IPv6 Policy Clarification - Initial allocation criteria "c)"
Dear Colleagues, As explained in the email sent on Mon, 14 Jun 2004: http://www.ripe.net/ripe/mail-archives/address-policy-wg/2004/msg00240.html This is the second mail request for clarification of the IPv6 Address Allocation and Assignment Policy. Below is an excerpt from the IPv6 Address Allocation and Assignment Policy: 5.1.1. Initial allocation criteria "c)" "To qualify for an initial allocation of IPv6 address space, an organisation must [...] plan to provide IPv6 connectivity to organisations to which it will assign /48s by advertising that connectivity through its single aggregated address allocation" LIRs who operate closed/private networks appear not to qualify because the address space in these networks will not be advertised. Was this the community's intention? Best Regards, Laura Cobley Registration Services RIPE NCC
On Tue 15 Jun 2004 16:36, Laura Cobley wrote:
"To qualify for an initial allocation of IPv6 address space, an organisation must [...] plan to provide IPv6 connectivity to organisations to which it will assign /48s by advertising that connectivity through its single aggregated address allocation"
Again, this seems to exclude mobile operators who may only want to assign /64s to their customers' handsets... rgds, Sascha -- Eirconnect | voice: 353 21 2307195 NSC Campus | fax: 353 21 2307197 Mahon, Cork | mailto:sascha@eirconnect.net Ireland | http://www.eirconnect.net
Sascha Luck wrote:
Again, this seems to exclude mobile operators who may only want to assign /64s to their customers' handsets...
I dont se how that assumption can be made. My mobile handset is connected to the public internet so that I can get updated exchange rates and read my email trough IMAP and so on. (and not trough NAT as my mobile operator has offered my the choice of not using NAT.) Iff this had been IPv6 and Iff Bluetooth had worked seemlessly on my other toys they could have had seemless access to the internet also. -hph
Hi, On Wed, Jul 13, 2005 at 11:29:21PM +0200, Hans Petter Holen wrote:
Again, this seems to exclude mobile operators who may only want to assign /64s to their customers' handsets...
I dont se how that assumption can be made. My mobile handset is connected to the public internet so that I can get updated exchange rates and read my email trough IMAP and so on.
(and not trough NAT as my mobile operator has offered my the choice of not using NAT.)
Iff this had been IPv6 and Iff Bluetooth had worked seemlessly on my other toys they could have had seemless access to the internet also.
Which would work fine with a /64 - "one big LAN with enough IPs"... This yields two questions: - is it likely that we'll see mobile handsets that provide connectivity to *two* (or more) independent IPv6 LAN networks? Like "bluetooth and WLAN, and not bridged"? - would a /60 suffice? Gert Doering -- NetMaster -- Total number of prefixes smaller than registry allocations: 71007 (66629) SpaceNet AG Mail: netmaster@Space.Net Joseph-Dollinger-Bogen 14 Tel : +49-89-32356-0 D- 80807 Muenchen Fax : +49-89-32356-234
Gert Doering wrote:
Hi,
On Wed, Jul 13, 2005 at 11:29:21PM +0200, Hans Petter Holen wrote:
Again, this seems to exclude mobile operators who may only want to assign /64s to their customers' handsets...
I dont se how that assumption can be made. My mobile handset is connected to the public internet so that I can get updated exchange rates and read my email trough IMAP and so on.
(and not trough NAT as my mobile operator has offered my the choice of not using NAT.)
Iff this had been IPv6 and Iff Bluetooth had worked seemlessly on my other toys they could have had seemless access to the internet also.
Which would work fine with a /64 - "one big LAN with enough IPs"...
This yields two questions:
- is it likely that we'll see mobile handsets that provide connectivity to *two* (or more) independent IPv6 LAN networks?
Yes - as I mentioned during the last RIPE meeting we are now seeing "Mobile broadband" products based on UMTS or EDGE and priced like xDSL in Norway - so if you are in a city and out of DSL coverage you can go mobile. In other words I do not think you can make the desicion on how many networks you need based on the transport technology - but there need to be some other cirteria.
