On Dec 18, 2009, at 10:32 AM, Steven Bellovin wrote:

Could you post a summary, in appropriate technical terms, of precisely what is being requested, and what changes to BGP they want?

Really. I can read tea leaves with the best of them, and the tea leaves I see tell me the reporter (in the story the blog points to) doesn't have a clue. What is the substance of the proposal? Depending on objectives, I would expect that this means that China wants to look at routers (which run BGP), and (a) use IPFIX-or-something to measure traffic rates and charge for trans-China transit, (b) use interface statistics to measure traffic rates and charge for trans-China transit, (c) tax Chinese ISPs for transit services they provide, or maybe (d) use IPFIX-or-something to map communication patterns. It would be (d) that the reporter might seriously want to worry about. But what is all this about "is the ITU interested in changing BGP"? If the word "metering" makes any sense in context, BGP doesn't meter anything.

On Dec 19, 2009, at 1:47 AM, Fred Baker wrote:

I can read tea leaves with the best of them, and the tea leaves I see tell me the reporter (in the story the blog points to) doesn't have a clue. What is the substance of the proposal?

The report seemed a reasonably accurate account of what went on in Kampala.

But what is all this about "is the ITU interested in changing BGP"? If the word "metering" makes any sense in context, BGP doesn't meter anything.

The Chinese delegation presented a dozen pages of formulae involving 20+ variables, infinite sums, and other mathematical goodies. Wowing the audience I guess. The whole way through "using BGP" was mentioned - in the sense of pulling data from, and adding data to BGP for the purposes of evaluating these formulae. It was clear that BGP would be used - and modified if need be - to achieve this. Mixing billing with the reachability information signalled through BGP just doesn't seem like a good idea. Interesting to note was that nowhere was the intent of all this mentioned, which is presumably to calculate the "value" each and every party's traffic traversing a link generates, and to apportion "costs" accordingly. Misguided, nonsensical, and unworkable ideas often gain traction. It's important that this one doesn't. Cheers, Jonny.

On Dec 19, 2009, at 2:24 AM, Jonny Martin wrote:

Mixing billing with the reachability information signalled through BGP just doesn't seem like a good idea.

This is done all the time via combinatorial BGP/NetFlow analysis, for peering/transit analysis reports, offnet/on-net billing differentials, etc. The merits (or lack thereof) of the 'proposal' in question aside, there's nothing evil or stupid about doing this on one's own network. ----------------------------------------------------------------------- Roland Dobbins <rdobbins@arbor.net> // <http://www.arbornetworks.com> Injustice is relatively easy to bear; what stings is justice. -- H.L. Mencken

On Dec 19, 2009, at 11:09 AM, James Hess wrote:

Otherwise, new router hardware could more easily provide suitable counters and IPFIX data (with suitable changes to ip flow export formats) to track the tarrifs due to all "tarrif payee IDs", or whatever that would be.

Existing hardware does this today with NetFlow, et. al. ----------------------------------------------------------------------- Roland Dobbins <rdobbins@arbor.net> // <http://www.arbornetworks.com> Injustice is relatively easy to bear; what stings is justice. -- H.L. Mencken

On Fri, Dec 18, 2009 at 10:09:32PM -0600, James Hess wrote:

On Fri, Dec 18, 2009 at 1:24 PM, Jonny Martin <jonny@pch.net> wrote: ..

...

modified if need be - to achieve this. ?Mixing billing with the reachability information signalled through BGP just doesn't seem like a good idea.

Indeed not.. but it might offer one advantage, if it was mandatory for any such tarrif/cost to be advertised there to be valid, and in the form of an ancillary BGP route attribute, rather than buried in some 500,000 page treaty that forces all ISPs to decipher it and try to figure out what their liabilities are.

Mainly because it makes any tarrif very visible, and easily understood. and offers an easy ability to automatically make decisions like discard reachability information that has any billing labels or "strings" attached to it, or has a cost greater than $X per million packets listed for 'source'... and easily allows an ISP to replace the next hop with null when a tarrif option has been listed, or use only a route not subject to tarrif.

I concur. Such visibility is efficient and drives simplification and automation from a data mining perspective, when analyzing accounting information. In such context, some care is required. Reachability information is destination based. Mixing accounting (ie. NetFlow) and reachability (ie. BGP) information is of good value for traffic delivered out of a routing domain but not for traffic received, ie. reverse reachability lookups can be a way although they are not truly deterministic due to routing asymmetries; a mix of ingress measurements, lookup maps and an export protocol supporting L2 information (ie. for same interface, multiple peers scenarios) give way a better chance to resolve which neighboring party is pulling which traffic into the observed domain. Cheers, Paolo

On Dec 19, 2009, at 6:42 PM, Randy Bush wrote:

this is particularly impressive given the complete absense of any facts about the alleged proposal.

