We are trying to setup IPv6 lab / testbed that will be used by companies who planning transistion to new protocol. Is there any one with suggestions or mybee real life experience. any comments are welcome.
It depends on what you are planning to test, and how. Is it operation of the applications ? Is it operation of the devices ? Is it related to user experience ? Throughput ? Manageability ? How do you foresee the typical test flow will look like ? Is it operation in dualstack environment or IPv6-only ?
Without these details, hard to say much, other than you'd get a native IPv6 routed to the testbed you are setting up - and also ensure there is a low-MTU scenario to verify the functioning of the setup in presence of tunnels.
In my home lab I have a linux box routing the IPv6 tunnel from SixXS - where its subnets either get routed to dualstack or IPv6-only segments. That works for my daily needs. Of course majority of the testing involves the functionality. For scalability/performance, one would need a different setups - but those might usually involve specialized traffic generators like Ixia/Spirent/BreakingPoint, so they are usually not done in a home lab
This testbed should help enterprises to try different transition technique and to choose the right one. Since they can not afford to do experiment in enteprise networks they will use this testbed to see, is it dual stack, NAT or some another transition technique best for them. We will try using virtualization to allow to test bussines application and different server versions in this IPv6 enviroment.
Then you probably might want to start building a lab that mimics the *current* environment of the enterprises - their address IPAM tools, applications, and processes. Once that is ready - then you can start adding various IPv6 transition mechanisms on top (BTW I assume you meant NAT64 as a transition technique - NAT by itself is merely IPv4 conservation technique, it does not transition to IPV6 at all - so it is part of dualstack).
Overall great and useful idea I think, but probably worth approaching it first from the angle of building a lab which is a good representation for current IPv4 environment in the enterprise - *then* adding IPv6 stuff to it.
You right, I did thought on NAT64. And idea is to build lab based on most common network enterprise arhitecture (DMZ, internal network, ASA). I am seeking for some advice regarding additional devices that could be used in enterprise network. Since based on project proposal this idea can be accepted, I do not want to setup lab without some important component that is missing. For applications and services ti is possible to offer cloud and virtualization system to create "copy" of bussines process, so knowledge from lab could easy be transfered to real life networks.
I am not sure that enterprises will switch their network infrastructure over night and start using new IP protocol. They prefer secure and trusted enviroment that they already builded. So we need to give them playground and after they became complitly confident, they will be ready for new protocol and transition.
I would recommend an architectural IPv6 assessment to determine your environments current support for IPv6, then once you have all of that information you will be able to evaluate your options as far as deployment models. For example, you may find you have legacy applications that do not support IPv6 and will never support IPv6 that will drive you to remain dual-stack until that application is replaced.
As Andrew stated it's impossible for us to provide you with a best practice lab other than stating that you should try and recreate your environment as best as possible, but again an architectural assessment is the first step in the process.
Depending on the size of your enterprise, my opinion is that IPv6 adoption is less technical and more of a political battle with an organizations acceptance to change and innovation. Sure the technical side of IPv6 integration is challenging, but in my experience executive buyoff and demonstrating the value is the most challenging. My point is, a test lab is fun and very informational but it will be the assessment and actual plan that will demostrate the magnitude of the project.