Like "bluetooth and WLAN, and not bridged"?
- would a /60 suffice?
My feeling today is that in most cases a /64 will be sufficient for any *personal* network and going to /63 or /62 for advanced homes - but if I am to look 100 years into the future it is harder to say. Hans Petter
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 On 2004-06-15, at 17.36, Laura Cobley wrote:
Below is an excerpt from the IPv6 Address Allocation and Assignment Policy:
5.1.1. Initial allocation criteria "c)"
"To qualify for an initial allocation of IPv6 address space, an organisation must [...] plan to provide IPv6 connectivity to organisations to which it will assign /48s by advertising that connectivity through its single aggregated address allocation"
LIRs who operate closed/private networks appear not to qualify because the address space in these networks will not be advertised. Was this the community's intention?
More or less yes. If they do not plan to advertise this space, they should go for the "unique-site-local-replacement-addresses-that-you-are-not-allowed-to- route-globally-ever" (or whatever they will be called). - - kurtis - -----BEGIN PGP SIGNATURE----- Version: PGP 8.0.3 iQA/AwUBQM+P8KarNKXTPFCVEQLCNQCfXZjr5+dl0X/Yi37xYlM0fs7SVdAAoN/0 dDnfaTnpuWSwxIx39mYLseek =YS11 -----END PGP SIGNATURE-----
Below is an excerpt from the IPv6 Address Allocation and Assignment Policy:
5.1.1. Initial allocation criteria "c)"
"To qualify for an initial allocation of IPv6 address space, an organisation must [...] plan to provide IPv6 connectivity to organisations to which it will assign /48s by advertising that connectivity through its single aggregated address allocation"
LIRs who operate closed/private networks appear not to qualify because the address space in these networks will not be advertised. Was this the community's intention?
The intention was that any (globally routable) address space be allocated in a way that aggregates well. We need to limit the size of the DFZ routing tables. That means: - no random prefixes to end sites - end sites get addresses from LIRs, so the LIR can aggregate _all_ the routes to the end sites it covers via a single prefix At the time the policy was developed, we explicitely did not include discussion of "closed/private" networks. So basically, that topic isn't really covered in the current policy. Speaking personally, I don't see a problem with making allocations for closed networks, _if_ there is the possibility/intention that at some _future_ time the address will be publically advertised. I.e., for the short term the address space will be closed, but the long-term intention is otherwise. But see below as well.
More or less yes. If they do not plan to advertise this space, they should go for the "unique-site-local-replacement-addresses-that-you-are-not-allowed-to- route-globally-ever" (or whatever they will be called).
For sites that will never be publically routed, this is a reasonable way to go. (draft-ietf-ipv6-unique-local-addr-05.txt is the current version.) Note that the "site local replacement" addresses are essentially owned by the end site, so there is no need to ever return those addresses. For addresses allocated from an LIR, the address space is effectively bound to the LIR, and is not "portable". Thus, if at some later date the end site wants global visibility, the visibility will be through the LIR, not some other ISP. End sites need to understand the implications of both approaches prior to selecting a type of address appropriate for them. Thomas
In my opinion, closed networks today, could be connected tomorrow, and consequently advertised, so why exclude them ? Excluding them will mean that if a network is disconnected, so no advertised, even by accident, they could miss the right for that allocation ? Regards, Jordi ----- Original Message ----- From: "Laura Cobley" <laura@ripe.net> To: <address-policy-wg@ripe.net> Sent: Tuesday, June 15, 2004 5:36 PM Subject: [address-policy-wg] IPv6 Policy Clarification - Initial allocation criteria "c)"
Dear Colleagues,
As explained in the email sent on Mon, 14 Jun 2004:
http://www.ripe.net/ripe/mail-archives/address-policy-wg/2004/msg00240.html
This is the second mail request for clarification of the IPv6 Address Allocation and Assignment Policy.