I think the whole brouhaha is the merely result of someone saying 'BGP-speaking routers' vs. saying 'peering/transit edge routers', combined with the notion of somehow cartelizing this on a national basis vs. the current individual network-to-network private/public basis. ----------------------------------------------------------------------- Roland Dobbins <rdobbins@arbor.net> // <http://www.arbornetworks.com> Injustice is relatively easy to bear; what stings is justice. -- H.L. Mencken

However, if they are after some consultancy time to write some useless documents then I will happily take their money. -- Leigh Porter -----Original Message----- From: Randy Bush [mailto:randy@psg.com] Sent: Sat 12/19/2009 11:42 AM To: North American Network Operators Group Subject: Re: Chinese bgp metering story i am truely in awe how deeply the implications and alternatives have been analysed. this is particularly impressive given the complete absense of any facts about the alleged proposal. randy

Paolo Lucente wrote:

On Fri, Dec 18, 2009 at 10:09:32PM -0600, James Hess wrote:

...

...

modified if need be - to achieve this. ?Mixing billing with the reachability information signalled through BGP just doesn't seem like a good idea. Indeed not.. but it might offer one advantage, if it was mandatory for any such tarrif/cost to be advertised there to be valid, and in

On Fri, Dec 18, 2009 at 1:24 PM, Jonny Martin <jonny@pch.net> wrote: .. the form of an ancillary BGP route attribute, rather than buried in some 500,000 page treaty that forces all ISPs to decipher it and try to figure out what their liabilities are.

Mainly because it makes any tarrif very visible, and easily understood. and offers an easy ability to automatically make decisions like discard reachability information that has any billing labels or "strings" attached to it, or has a cost greater than $X per million packets listed for 'source'... and easily allows an ISP to replace the next hop with null when a tarrif option has been listed, or use only a route not subject to tarrif.

I concur. Such visibility is efficient and drives simplification and automation from a data mining perspective, when analyzing accounting information.

In such context, some care is required. Reachability information is destination based. Mixing accounting (ie. NetFlow) and reachability (ie. BGP) information is of good value for traffic delivered out of a routing domain but not for traffic received, ie. reverse reachability lookups can be a way although they are not truly deterministic due to routing asymmetries;

deliberate tunning for purposes of TE, use of default. will all contribute to ingress path not resembling egress...

a mix of ingress measurements, lookup maps and an export protocol supporting L2 information (ie. for same interface, multiple peers scenarios) give way a better chance to resolve which neighboring party is pulling which traffic into the observed domain.

Cheers, Paolo

From the BBC article quoted in the isoc-ny.org link:

An ITU spokesman said: "The ITU has no plans to modify the BGP protocol, which is not an ITU-T standard. "A proposal has been made, and is being studied, to use BGP routers to collect traffic flow data, which could be used, by bilateral agreement, by operators for billing purposes." ---- I read this to mean, no news here. If you want to move traffic, you need a bilateral agreement. That already exists. Where/if money flows, we know circuits don't build themselves for free, so the question of using money isn't a question. The only question is whether you are adjusting based on usage, or ports, or total speed, or direction of bits. ITU is already acknowledging that BGP isn't its baby, so it has nothing to say there. Deepak

On Fri, 18 Dec 2009, Deepak Jain wrote: > ITU is already acknowledging that BGP isn't its baby, so it has nothing to say there. Yes, that was the successful (for us) outcome of the meeting, which would not have been the case had we not been prepared and had people there. Just to explain the general danger here... The ITU is the standards body in which international spectrum allocations and satellite lots are negotiated. No industrialized country will withdraw from that. Because it's an international treaty organization, member countries are bound to enforce the outcome of its decisions within their jurisdictions, regardless of whether they agreed with the decision or not. If the ITU had decided to take the BGP spec from the IETF, the IETF could easily have told them to take a hike, but national governments could not have done so, and that would put national governments in the very uncomfortable position of having to try to enact or support that change in law somehow. With the BGP spec, this all seems a bit ridiculous and abstract, but with IP allocation, the danger is a little more immediate. The decision on that will mostly be made in mid-March. -Bill

On Fri, 18 Dec 2009, Dobbins, Roland wrote: > > The decision on that will mostly be made in mid-March. > By whom? A working group of the ITU Council. > The RIRs aren't just going to say, "OK, ITU folks, it's all yours," heh. Indeed not. However, the RIRs don't have a voice in the decision. This is an intergovernmental decision within the ITU Council. If the ITU Council were to decide that it's a good idea for the ITU to take over IP addressing and break it, they would then take it to the ITU Plenipotentiary. At that point, it could become policy of the treaty organization, and then member country governments would become bound to support the policy in their own legal structures. Odds are that would be expressed in laws similar to that of Korea, where it's illegal for network operators to get IP addresses from APNIC, their RIR, and they must instead get them from KRNIC, a Korean governmental agency. Which, in turn, proxies their votes in the APNIC elections, but that's another story. :-) -Bill

Why can't we carry price per "kilosegment" on BGP ? And don't be so hard on the ITU folks, the only thing they want to break is the monopoly of IP address allocation. J

On Fri, Dec 18, 2009 at 2:53 PM, Fred Baker <fred@cisco.com> wrote:

On Dec 18, 2009, at 12:43 PM, Jorge Amodio wrote:

...

