RE: [address-policy-wg] getting second IPv6 PA as a LIR
And why wouldn't the Internet work with 600,000 prefixes in the DFZ?
Now all of a sudden for instance Cisco 7600 3CXL isn't enough to old a full table (at around 750k).
Also, IPv6 uses twice the TCAM resources as IPv4...
so, basically what you are saying is that you know that your routers need an upgrade in 5 years and you don't want to pay for an upgrade or you can't figure out a business plan which covers the costs for that? But you are telling small ISPs/NCOs/"hobbyusers"/whatever THEY don't get their business plan right if they don't can afford paying $$$ for PI space or rather would prefer to pay other bills with the money? WTF?! ------ No, the problem that the small ISPs you are speaking about will have to spend that money to swallow such routing table. And it is not $2000, "slightly" more...
On Tue, 3 May 2011, poty@iiat.ru wrote:
so, basically what you are saying is that you know that your routers need an upgrade in 5 years and you don't want to pay for an upgrade or you can't figure out a business plan which covers the costs for that?
Paying for the upgrade by having higher prices towards end-customer is what is going to happen all across the world if we get hundreds of thousands of Ipv6 PI. So everybody pays, just not the ones causing the problem. Classic externalising of costs. See my earlier links.
But you are telling small ISPs/NCOs/"hobbyusers"/whatever THEY don't get their business plan right if they don't can afford paying $$$ for PI space or rather would prefer to pay other bills with the money? WTF?!
I guess you don't belive in "polluter pays". -- Mikael Abrahamsson email: swmike@swm.pp.se
On Tue, May 3, 2011 at 16:03, Mikael Abrahamsson <swmike@swm.pp.se> wrote:
Paying for the upgrade by having higher prices towards end-customer is what is going to happen all across the world if we get hundreds of thousands of Ipv6 PI. So everybody pays, just not the ones causing the problem. Classic externalising of costs. See my earlier links.
You did not establish any proof that there will be hundreds of thousands of additional PI prefixes in global routing tables. I find this claim questionable, at best.
I guess you don't belive in "polluter pays".
It seems to me that € 50 / year * PI prefix is a cost that's being paid today. Of course, you are free to argue that this is not enough. Maybe more people would agree with you if you wouldn't use ridiculously inflated numbers, both for the projected PI prefixes and your desired price of € 2000 / year * PI prefix. Or tried to write your emails in a slightly less heated manner. Best regards, Richard
On Tue, 3 May 2011, Richard Hartmann wrote:
Maybe more people would agree with you if you wouldn't use ridiculously inflated numbers, both for the projected PI prefixes and your desired price of € 2000 / year * PI prefix. Or tried to write your emails in a slightly less heated manner.
I am not stupid, I of course know RIPE is never going to charge 2kEUR per year for PI even if I want them to. Also, I am going to save this email and bring it out in 5 years and let's see how many 50EUR/year PIs there are in the world then. I guess that's the only way to find out. I am not heated, I am resigned about nobody here caring about the common good. Seems shared cost is "someone elses problem". As long as "my hobby project" is important to me, I should get to pollute without paying. And at 50EUR per year, heck, I'd PI my home connection in a blink. Why wouldn't anyone? -- Mikael Abrahamsson email: swmike@swm.pp.se
Hi, On Tue, May 03, 2011 at 04:20:42PM +0200, Mikael Abrahamsson wrote:
And at 50EUR per year, heck, I'd PI my home connection in a blink. Why wouldn't anyone?
Well, it's a bit more to it than just the 50 EUR - as of today, most cheap end user DSL connections will not route arbitrary PI networks for customers, but instead will insist on automatic assignment of addresses (coming from a block aggregated by the ISP, whatever you call it) to the end users. So you need a "business connection" and that will be more than 50 EUR/year. Of course one of the big Telcos might decide that this is *the* product to compensate for loss of customers due to <whatever excuse>, but I seriously don't really see it happen any time soon, not for the large masses of end customers. We'll have some "PI" in the global table, of course. But please look at Alex Le Heux' numbers that will be presented on Thursday morning (about 10:00) on what makes up the bulk of the current IPv4 routing table - and you might be surprised. Gert Doering -- APWG chair -- did you enable IPv6 on something today...? SpaceNet AG Vorstand: Sebastian v. Bomhard Joseph-Dollinger-Bogen 14 Aufsichtsratsvors.: A. Grundner-Culemann D-80807 Muenchen HRB: 136055 (AG Muenchen) Tel: +49 (89) 32356-444 USt-IdNr.: DE813185279
We'll have some "PI" in the global table, of course. But please look at Alex Le Heux' numbers that will be presented on Thursday morning (about 10:00) on what makes up the bulk of the current IPv4 routing table - and you might be surprised.