Below is an excerpt from the IPv6 Address Allocation and Assignment Policy:
5.1.1. Initial allocation criteria "c)"
"To qualify for an initial allocation of IPv6 address space, an organisation must [...] plan to provide IPv6 connectivity to organisations to which it will assign /48s by advertising that connectivity through its single aggregated address allocation"
LIRs who operate closed/private networks appear not to qualify because the address space in these networks will not be advertised. Was this the community's intention?
Best Regards,
Laura Cobley Registration Services RIPE NCC
********************************** Madrid 2003 Global IPv6 Summit Presentations and videos on line at: http://www.ipv6-es.com This electronic message contains information which may be privileged or confidential. The information is intended to be for the use of the individual(s) named above. If you are not the intended recipient be aware that any disclosure, copying, distribution or use of the contents of this information, including attached files, is prohibited.
Hi, On Sun, Jun 20, 2004 at 01:05:19PM +0200, JORDI PALET MARTINEZ wrote:
In my opinion, closed networks today, could be connected tomorrow, and consequently advertised, so why exclude them ?
A *LIR* usually doesn't (purposely) change back and forward from "operating a non-public network" and "being connected to the Internet". End-user networks do, but the policy isn't about end-users anyway - what end-users do depends on their contractual relation to the LIR of their choice and trust. If they have no LIR available, and are not connected to the Internet, they can use non-publically-routed-global-unique IPv6 space (IIRC it was Geoff Houston's draft).
Excluding them will mean that if a network is disconnected, so no advertised, even by accident, they could miss the right for that allocation ?
Nobody will take an allocation away just because your core-router died. OTOH, if someone receives an allocation and it's not visible after two years (or any other reasonable time), one might start asking questions. Gert Doering -- NetMaster -- Total number of prefixes smaller than registry allocations: 60210 (58081) SpaceNet AG Mail: netmaster@Space.Net Joseph-Dollinger-Bogen 14 Tel : +49-89-32356-0 80807 Muenchen Fax : +49-89-32356-299
In IPv6 this could happen in the future, is true that may be not with LIRs, but with big end customers, but this big end customers will sooner or later claim for PI and sooner or later we should provide a solution for that. I think APNIC is working in something similar to "intermittently disconnected networks", but may be I'm wrong. Anyway the point probably is, as indicated in my previous email, to facilitate the allocations and not being restrictive regarding time or number of customers, but more looking at the real market trends. Difficult to measure ? not really if we look to the average situation and then we take a non-drastic approach, at least not initially with those not announcing their prefix even when the majority of the market is already doing so. They will lose the customers and probably kill themselves ... so if they are smart will not do so. Regards, Jordi ----- Original Message ----- From: "Gert Doering" <gert@space.net> To: "JORDI PALET MARTINEZ" <jordi.palet@consulintel.es> Cc: <address-policy-wg@ripe.net> Sent: Monday, June 21, 2004 6:09 PM Subject: Re: [address-policy-wg] IPv6 Policy Clarification - Initial allocation criteria "c)"
Hi,
On Sun, Jun 20, 2004 at 01:05:19PM +0200, JORDI PALET MARTINEZ wrote:
In my opinion, closed networks today, could be connected tomorrow, and consequently advertised, so why exclude them ?
A *LIR* usually doesn't (purposely) change back and forward from "operating a non-public network" and "being connected to the Internet".
End-user networks do, but the policy isn't about end-users anyway - what end-users do depends on their contractual relation to the LIR of their choice and trust. If they have no LIR available, and are not connected to the Internet, they can use non-publically-routed-global-unique IPv6 space (IIRC it was Geoff Houston's draft).
Excluding them will mean that if a network is disconnected, so no advertised, even by accident, they could miss the right for that allocation ?
Nobody will take an allocation away just because your core-router died.
OTOH, if someone receives an allocation and it's not visible after two years (or any other reasonable time), one might start asking questions.
