- NANOG - lists.nanog.org

Game Theory (was: RE: Sprint peering policy)
by Scott A Crosby 01 Jul '02

01 Jul '02

On Mon, 1 Jul 2002, David Schwartz wrote: > On Mon, 1 Jul 2002 12:11:46 -0500, Paul A Flores wrote: > > >Since it seems we are speaking of 'zero cost' interconnects, if Either X OR > >Y feel like they are getting ripped, they won't (and shouldn't) do it. If > >party X feels that party Y is gaining more from the interconnect that they > >are, X might feel the need to lay some surcharges of some time on the > >connection, which is only fair, if they feel they aren't receiving value for > >value. > > Suppose that X is presented with a take it or leave it deal. Suppose further > that X, on net, benefits from this deal. Why should they care how much or how > little Y, Z, or T benefit from the deal? What kind of business sense does > that make? > > >Otherwise, esp. now that enough people have gotten their hands caught in the > >cookie jars, why would they GIVE away 'free' services for nothing in return? > > We're not talking about "nothing in return". We're talking about an > arrangement between two parties that both benefit from. Why should one party > care how much the other benefits? (Except, of course, as possible leverage to > negotiate a better deal.) > What you're describing is a game theory problem.. The situation you describe, where neither party charges for as long as both parties benefit, this only occurs when the parties have no information on each other. Lets say you have two parties, A&B, and zero transaction costs. Furthermore, both parties wish to obtain the MAXIMUM benefit from the deal. A benefits $ 50 B benefits $100 Now, there are 4 situations: 1. A knows nothing about how much B will benefit and vice versa. In this case, if either party tries to charge the other, they run a risk of having the deal fall through, and getting no benefit. The optimal situation is for neither party to request any payment from the other. (This situation is theoretical, in real business, you could make guesses as to the benefit obtained by another party.) Total benefit is $50 & $100 2. B knows that A will benefit by $50. A knows nothing. Then, B could demand a charge of $49. A would choose to pay it, for they would obtain a net benefit of $1. Note that in this case, B who's already benefitting the most would get paid by A! Total benefit is $1 & $149 3. A knows that B will benefit by $100. B knows nothing. Then A could demand a charge of $99. B would choose to pay it, for they would obtain a net benefit of $1. Total benefit is $149 and $1. 4. A and B both know how much the other benefits. Then.. I'm not exactly sure how game theory reasons this out. In reality, there is more than one party, and imperfect information on respective benefits. I believe one would use an economic and statistical argument, and the expected result would be that B paid A $25. > Actually, I think you can make peering fair in a much more simple way. > Simply explain to people a sensible and rational way to evaluate their > peering decisions. "Does it benefit me as much as or more than anyone else" > is neither sensible not rational. "Does it benefit me more than it costs me?" > and "Is it the best deal I can negotiate for this amount of benefit?", on the > other hand, are sensible, rational, and fair. Knowing how much the other benefits from a particular transaction can lead to a remarkable difference in how valuable the transaction is to the respective parties. And maximizing revenue is the purpose of business. Scott

1 0

RE: Sprint peering policy
by Paul A Flores 01 Jul '02

01 Jul '02

> On 29 Jun 2002 02:32:03 +0000, Vijay Gill wrote: > > > >Mike Leber <mleber(a)he.net> writes: > > > >>Sprint's peers aren't equal to Sprint or each other when > considered by > >>revenue, profitability, number of customers, or > geographical coverage. > > > >A good proxy for the above is to ask the question: > > > >Do X and Y feel they derive equal value (for some value of equal) by > >interconnecting with each other? > > > >If they think they do, then an interconnection is set up between X > >and Y. However, if one party feels that they do NOT derive equal > >value by interconnecting with the other, than that party usually > >balks. > > This doesn't make any sense at all. Why should X care > how much value Y gets > out of the deal at all?! This is like saying that Burger King > should charge > hungrier people more for a Whopper. > Don't you think they would if they could? :) Since it seems we are speaking of 'zero cost' interconnects, if Either X OR Y feel like they are getting ripped, they won't (and shouldn't) do it. If party X feels that party Y is gaining more from the interconnect that they are, X might feel the need to lay some surcharges of some time on the connection, which is only fair, if they feel they aren't receiving value for value. Otherwise, esp. now that enough people have gotten their hands caught in the cookie jars, why would they GIVE away 'free' services for nothing in return? Peering with anyone is a pain, but a necessary one. If you don't have something anyone wants, your not going to get peered (for free) with anyone. You have ZERO value to anyone else, therefore, expect to pay for your connections. Does it really have to be more complicated than that? Since this is basically a financial issue (and not really a regulatory issue), the only way you could make it 'fair' is to have some kind of mandate from a government body to MAKE peering 'fair'. The only way _I_ would buy off on that, would be to have some kind of subsidy paid from tax dollars to the carriers in question to 'force' them to peer with people who have no other redeeming value. This way, I get paid, Y gets to brag to his peers that he is hooked up with X and my tax bill goes up... Talk about false progress. Isn't fudging the books how we got here in the 1st place? No thanks. -PF

3 2

Allocated IP blocks
by Mike Batchelor 01 Jul '02

01 Jul '02

Is there a list anywhere of allocated IP blocks? I need to update my IDS sensor's table of valid blocks. It's alarming on some traffic coming from 67, 68 and 219, which I know were not allocated until fairly recently. Is there a simple list somewhere, i.e: 4.0.0.0/8 6.0.0.0/8 .... 219.0.0.0/24 etc... ?? --- "The avalanche has already begun. It is too late for the pebbles to vote." -- Kosh

