Re: [v6ops] New version of draft-lopez-v6ops-dc-ipv6

["Diego R. Lopez" <[e-mail]>]

Hi Wes,

Thanks for the comments. A few more detailed replies inline

On 26 Oct 2012, at 16:54 , George, Wes wrote:
> I've read this version of the draft, and I have a few comments.
>
> I would not refer to your last transition stage as "Next Generation" as that is not a particularly descriptive term for the stage that you describe, and it is also a fairly overloaded term - my "next generation" DC may well be the one where I increase the bandwidth, or build out a fabric, start using some new widget, or whatever. I would maybe call it IPv6-only or Single-stack IPv6, since that is really the desired end state, where as much of the DC as possible is running IPv6 for all functions, and IPv4 connectivity is handled via transition and translation technologies at the edges of the DC (eg loadbalancers, etc).

It makes sense, though I can understand that using something we could call
an euphonic name would help in sending the message that's the desirable
stage to achieve. I'd suggest to have some live discussion about this in
Atlanta...

> Section 2.3 should mention the fact that this stage may be driven by either a lack of enough IPv4 resources (whether private or globally unique) or a need to reclaim IPv4 resources from portions of the network which no longer need them. There is a point at which dual stack is simply not possible anymore, and once that point has been reached, a careful evaluation of what still needs to speak IPv4 and what does not will need to happen to ensure judicious use of the remaining IPv4 resources.

Agreed. Mentioning this will add more weight to the reasons to achieve this
stage. Will update the document in this respect.

> It might also be helpful to discuss the different classes/categories of things that are in an average data center, as those may have a bearing on how the transition proceeds, and indeed may be at different phases of the transition at different times:
>
> Management systems (provisioning, alarms/PMs, software distro, etc)
> Fabric (and perhaps even discuss the different layers, ToR, agg, core, etc)
> Hypervisor (though this is already mentioned in a few places in the draft)
> Machine-to-machine communications (one application talking to another over an API)
> End-user applications/content
> External services
> Etc
>
> The key here is to keep it generic enough so that it's widely applicable, rather than getting so specific that you have to have multiple different scenarios to represent different potential DC applications and designs.

This is a good point, though I think we should agree in limiting the number
of such items, and clearly identify how the application of v6 affect them and
in which respect they are specific to the DC environment. Certainly
management systems are, as considerations on the fabric and hypervisors.
In the case of end-user applications and content I think they are mostly considered in other operational guideline: (draft-chkpvc-enterprise-incremental-ipv6) and we could follow an approach similar to the used in
the security considerations.

When it comes to M2M and external services, I'd like to have a clearer idea
of what you are referring to...

> Security considerations: draft-vynke-opsec has been replaced by draft-ietf-opsec. However...

Right. Will correct.

> The referenced draft covers a lot of ground, and simply referring to it by itself might be daunting to potential readers. It might be worth highlighting specific sections that are especially pertinent to DCs, such as the discussion of ND cache exhaust, etc.

That certainly makes sense. Will take care of it for the coming version

Be goode,

--
"Esta vez no fallaremos, Doctor Infierno"

Dr Diego R. Lopez
Telefonica I+D
http://people.tid.es/diego.lopez/

e-mail: [e-mail]
Tel:    [phone]
Mobile: [phone]
-----------------------------------------


________________________________

Este mensaje se dirige exclusivamente a su destinatario. Puede consultar nuestra política de envío y recepción de correo electrónico en el enlace situado más abajo.
This message is intended exclusively for its addressee. We only send and receive email on the basis of the terms set out at:
http://www.tid.es/ES/PAGINAS/disclaimer.aspx