And don't be so hard on the ITU folks, the only thing they want to break is the monopoly of IP address allocation.

With all due respect, they don't want to break said monopoly, assuming one agrees that it is a monopoly (I think there's a lot more to the story than that, but that's another discussion). They want to *be* said monopoly.

Indeed !!! I was being sarcastic, I was watching live the last IGF meeting when by proxy ICANN's CEO got grilled with the question about IPv6 address allocation. Jorge

Nobody here remembers ICAIS? This is actually an old story/ambition, which started elsewhere, and not long after the the 1997-1998 "rebalancing" of ITU-mediated switched telecom settlements. Two nuggets from the history books pasted in below. Of course, just because it's not new doesn't mean that it's not newsworthy. As I recall, this issue precipitated a fairly titanic behind-the-scenes struggle last time around... TV _____ AAP NEWSFEED July 15, 1999, Thursday Telstra chief calls for equitable Net traffic cost sharing SYDNEY, July 15 AAP - Telstra Corp Ltd chief executive Ziggy Switkowski today called for an equitable arrangement for sharing the cost of carrying Internet traffic to and from the United States.In an address to the Asia-Pacific Economic Cooperation (APEC) business conference here, Dr Switkowski said US operators were currently enjoying an implied subsidy of 30 per cent of the costs of international Internet connection... The charging system operates on a similar principle to that used in international phone charging arrangements, he said. "For Australia alone, that represents approximately $50 million a year, and the sum varies from country to country depending on usage," Dr Switkowski said. "Telstra's view is that the future of e-commerce could be undermined if investment in capacity growth does not match growth in demand. "But infrastructure providers outside the US need to have sufficient confidence in cost sharing to invest in new capacity to meet the exploding demand for bandwidth"... _____ Economist October 19, 1996 Too cheap to meter? The fact that the Internet seems free to many of its users has been one reason for its success. Now it may have to change. But how? ...If the costs of the telephone companies and the Internet are similar, why are their methods of pricing different? The answer is that telecoms charges bear little relation to costs. The telephone industry is regulated nearly everywhere and in most countries prices are set by bureaucrats and commissions; real costs are hidden by a layer of crosssubsidies. The Internet, on the other hand, is essentially unregulated. At present, telephone companies typically make less than half their revenue from fixed charges rather than from the price of each call. Tim Kelly, of the International Telecommunication Union in Geneva, reckons that the share of revenue from connection charges and monthly rentals has risen in the past decade from about 33% to 40%; he expects an increase to around 60% over the next ten years. The companies are not keen on such "rebalancing", since it usually involves reducing lucrative call charges rather than increasing fixed charges. But without it, they are vulnerable to competition, including competition from the Internet, which can offer rival services far less expensively... ...Such settlements are a source of endless argument: America's long- distance carriers complain that local telephone companies overcharge them. Moreover, they transfer huge sums of money between countries: in 1994, carriers based in the United States handed over a net $ 4.3 billion to foreign carriers. Because countries in which telephoning is cheap (such as America) tend to ring countries where calls are dearer, American carriers grumble that they are subsidising the inefficient and uncompetitive. Gradually, therefore, telephone companies are moving towards a "sender-keeps-all" system, where they will charge each other a flat fee for access to a certain amount of transmission capacity, rather than bill each other on the basis of use. That would bring them increasingly into line with what happens on the Internet, where settlement is rudimentary. There are payments between each of the Internet's hierarchy of links: access providers pay their regional network and regional networks pay the companies that operate the high-capacity long-distance parts, the backbone of the system. But such payments are mostly based on the availability of capacity, not its use: service providers simply agree to carry each other's traffic without totting up precise bills. This encourages a "hot-potato" approach: Internet access providers hand traffic on as quickly as possible to the carrier taking it to its ultimate destination. That benefits small service providers and irritates big ones, who say they get little reward for the effort of carrying the traffic for most of its journey. In turn, this lessens their incentive to invest in new capacity. The problem of settlement is worse for access providers outside America. Led by Singapore Telecom and Australia's Telstra, they complain that they have to pay all the cost of leasing lines between their country and the United States. "The rest of the planet subsidises the United States," argues Barry Greene, who works for Cisco, a maker of routers, but was previously with Singnet, the Internet arm of Singapore Telecom. The high cost of leasing international lines means that upgrading them to ease congestion can cost a non-American company ten times more than in America.