Gert Doering -- NetMaster -- Total number of prefixes smaller than registry allocations: 60210 (58081)
SpaceNet AG Mail: netmaster@Space.Net Joseph-Dollinger-Bogen 14 Tel : +49-89-32356-0 80807 Muenchen Fax : +49-89-32356-299
********************************** Madrid 2003 Global IPv6 Summit Presentations and videos on line at: http://www.ipv6-es.com This electronic message contains information which may be privileged or confidential. The information is intended to be for the use of the individual(s) named above. If you are not the intended recipient be aware that any disclosure, copying, distribution or use of the contents of this information, including attached files, is prohibited.
Hi, on your other wish for clarification: On Tue, Jun 15, 2004 at 05:36:28PM +0200, Laura Cobley wrote:
"To qualify for an initial allocation of IPv6 address space, an organisation must [...] plan to provide IPv6 connectivity to organisations to which it will assign /48s by advertising that connectivity through its single aggregated address allocation"
LIRs who operate closed/private networks appear not to qualify because the address space in these networks will not be advertised. Was this the community's intention?
At the time this was written, "site-local" addresses where considered as the solution for these networks. Since then, they have mostly been deprecated, but a new solution (draft-ietf-ipv6-unique-local-addr-04.txt) seems to be in sight. I've seem some comments to the extent of "we must be very liberal here". The more interesting question is: does it make any difference? Of course there are large numbers of enterprises that operate closed networks, but does anyone have numbers about the number of *LIRs* that purposely and permanently do not connect their PA-allocated network blocks to "the Internet", while still paying yearly RIR membership fees? Gert Doering -- NetMaster -- Total number of prefixes smaller than registry allocations: 60210 (58081) SpaceNet AG Mail: netmaster@Space.Net Joseph-Dollinger-Bogen 14 Tel : +49-89-32356-0 80807 Muenchen Fax : +49-89-32356-299
On Thu, Jun 24, 2004 at 10:11:47PM +0200, Gert Doering wrote:
LIRs who operate closed/private networks appear not to qualify because the address space in these networks will not be advertised. Was this the community's intention? At the time this was written, "site-local" addresses where considered as the solution for these networks. Since then, they have mostly been deprecated, but a new solution (draft-ietf-ipv6-unique-local-addr-04.txt) seems to be in sight.
Which is a good thing. Site-local would bring up the same problems, RfC1918 space did (to organisations connecting to each other, both using 10/8 => Double NAT). With globally unique local addresses this would be solved. I just did a quick read of the document (did not have the time for slow reading), but I think the most interesting question still open is "how do I get my globally unique local address prefix". But that concept would sort out a lot of the existing problems we have with IPv4 today.
Of course there are large numbers of enterprises that operate closed networks, but does anyone have numbers about the number of *LIRs* that purposely and permanently do not connect their PA-allocated network blocks to "the Internet", while still paying yearly RIR membership fees?
I do not have real numbers, but know a few providers doing that. When they manage their customers networks, they use globally unique addresses for that. Currently this can only be done with IPv4 PA/PI space, as RfC1918 addresses are not unique. So they use addresses out of their assignment to have guaranteed uniqueness. Also I know of some of the big customers of ours doing it. Reality is though, mostly they assign the complete block and then just nullroute it, if traffic comes from the outside, as I think of it. So the block is is announced, but not reachable, though not "connected to the internet". But if globally unique local addresses (can I get an acronym for that, that is way too long ... and GULA even sounds nice) become reality this would be an alternative, as uniqueness seems to be the only reason for using these addresses. Nils -- Schützt ungeborenes Leben -- esst weniger Obst
On Thu, Jun 24, 2004 at 05:31:20PM -0400, Nils Ketelsen wrote:
Reality is though, mostly they assign the complete block and then just ^^^^^^
This should read "announce" Nils -- "Kommt Schrot kommt Not" [Torfrock in dem epochalen Werk 'Wildsau']
participants (9)
-
Gert Doering -
Hans Petter Holen -
Hans Petter Holen -
JORDI PALET MARTINEZ -
Kurt Erik Lindqvist -
Laura Cobley -
Nils Ketelsen -
Sascha Luck -
Thomas Narten