Open Stack

On 7 September 2015 at 03:34, Sean M. Collins <sean at coreitpro.com> wrote:
> On Sun, Sep 06, 2015 at 04:25:43PM EDT, Kevin Benton wrote:
>> So it's been pointed out that http://169.254.169.254/openstack is completed
>> OpenStack invented. I don't quite understand how that's not violating the
>> contract you said we have with end users about EC2 compatibility under the
>> restriction of 'no new stuff'.
>
> I think that is a violation. I don't think that allows us to make more
> changes, just because we've broken the contract once, so a second
> infraction is less significant.

I see the OpenStack part of the metadata service a different
interface, that happens to be accessed in a similar way to EC2.

>> If we added an IPv6 endpoint that the metadata service listens on, it would
>> just be another place that non cloud-init clients don't know how to talk
>> to. It's not going to break our compatibility with any clients that connect
>> to the IPv4 address.
>
> No, but if Amazon were to make a decision about how to implement IPv6 in
> EC2 and how to make the Metadata API service work with IPv6 we'd be
> supporting two implementations - the one we came up with and one for
> supporting the way Amazon implemented it.

Yes, thats the cost of moving first.
Honestly, I would assume we end up implementing two access routes, if
we support IPv6 first.

Thanks,
John