Re: [dc] draft-khasnabish-vmmi-problems-00.txt

[Truman Boyes <[e-mail]>]

On Tue, Jan 31, 2012 at 10:56 PM, <[e-mail]> wrote:

>
> Thomas, thank you for your comments!
>
> About VM migration in the draft, we did not demand forcibly between the
> different vender(or SPs), so your concerns may not
> exist. As you say, "let's be realistic", our intention is to improve the
> flexibility of VM migration, as well as the breadth
> of applications under the premise of market heavyweights are not opposed
> to it.In order to achieve these goals, analyse possible problems, discuss
> and resolve these problems, such as: VM migration is due to a non-public
> sector energy-efficient needs, rather than to public access without
> interruption, or business needs of the user's desktop migration, this
> demand may exist within a same service provider, or you say a same vender
> with mixed network, for example:
>
> With the promotion of IPv6 technology, the existing IPv4 networks will be
> more and more IPv6 hosts, these applications driven
> a series of tunnel technologies to provide solutions, such as: 6to4 tunnel
> technology, ISATAP tunnel technology, and so on.
> Virtual machine migration technology will also be the basis of these
> network environments,in the transition network using tunneling transition
> technique, the connections between the subnets and the backbone network are
> achieved through the tunneling gateway. In the IPv4/IPv6 transition period,
> a variety of tunnels coexist. The establishment of the tunnel varies with
> different gateways. The traditional tunneling gateway only establishes
> tunnels for communication with the same type of gateway, the different
> types of traditional tunneling gateway cannot communicate with each other,
> which cannot meet the requirements of VPN communications in the transition
> period. A multi-tunnel VPN gateway is used to solve the problem of
> establishing the tunnel between the heterogeneous gateways.
>
> Many thanks for guidance
>
> Regards,
>
> Bin Liu
>
> [e-mail]
> [e-mail]
>
>
>
I understand the first section in the above comments regarding virtual
machines. In the context of IPv6 adoption for virtual machines or rather
generically in data centers, I see native IPv6-only or dual stack as being
the choice for numbering VMs. The tunneling technologies mentioned (6to4,
ISATAP, etc) should be kept out of data centers.

-- 
--truman