society could be a good lead. There are a lot of fabulous people over at the Internet Society, like my friend Frederic Donck, who has a lot of experience as negotiator in technology policy areas.

    On another note: I’m relative outsider to the NN debate, so this may be a dumb question. Has anyone discussed botnets in the context of NN?

    If ISPs are prohibited from throttling any specific application or type of application, would that unnecessarily limit the ability of ISPs to play an important role in responding to DDOS attacks? Do people think that the current legislative proposals have loopholes big enough to allow ISPs to react to bot net attacks when they see some of the infected machines on their own network? Could a black hat use existing P2P or other technologies to launch an attack that an ISP would be legally bound NOT to stop?

  • http://bennett.com/blog Richard Bennett

    Mark – re: botnets, the “network management exception” actually originated because of the realization that DDOS attacks and spamming had to be managed.

  • http://felter.org/wesley/ Wes Felter

    While filtering outbound spam and bot traffic would undoubtedly benefit the Internet, cutting the customers off might give them more incentive to clean things up. :-)

    Botnets also provide a good counter-argument for metered service, since customers will demand refunds for bot traffic, but unless they have detailed traffic accounting (DPI!), ISPs will just have to refund all the overage fees or cancel the customer.

  • http://enigmafoundry.wordpress.com/2008/05/25/ministry-of-truth-at-the-tlf/ enigma_foundry

    One of the points that needs to be made here is that much of the uproar against Comcast has been a result of their own doing–especially about them not being transparent and clear to their customers about exactly what they are doing to manage their network.

    Rather than explicitly mandate net neutrality, I’d proposed that we mandate transparency and full disclosure, on the part of the ISP’s.

    The threat of NN legislation will probably keep the networks more or less neutral–but the requirement to disclose how the networks are being managed will prevent 90% of the abuse.

    It will, of course be argued that this mandate of transparency is overburdensome–but given it is infrastructure we are talking about here–and infrastructure that is the place where much public discourse occurs–I think it’s safe to say that burden is more than justified by the public policy goals of fostering free speech and open and unfettered debate and dissemination of information.

  • http://enigmafoundry.wordpress.com/2008/05/25/ministry-of-truth-at-the-tlf/ enigma_foundry

    On a smaller scale, Comcast and Bit Torrent did this in a voluntary, bilateral fashion when they sat down to hammer out a collaborative agreement in March. As their press announcement noted:

    Comcast Corporation and BitTorrent, Inc. announced today that they will undertake a collaborative effort with one another and with the broader Internet and ISP community to more effectively address issues associated with rich media content and network capacity management. While BitTorrent and Comcast are talking directly, they are also in discussions with other parties to help facilitate a broader dialogue and cooperation across industries.

    Please. Who believes that comcast voluntarily sat down with Bit-torrent? They did so AFTER they realized they were in trouble for their network management policies. It’s a system of checks and balances, and in the comcast case it worked. Part of the reason it worked was the looming threat of NN legislation; nothing should be done which would preclude that threat.

  • http://bennett.com/blog Richard Bennett

    Mark – re: botnets, the “network management exception” actually originated because of the realization that DDOS attacks and spamming had to be managed.

  • http://felter.org/wesley/ Wes Felter

    While filtering outbound spam and bot traffic would undoubtedly benefit the Internet, cutting the customers off might give them more incentive to clean things up. :-)

    Botnets also provide a good counter-argument for metered service, since customers will demand refunds for bot traffic, but unless they have detailed traffic accounting (DPI!), ISPs will just have to refund all the overage fees or cancel the customer.

  • http://bennett.com/blog Richard Bennett

    Don’t ignore the power of bad press to change behavior, anonymous one.

  • http://enigmafoundry.wordpress.com/2008/05/25/ministry-of-truth-at-the-tlf/ enigma_foundry

    One of the points that needs to be made here is that much of the uproar against Comcast has been a result of their own doing–especially about them not being transparent and clear to their customers about exactly what they are doing to manage their network.