4 4

how is cold-potato done?
by Ralph Doncaster 01 Jul '02

01 Jul '02

If I peer with network X in cities A and B, and receive the same route in both cities with an AS-path of X, how do I know which city to use for an exit? I can understand how if X uses communities to tag the geographic origin of the traffic, but I'm not aware of many networks that do this. Lots of networks claim to use cold-potato routing though, so how do they do it? Ralph Doncaster principal, IStop.com

11 12

Re: Sprint peering policy
by alex＠yuriev.com 01 Jul '02

01 Jul '02

> >If they think they do, then an interconnection is set up between X > >and Y. However, if one party feels that they do NOT derive equal > >value by interconnecting with the other, than that party usually > >balks. > > This doesn't make any sense at all. Why should X care how much value Y gets > out of the deal at all?! This is like saying that Burger King should charge > hungrier people more for a Whopper. > It is called price descrimination and it does happen, not at burger king though. For example, those who arrive at certail site via Overture referals end up paying more for goods, compared to those who come from comparsion sites. Alex

1 0

valid IP address
by David Barak 01 Jul '02

01 Jul '02

Hi JothirLatha, Clearly it's possible to do this: 4500_3#conf t Enter configuration commands, one per line. End with CNTL/Z. 4500_3(config)#int lo1 4500_3(config-if)#no ip addr 4500_3(config-if)#ip addr 10.1.1.0 255.255.255.255 4500_3(config-if)#int lo2 4500_3(config-if)#ip addr 10.1.1.255 255.255.255.255 4500_3(config-if)#^Z Given that this is the group for network operators, why didn't you just try it? -David Barak "Quis custodes ipsos custodiet?" - Juvenal Date: Mon, 1 Jul 2002 07:53:32 -0700 (PDT) From: JothirLatha Jaganathan <jyothi_j(a)yahoo.com> To: nanog(a)merit.edu Subject: Valid ip address ? Can any one tell me if the last byte of a point to point or loopback ipaddress with 32 bit mask can be 0 or 255? For e.g.: Is 10.1.1.0/32 a valid interace address? __________________________________________________ Do You Yahoo!? Yahoo! - Official partner of 2002 FIFA World Cup http://fifaworldcup.yahoo.com

1 0

Re: Sprint peering policy
by Rizzo Frank 01 Jul '02

01 Jul '02

"Mitchell, Dan" wrote: > Plan on seeing an SLA of 99.999% Better-than-PSTN reliability, coming soon to an ALGX salesrep near you! What next, Dan, 6 9's? Rizzo, Frank

1 0

Re: Cook Report down?
by Rizzo Frank 01 Jul '02

01 Jul '02

> Link isn't working this morning. Thanks Jane. I've taken the liberty of copying all interested parties. Frank

1 0

Cook Report down?
by Pawlukiewicz Jane 01 Jul '02

01 Jul '02

Link isn't working this morning. Jane

2 1

RE: Sprint peering policy
by Mitchell, Dan 01 Jul '02

01 Jul '02

Here's what I've "heard" about Sprint and peering... Sprint's minimum to peer is 6 x OC-12's. To peer with them, you have to undergo a 90-day trial period. At the end of the 90 days, they will determine whether or not they would like to peer with you. Here's the catch: During the 90-day trial period, there is a NON-REFUNDABLE $125k/port fee (for each OC-12). So, you're looking at laying out $750k per month for 3 months, or $2.25 MILLION over the 90-day "trial" period, and even then, they MIGHT peer with you. But that's just what I've "heard". On a side note: Folks should keep an eye on Allegiance Internet (the folks I work for). They picked up the old Digex / Intermedia Business Internet backbone and are in the process of integrating it (under AS2548). Financially solid company, very easy to work with, and have some strong uptimes. (Plan on seeing an SLA of 99.999%/99.75%/65ms with a guaranteed 15 minute response time for loss of connectivity to customer prem and a guaranteed 40 business day install). For peering, I'm pretty sure the policy is Tier 1 = 4 points OC-3 or better, Tier 2 = 4 points DS-3 or better (Tier 2 peer pays loop charge if applicable). Thanks, Dan Dan Mitchell Internet Systems Engineer - Boston AllegianceInternet / Hosting.com 781.478.1857 Direct Line dan.mitchell(a)hosting.com -----Original Message----- From: Jeff Nelson [mailto:jnelson@rackspace.com] Sent: Wednesday, June 26, 2002 1:04 PM To: Majdi S. Abbas; Ralph Doncaster Cc: nanog(a)merit.edu Subject: Re: Sprint peering policy http://www.sprintlink.net/policy/index.html --what they publish $$$$$--what they don't jeff "Be liberal in what you accept, and conservative in what you send." --Jon Postel ----- Original Message ----- From: "Majdi S. Abbas" <msa(a)samurai.sfo.dead-dog.com> To: "Ralph Doncaster" <ralph(a)istop.com> Cc: <nanog(a)merit.edu> Sent: Wednesday, June 26, 2002 11:57 AM Subject: Re: Sprint peering policy > > On Wed, Jun 26, 2002 at 12:39:11PM -0400, Ralph Doncaster wrote: > > While many other tier-1's have publicly listed their peering policies, > > I've never seen anything for 1239. Not that I'd stand a chance, but does > > anyone know what their peering requirements are? > > sprintlink.net# grep peering /etc/aliases > peering: /dev/null > sprintlink.net# > > --msa >

4 3