Is this presentation available online somewhere for those who can't be in Amsterdam this week and for whom the time of the prezo is not so convenient to participate remotely? There doesn't appear to be any pointer to the presentation material on the RIPE 62 meeting plan. I'm curious to see whether the drivers for global routing state have changed since I gave similar presentations a few years ago. FWIW, I'd assert that the dual role of IP addresses as both endpoint identifiers and routing locators means that finding a scalable solution that allows both provider independence and scalable routing is basically impossible without implementing separation of those roles into two separate numbering spaces. Thanks, --Vince
Hi, sorry for the slow response: On Wed, May 04, 2011 at 11:50:10AM -0700, Vince Fuller wrote:
We'll have some "PI" in the global table, of course. But please look at Alex Le Heux' numbers that will be presented on Thursday morning (about 10:00) on what makes up the bulk of the current IPv4 routing table - and you might be surprised.
Is this presentation available online somewhere for those who can't be in Amsterdam this week and for whom the time of the prezo is not so convenient to participate remotely? There doesn't appear to be any pointer to the presentation material on the RIPE 62 meeting plan.
Actually, when clicking on a working group's title, it should have showed you the agenda on the left side, and all the presentations on the right side. Now all the presentations are available in a central archive: http://ripe62.ripe.net/presentations/presentation-archive and specifically, Alex' presentation is here: http://ripe62.ripe.net/presentations/160-apwg.key (keynote only, pdf conversion will be provided later, as far as I understand) Gert Doering -- NetMaster -- did you enable IPv6 on something today...? SpaceNet AG Vorstand: Sebastian v. Bomhard Joseph-Dollinger-Bogen 14 Aufsichtsratsvors.: A. Grundner-Culemann D-80807 Muenchen HRB: 136055 (AG Muenchen) Tel: +49 (89) 32356-444 USt-IdNr.: DE813185279
Mikael, On Tue, May 3, 2011 at 10:20 AM, Mikael Abrahamsson <swmike@swm.pp.se> wrote:
On Tue, 3 May 2011, Richard Hartmann wrote:
Maybe more people would agree with you if you wouldn't use ridiculously inflated numbers, both for the projected PI prefixes and your desired price of € 2000 / year * PI prefix. Or tried to write your emails in a slightly less heated manner.
I am not stupid, I of course know RIPE is never going to charge 2kEUR per year for PI even if I want them to.
Also, I am going to save this email and bring it out in 5 years and let's see how many 50EUR/year PIs there are in the world then. I guess that's the only way to find out.
I am not heated, I am resigned about nobody here caring about the common good. Seems shared cost is "someone elses problem". As long as "my hobby project" is important to me, I should get to pollute without paying.
And at 50EUR per year, heck, I'd PI my home connection in a blink. Why wouldn't anyone?
Why don't you? :) I'll just re-iterate an earlier point I made in this thread: If a routing protocol can't scale in the core, it's either applied wrong or there is a problem with the protocol design itself. I don't think taxing people away is a good sustainable/stable approach to solving the problem of the routing protocol in question. In the context of a potential (not certain) explosion of *both* IPv4 and IPv6 prefixes in the BGP DFZ's, LISP or other specific solutions to the expensive core *may* become necessary. My view is that *if* this explosion comes, it does so fully regardless of RIPE's PI/PA IPv6 policies. Regards, Martin
On Tue, May 3, 2011 at 16:20, Mikael Abrahamsson <swmike@swm.pp.se> wrote:
I am not stupid, I of course know RIPE is never going to charge 2kEUR per year for PI even if I want them to.
Wouldn't it make sense to argue for goals you have a chance of achieving instead of dragging out a fringe discussion about things that will not happen anyway?