    Rather than explicitly mandate net neutrality, I’d proposed that we mandate transparency and full disclosure, on the part of the ISP’s.

    The threat of NN legislation will probably keep the networks more or less neutral–but the requirement to disclose how the networks are being managed will prevent 90% of the abuse.

    It will, of course be argued that this mandate of transparency is overburdensome–but given it is infrastructure we are talking about here–and infrastructure that is the place where much public discourse occurs–I think it’s safe to say that burden is more than justified by the public policy goals of fostering free speech and open and unfettered debate and dissemination of information.

  • http://enigmafoundry.wordpress.com/2008/05/25/ministry-of-truth-at-the-tlf/ enigma_foundry

    On a smaller scale, Comcast and Bit Torrent did this in a voluntary, bilateral fashion when they sat down to hammer out a collaborative agreement in March. As their press announcement noted:

    Comcast Corporation and BitTorrent, Inc. announced today that they will undertake a collaborative effort with one another and with the broader Internet and ISP community to more effectively address issues associated with rich media content and network capacity management. While BitTorrent and Comcast are talking directly, they are also in discussions with other parties to help facilitate a broader dialogue and cooperation across industries.

    Please. Who believes that comcast voluntarily sat down with Bit-torrent? They did so AFTER they realized they were in trouble for their network management policies. It’s a system of checks and balances, and in the comcast case it worked. Part of the reason it worked was the looming threat of NN legislation; nothing should be done which would preclude that threat.

  • http://bennett.com/blog Richard Bennett

    Don’t ignore the power of bad press to change behavior, anonymous one.

  • http://pobox.com/~adamm Adam Marcus

    As Vint says, “the real question for today’s broadband networks is not whether they need to be managed, but rather how.” In a sense, managing the Internet is like managing highways. As more highways (and more lanes) are built, people move further from where they work and commute for longer distances. It’s called induced demand. Similarly, as the supply of bandwidth increases, the demand for bandwidth increases. See here and here. As Berin Szoka put it, “trying to solve network congestion problems simply by increasing the amount of bandwidth available is like a pie-eating contest where the prize is… more pie.” Just as full-screen standard-definition video over the Internet is becoming a reality, everyone starts clamoring for high-definition video over the Internet. So at some point (whether because the network isn’t built fast enough, a natural disaster makes everyone pick up their phone, or Victoria’s Secret does an online lingerie show), traffic congestion will force routers to delay or drop packets. When that happens, following a simple “first in, first out” approach just doesn’t make sense. Just as we want ambulances and fire engines to be given priority on our roadways, most people would rather their VoIP conversations get through than someone’s file download finishes a few seconds or minutes faster.

    Vint suggests that “Internet traffic should be managed with an eye towards applications and protocols.” In a sense, Comcast was doing exactly that by limiting BitTorrent traffic. BitTorrent is a specific application with a specific protocol. It is not time-sensitive like VoIP so it makes sense to limit BitTorrent traffic first when network congestion becomes a problem. But Comcast (according to the FCC’s press release) was limiting BitTorrent traffic at all hours of the day and in all sorts of areas, presumably even in areas where network congestion wasn’t a problem. Vint clarifies his position: “such prioritization should be applied across the board to all low latency traffic, not just particular application providers.” So managing a specific application is bad, but prioritization in general is good. What’s needed is an agnostic way to prioritize packets.

    There *is* such an agnostic method for controlling traffic. Every TCP/IP packet can be flagged with a certain priority level, and routers can be programmed to handle packets differently depending on their priority. But this sort of packet prioritization has not been widely implemented. Maybe it should be.

    Let’s assume packet prioritization was widely implemented and application providers abused it by assigning all of their packets a high priority. Would an ISP run afoul of the FCC if it lowered the priority of packets based on application type, for example by lowering the priority of BitTorrent traffic?

