Summary on comments on "3GPP analysis" doc so far (since December 10th)

[<juha.wiljakka@nokia.com>]

 Hi all!

I have tried to collect a summary on the discussion on "3GPP analysis" document so far.

----------------------

1) Thakur
The na(p)t-pt issues (magnified in a 3gpp network) exist anywhere and everywhere when deployment of v6 gathers momentum, solution to nat-pt (dns alg) problems must be found.

2) Thakur
Not sure if manually configured tunnels provide a good solution in the years to come. Worries how to manage a great number of manually configured tunnels

=> The selection of the tunneling mechanism is up to the operator/ISP deployment scenario and we can only give generic recommendations in our draft. A basic question is how many tunnels we need to set up - if the number is not so big, it is no problem setting up manually configured tunnels in the network. It is good to remember that use of automatic tunneling (e.g. 6to4) in very big networks is not trouble-free either.
=> Some edits in the text may be needed

3) Thakur
Recommends (as in 3.4) the usage of application proxies, when possible.

=> Use of application proxies (when possible) is not a bad idea at all. But in the cases like IMS scenario 1 we need also IP level protocol translation.

4) Thakur
Worries that all tunneling / translation mechanisms in existence today are far from perfect. We need to find a transition mechanism that scales well, is usable whether there are small islands of v6 in a v4 ocean or vice-versa, and preferably can be implemented in all or at least most of the scenarios mentioned in this draft.

=> One main goal in this exercise is to see what existing tr. mechanisms are usable for 3GPP scenarios and document any gaps. Pointing out possible problems in the existing mechanisms also includes in this work. However, we are *not* going to specify any new mechanisms. We don't need just one or two mechanisms that would fit in all cases, but it would be nice to have a compact set of solutions (let's call it transition toolbox) that the operators can deploy IPv6.

5) Thakur
DNS functionality related questions (especially name space fragmentation in GPRS scenario 2)

=> comment from Jasminko Mulahusic: This is not an issue as long as you have a continuous name space. If someone is deploying ipv6-only networks, it doesn't automatically mean the name space is fragmented.
=> Brian Carpenter: AAAA records travel just fine over an IPv4 network

6) Thakur
Usage of IPv4 PDP context instead of IPv6 PDP context (by the dual stack UE) in the case the network and the peer node is IPv4.

=> Not an issue, v4 / v6 PDP contexts can be flexibly used in 3GPP networks based on application needs, availability of public IPv4 addresses is a problem.

7) Beloeil
Using dual stack solution to solve IMS scenario 1

=> This is not feasible solution according to current 3GPP specifications stating that *IMS is IPv6-only*.
- IMS is an extensive concept; when saying that IMS is "exclusively IPv6", it is a valid statement for all elements that are connected to the IMS / considered to be a part of the IMS. Both signalling and user data traffic belong to the IMS. E.g. the connections between two IMS UEs belong to the IMS. Connections can only be IPv6, not IPv4 and not IPv4 & IPv6 simultaneously. 
- Tunneling traffic (IPv4 in IPv6) through the IMS and IPv4 connection from the IMS UE to the IPv4 correspondent node is not a feasible solution for this problem.

8) Beloeil
Static tunnels become quickly boring. Would prefer a 6to4-based solution (however, the security issues are remaining).

=> Static tunneling makes sense when the number of tunnels can be kept limited.

9) Savola
Comments on separating internal / external cases

=> Based on the discussion it was agreed to more clearly separate
"inside the operator's network" and "outside the operator's network" cases in
the next revision! This is valid (at least) for GPRS scenario 2.

10) Savola
Operator backbone IPv6 deployment status (how much IPv6 is already deployed today?)

=> Many GPRS operators already have IPv4 backbone networks (gradually migrating them while introducing  IPv6 islands). And native v6 backbones are quite rare in the first phases of the transition.
=> Discussion on the actual deployment status continues and maybe some clarifying text is needed in the scenarios or analysis document

11) Savola
Applicability of [BGP] / [IGP] tunneling (scalability issues; applicable for smaller-than-Internet scopes - not applicable when making connections over the Internet). Basic recommendation for the GPRS scenario 2 is to use a couple of configured tunnels.

=> can be applicable for 'inside the operator's network' case / for the cases when operator's backbone is based on IPv4.
=> some more text / a bit clearer applicability statement is needed
=> Karim El-Malki: Applicability of those mechanisms depends on
the exact deployment and operator requirements so we can only make
recommendations. I think it is correct to recommend static or EGP/IGP
mechanisms if the operator wants to connect its v6 islands over v4.

12) Savola
What parts of the transition is done by the ISPs and what is done by 3GPP operators. How many ISP functions does the 3GPP operator wants to perform (indeed, believes that most of them are ISP's of some sort themselves)

=> add some clarifying text on the environment

13) Savola
Max 10-15 isolated IPv6 islands (to be connected using configured tunnels) in the 3GPP operator's network should be sufficient (~ reaching one limit where in some cases configured tunneling could become more difficult to maintain).

14) Laloux
[BGP] defines 2 different deployment options: the first one assumes MP-BGP implementation over IPv4, while the second one assumes MP-BGP implementation over IPv6, thus needing to rely on some existing tunneling technique (as you mentioned before). We could just consider the first approach, avoiding the need for any additional tunneling mechanisms.

15) Thakur
The transition scenarios are not 3gpp specific, but extra issues and points arise in the 3gpp case: the presence of an extraordinary number of mobile nodes accessing the Internet and blowing up of NAT-PT / DNS ALG issues.

16) Savola
The long paragraph in chapter 3.2 should be split in about 5.

---------------

I will return to this discussion in January and compose revision -01 of the 3GPP Analysis doc around January 20th.

Merry x'mas & a happy new IPv6 year 2003!

Best Regards,
		-Juha W.-