Also, I am going to save this email and bring it out in 5 years and let's see how many 50EUR/year PIs there are in the world then. I guess that's the only way to find out.
Fine by me. To put my money where my mouth is and as "hundreds of thousands" implies 200,001 or more, how about this bet: If, on the 1st of May, 2016, there are 200,001 or more IPv6 PI routes (1/5th of which need to be from RIPE space) in the global routing table, I will owe you one good bottle of rum. If this condition is not met, you owe me a good bottle of rum. "Good" means all rums in the blend need to be at least 8 years old and it must not be from any company manufacturing industrial paint solvents like Bacardi or Havanna Club. I maintain a separate calendar for bets so this would not get lost.
I am not heated, I am resigned about nobody here caring about the common good.
Your definition of the common good seems to diverge from the one most of the other people posting in this thread. Disagreement is fine, implying others are careless and ignorant, not so much.
Seems shared cost is "someone elses problem". As long as "my hobby project" is important to me, I should get to pollute without paying.
Again, € 50 / year * PI prefix is not free.
And at 50EUR per year, heck, I'd PI my home connection in a blink. Why wouldn't anyone?
Mainly because you do not have a need and will not get it routed on a normal home connection. Same as for almost all other home, or business, connections. Richard PS: Arguably, "hundreds of thousands" means 300,001 or more, but I am confident 200,001 will not be reached so the point is moot.
* Mikael Abrahamsson:
I am not heated, I am resigned about nobody here caring about the common good.
Equal access to IP addresses for a nominal fee (covering registration processing costs) is a common good, too. 8-) -- Florian Weimer <fweimer@bfk.de> BFK edv-consulting GmbH http://www.bfk.de/ Kriegsstraße 100 tel: +49-721-96201-1 D-76133 Karlsruhe fax: +49-721-96201-99
On Tue, May 03, 2011 at 04:20:42PM +0200, Mikael Abrahamsson wrote:
I am not heated, I am resigned about nobody here caring about the common good.
The common good is an Internet which allows nondiscriminatory access to whoever wants, with no ARTIFICIAL (and that's your PI tax) barriers. No matter how financially potent they are, wether they are a business or non-commercial org or even just a private person. People tend to forget that Internet is more than "ISPs trying to get revenue out of the big content players and the million numb eyeballs". Regards, Daniel -- CLUE-RIPE -- Jabber: dr@cluenet.de -- dr@IRCnet -- PGP: 0xA85C8AA0
Hay, Am 03.05.2011 um 16:03 schrieb Mikael Abrahamsson:
On Tue, 3 May 2011, poty@iiat.ru wrote:
so, basically what you are saying is that you know that your routers need an upgrade in 5 years and you don't want to pay for an upgrade or you can't figure out a business plan which covers the costs for that?
Paying for the upgrade by having higher prices towards end-customer is what is going to happen all across the world if we get hundreds of thousands of Ipv6 PI. So everybody pays, just not the ones causing the problem. Classic externalising of costs. See my earlier links.
It's exactly the other way round. The big ISPs sell Internet at ridiculous low price points, just to have the most customers in their market or just to win a bid over a multi million $$$ project over their competitors. Now they complain that they cannot keep up with the growth of the Internet/the DFZ and want to stop new people or even competitors to join in? (Not to forget: Try to make content providers pay for transporting date to their own customers, see "net neutrality...") And you really think you have a valid argument there that the prices need to be raised for everyone? They shouldn't have gone down so far in the first place if you had a good business plan! But of course, now that is too late. It's like the financial crisis all over again... big companies complaining they are out of money, please someone else pay for us and protect us from new competitors.
But you are telling small ISPs/NCOs/"hobbyusers"/whatever THEY don't get their business plan right if they don't can afford paying $$$ for PI space or rather would prefer to pay other bills with the money? WTF?!
I guess you don't belive in "polluter pays".
You first have to prove to us that adding prefixes to the DFZ is pollution as long as it's not useless de-aggregates. Who are you to decide which entity is allowed to be in the DFZ and which isn't - the god of the internet? If you decide to get multihomed in your bedroom, get two BGP Upstreams and multihome. I don't care. I guess you are educated enough that you know what you are doing and why you need to do this and the RIPE NCC hostmas...i mean... IPRAs :-) will have double checked that you have valid reasons so i'm fine! Probably you will set up the next Facebook or the next Twitter in your bedroom...wouldn't that be great for the Internet? -- Mit freundlichen Grüßen / Kind Regards Sascha Lenz [SLZ-RIPE] Senior System- & Network Architect
On Tue, May 3, 2011 at 9:37 AM, <poty@iiat.ru> wrote:
And why wouldn't the Internet work with 600,000 prefixes in the DFZ?
