[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: [v4tov6transition] Some opinions about establish a new WG
Hi all,
I support Tina's proposal to have a FAQ document on IPv6 deployments.
IETF is in my understanding the right place to work on official answers to these questions, in an easily readable format.
If v6ops isn't the place to do this work, and write consensus answers to ALL questions that people involved in IPv6 deployments may ask, then a new WG is IMHO the right solution.
- Answers may be pointers to other bodies.
- Questions, shouldn't be limited to Operations-and-Management considerations; they should include some customer-oriented ones.
Regards,
RD
> Hi all,
> I totally agree with Cancan opinion. In ISP's point of view, there are so many problems we need to handle when transitioning v4 to v6, and some of these problems may be not "very" technical, like how to do IPv6 address planning and what elements should be considered, or in the case of China Telecom's network, which technics should be selected to deploy to CT's network and what elements should be considered, etc.
>
> I think the answers for these problems are very valuable for other ISPs, and worth our discussion.
>
> On Fri, Aug 27, 2010 at 12:46 PM, huang cancan <cancanhuang110@gmail.com> wrote:
> hi, folks:
> As a person who have worked in telecom operator for many years, I do appreciate the thousands of rfcs the IETF provide us and they do help us a lot when we deploy technologies in our network. However, I am here to argue that is very necessary to establish a new WG which is really focus on the operator's need and solve our problems.
>
> At first, I want to make it clear that what is the operator really need when they starting v4 to v6 transition. Why I emphasize the topic of v6 transition is because this item is quite different from any other projects. It is just like to establish a new Internet world!!! It is definitely a huge systematic work other than how to deploy DS-lite in the network. When we start this project, we first have to show our boss the migration strategies and paths, tell him what will happen at what time and what we can do to solve that problems step by step. When we start to do a thing we need a road map, isn't it? So, the real requirement of a operator to start his tour on IPv6 transition is to work out the strategies. And this is what 6ops cannot provide us because they say it is our own business problem. However, hey, the business problem is the foremost problem, isn't it? Without solve this problem, we even cannot start our tour~~~ The evidence is as below:
> ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- How to decide the scale of the IPv6 address space we need? For example, China telecom will have 100+ millions subscribers in the next 3 years, and M2M service will come soon, so we should apply for /24 or shorter?
> Since IPv6 address space is large, you can start with a longer prefix and grow. If the APNIC’s policy allows you to get a /24, you can ask for it.
>
> What kind of address allocation schemes should be deployed in enterprise networks? Provider aggregatable address, provider independent address or local address? If PA address, how to avoid renumbering when the enterprise network change site or provider? If PI address, how to reduce routing table? If local address, how to make sure all of the computers can access internet, NAT66 is a good choice or not?
> I think this is more like a business decision than technology decision. An ISP can definitely offer service to give enterprises an IP prefix from its own aggregate. However, some enterprise won’t like it because of the renumber problem you mentioned by switching provider.
>
> ---------------------------------------------------------------------------------------------------------------------------------------------------------
> Well, it is not the answer we need. What we need is what length of prefix we should apply, which is appropriate for us. You cannot suggest every operator that them can start a longer prefix...... What we need is: is there any model or formula to calculate the prefix length that we need or address architecture planning with the considerations of the subscriber scale and increase rate? And how can operators to utilize the address bits before /56 and after /56 to distinguish different type of service or different metropolitan area network in order to optimize the routing or management.
> You may said it is our own business problem,depend which one we like...... I do believe this answer can not solve the operator's problem. If 6ops continues gave us that answer, I don't believe the final guideline 6ops provide us will meet our requirements.
> If the so called business problem,which is the most important factor to make a strategy, is not concerned by 6ops, can we gather some people working in the operator who is interest about that to discuss our business problem in another WG, in which business problem will not be ignored?
>
> Can-can Huang
>
>
> _______________________________________________
> v4tov6transition mailing list
> v4tov6transition@ietf.org
> https://www.ietf.org/mailman/listinfo/v4tov6transition
>
>
> _______________________________________________
> v4tov6transition mailing list
> v4tov6transition@ietf.org
> https://www.ietf.org/mailman/listinfo/v4tov6transition