Dear Sir/Madam IPv6 is one of the key research areas within NC3A and we have come across an issue on addressing. I have presented my questions to SURFnet ( which you can find below ) Since Surfnet has not yet presented their views yet I continued my search. I came acros the ECP IPv6 portal and a presentation by Rumy Kanis from RIPE NCC. http://www.ecp.nl/ipv6/docs/Rumy_Kanis_27apr05#1 which refers to IPv6 plicy and this seems relevant to our problem. I am wondering if you can lead me to the right channel for some more guidance on the issue of global IPv6 addres assignments for large geographical distributed organisations like NATO , We are also curious in which database tools are used by RIRs ? Could you help me in finding the right sources or people to get more clearness on this issue so we can further expand our IPv6 work. Regards Aad van der Zanden.
Date: Fri, 05 Aug 2005 09:34:43 +0200 To: Jasper Distelbrink <jasper.distelbrink@surfnet.nl>,Rogier Spoor <rogier.spoor@surfnet.nl> From: Aad van der Zanden <Aad.van.der.Zanden@nc3a.nato.int> Subject: Re: [AA] Aanvraag IPv6 adressen Cc: selm,jan-willem.smaal@nc3a.nato.int,levente.janovszki@nc3a.nato.int,goode,martin.gerritsen@nc3a.nato.int
Dear Jasper /Rogier
In March '05 our first official IPv6 block has been assigned to us; 2001:0610:0B20::/48. As it turns out, the policy is that even any home user gets a /48 assigned as a minimum. One starts wondering whether a /48 is then a proper size if one is allocating v6 addresses to an organisation that has currently approx 500 people as well as numerous pilot networks spread over several locations.
So after a review on the IPv6 address assignment ( and tryiing to sqeeuze all in the /48 ) we have come across a much broader view on assignment of IPv6 address space. Our Agency takes part or iniitated numerous networking initiatives that are ( or becoming ) pilots to larger scale implementations. These network can have external connectivity ( sometimes exceeding international borders as well). Next to this these network may have several security classifications levels/layers.
As an example: We have a network called NGCS. It now simulates only 5-7 Headquarter locations; the network is built upon two or more security levels/layers. So every "Headquarter" has at least these layers. - An "open layer" that is still a closed IP environment but may be using the internet as bearer one day - still it needs IPv6 capability now - a closed encrypted ( by means of IP crypto ) layer of networks - still it requires IPv6 capability now - some closed usergroups built on top by means of firewalls - also requires IPv6 capability to be studied This pilot network is the first phase to the generic NATO IPv4/IPv6 provisioning network so will scale to be a very large network in the end.
How should we approach the IPv6 address assignment for this network in your view? Could we currently apply for a larger IPv6 address space for the pilot for the open layer too from SURFnet? And what about the closed IP layer in the pilot? What if the deployment becomes a fact - who will provide the NATO IPv6 addresses if some of those Headquarters are not directly ISP connected Who is going to provide the closed layer IPv6 addresses? Since there is no private IPv6 address space, where should this address block come from ? ( so in the early pilot stage and in the "operational" stage?)
Other networks exist which are built upon (NC3A private ) infrastructure which crosses international borders to regional offices. So if we plan to deploy IPv6 on these networks too. How do we go about implementing IPv6 addressing here?
It would help if you could outline the SURFnet approach to these interesting issues. For now - can we apply for more IPv6 address space ( or a larger block ) to cover all these pilot networks ?
Regards Aad