Now all of a sudden for instance Cisco 7600 3CXL isn't enough to old a full table (at around 750k).
Also, IPv6 uses twice the TCAM resources as IPv4...
so, basically what you are saying is that you know that your routers need an upgrade in 5 years and you don't want to pay for an upgrade or you can't figure out a business plan which covers the costs for that? But you are telling small ISPs/NCOs/"hobbyusers"/whatever THEY don't get their business plan right if they don't can afford paying $$$ for PI space or rather would prefer to pay other bills with the money? WTF?!
------ No, the problem that the small ISPs you are speaking about will have to spend that money to swallow such routing table. And it is not $2000, "slightly" more...
Not really; normal PC RAM is pretty cheap. 16GB ECC/REG ~400 EUR. /M
poty@iiat.ru wrote:
And why wouldn't the Internet work with 600,000 prefixes in the DFZ? Now all of a sudden for instance Cisco 7600 3CXL isn't enough to old a full table (at around 750k). Also, IPv6 uses twice the TCAM resources as IPv4...
so, basically what you are saying is that you know that your routers need an upgrade in 5 years and you don't want to pay for an upgrade or you can't figure out a business plan which covers the costs for that? But you are telling small ISPs/NCOs/"hobbyusers"/whatever THEY don't get their business plan right if they don't can afford paying $$$ for PI space or rather would prefer to pay other bills with the money? WTF?!
------ No, the problem that the small ISPs you are speaking about will have to spend that money to swallow such routing table. And it is not $2000, "slightly" more...
frankly... small ISPs run on Linux/BSD... 4 core, 8 core, 16 core, 64G of memory, more and more stable kernels, drivers etc... If they don't want to pay LIR expanses, they will also resist to buy Cisco/Juniper/Ericssn/whatever based on ASICs.. Regards, Marcin
Hay, Am 03.05.2011 um 15:37 schrieb <poty@iiat.ru> <poty@iiat.ru>:
And why wouldn't the Internet work with 600,000 prefixes in the DFZ?
Now all of a sudden for instance Cisco 7600 3CXL isn't enough to old a full table (at around 750k).
Also, IPv6 uses twice the TCAM resources as IPv4...
so, basically what you are saying is that you know that your routers need an upgrade in 5 years and you don't want to pay for an upgrade or you can't figure out a business plan which covers the costs for that? But you are telling small ISPs/NCOs/"hobbyusers"/whatever THEY don't get their business plan right if they don't can afford paying $$$ for PI space or rather would prefer to pay other bills with the money? WTF?!
------ No, the problem that the small ISPs you are speaking about will have to spend that money to swallow such routing table. And it is not $2000, "slightly" more...
Small companies start with small routers, PC based Linux Quagga Boxes, or Routerboards, or Juniper J-Series or whatever - not really much costs here (see other replies). If they become bigger, they need to take into account that bigger boxes cost more money, or they decide to stay small. In any way, PI space is about END-USERS, not about ISPs so much anyways. So if you're a small ISP, with this thinking in mind that you don't want more prefixes in the DFZ, you are basically saying that you don't want any more highlevel customers by raising the entry level for such endusers who require this kind of service (full BGP redundancy)? You rather serve homeusers for cheap money which never get you enough revenue? You don't want new startups competing with your company? ...or what? What's the point about all you guys who want to prevent new people to be in the DFZ? Is there any reasonable point which isn't selfish? I just don't get it for some years now. -- Mit freundlichen Grüßen / Kind Regards Sascha Lenz [SLZ-RIPE] Senior System- & Network Architect
participants (10)
-
Daniel Roesen
-
Florian Weimer
-
Gert Doering
-
Marcin Kuczera
-
Martin Millnert
-
Mikael Abrahamsson
-
poty@iiat.ru
-
Richard Hartmann
-
Sascha Lenz
-
Vince Fuller