  • http://pobox.com/~adamm Adam Marcus

    As Vint says, “the real question for today’s broadband networks is not whether they need to be managed, but rather how.” In a sense, managing the Internet is like managing highways. As more highways (and more lanes) are built, people move further from where they work and commute for longer distances. It’s called induced demand. Similarly, as the supply of bandwidth increases, the demand for bandwidth increases. See here and here. As Berin Szoka put it, “trying to solve network congestion problems simply by increasing the amount of bandwidth available is like a pie-eating contest where the prize is… more pie.” Just as full-screen standard-definition video over the Internet is becoming a reality, everyone starts clamoring for high-definition video over the Internet. So at some point (whether because the network isn’t built fast enough, a natural disaster makes everyone pick up their phone, or Victoria’s Secret does an online lingerie show), traffic congestion will force routers to delay or drop packets. When that happens, following a simple “first in, first out” approach just doesn’t make sense. Just as we want ambulances and fire engines to be given priority on our roadways, most people would rather their VoIP conversations get through than someone’s file download finishes a few seconds or minutes faster.

    Vint suggests that “Internet traffic should be managed with an eye towards applications and protocols.” In a sense, Comcast was doing exactly that by limiting BitTorrent traffic. BitTorrent is a specific application with a specific protocol. It is not time-sensitive like VoIP so it makes sense to limit BitTorrent traffic first when network congestion becomes a problem. But Comcast (according to the FCC’s press release) was limiting BitTorrent traffic at all hours of the day and in all sorts of areas, presumably even in areas where network congestion wasn’t a problem. Vint clarifies his position: “such prioritization should be applied across the board to all low latency traffic, not just particular application providers.” So managing a specific application is bad, but prioritization in general is good. What’s needed is an agnostic way to prioritize packets.

    There *is* such an agnostic method for controlling traffic. Every TCP/IP packet can be flagged with a certain priority level, and routers can be programmed to handle packets differently depending on their priority. But this sort of packet prioritization has not been widely implemented. Maybe it should be.

    Let’s assume packet prioritization was widely implemented and application providers abused it by assigning all of their packets a high priority. Would an ISP run afoul of the FCC if it lowered the priority of packets based on application type, for example by lowering the priority of BitTorrent traffic?

  • Bret Swanson

    Cerf’s acknowledgement of obvious truths about the way networks work is welcome news and could be a constructive step toward the industry cooperation Adam suggests. Perhaps we can head off further bungling by the FCC. But it would have been nice for opponents of “traffic management” to acknowledge these obvious facts — “So the real question for today’s broadband networks is not whether they need to be managed, but rather how” . . . “Network capacity (bits per second or data rate) is a limiting factor in all communications networks. Users cannot send traffic faster than the amount of network capacity available to them. But when users’ aggregate demand exceeds the available capacity of the network, network operators naturally seek to manage the traffic loads” — *before* the FCC handed down its technically, legally, and procedurally wanting traffic management directive. I’m all for industry collaboration and self-regulation, but not with a heavy hand tipping the scales in any “negotiations.”

  • Bret Swanson

    Cerf’s acknowledgement of obvious truths about the way networks work is welcome news and could be a constructive step toward the industry cooperation Adam suggests. Perhaps we can head off further bungling by the FCC. But it would have been nice for opponents of “traffic management” to acknowledge these obvious facts — “So the real question for today’s broadband networks is not whether they need to be managed, but rather how” . . . “Network capacity (bits per second or data rate) is a limiting factor in all communications networks. Users cannot send traffic faster than the amount of network capacity available to them. But when users’ aggregate demand exceeds the available capacity of the network, network operators naturally seek to manage the traffic loads” — *before* the FCC handed down its technically, legally, and procedurally wanting traffic management directive. I’m all for industry collaboration and self-regulation, but not with a heavy hand tipping the scales in any “negotiations.”

  • Pingback: you can look here

  • Pingback: quality seo services

  • Pingback: garcinia cambogia, about garcinia cambogia, all natural garcinia cambogia, all natural garcinia cambogia extract, all natural weight loss, all natural weight loss supplement, amazon garcinia cambogia extract, amazon garcinia cambogia extract pure, appetit

  • Pingback: premier league philippines

  • Pingback: prix de l'immobilier

Previous post:

Next post: