[occi-wg] Storage Verbs (was Re: OCCI Dashboard)

[Sam Johnston]

On Thu, Jun 25, 2009 at 4:51 PM, Gary Mazz <garymazzaferro at gmail.com> wrote:

>
> If we move to adding life cycle management to storage, it will open a can
> of worms. If anything, we should adopt a quality of service metric for
> storage and let the provider figure out the best way to implement.
>
> Items like capacity management, attaches, disconnects and failure
> management are OS features out of scope for this spec.
>

I generally agree (especially about having "what" rather than "how"
specifications), but people are going to need to have levers to pull which
will largely depend on the implementation. We only need to lock down that
which we care about for interop right now (much of which is already handled
by CRUD operations).

Sam


> Sam Johnston wrote:
>
>> On Thu, Jun 25, 2009 at 3:45 PM, <shlomo.swidler at gmail.com <mailto:
>> shlomo.swidler at gmail.com>> wrote:
>>
>>    2. Amazon has distinct nouns for "Volumes" and "Snapshots". Amazon's
>>    snapshots are not mountable, you can only "copy"/"clone" (not sure
>>    which OCCI verb is the correct one here) an Amazon EBS snapshot into a
>>    new EBS drive. The OCCI API seems to imply that the result of the verb
>>    "image" on a drive yields another drive. So there may be a need to
>>    introduce either a new noun "drive snapshot" that is not linkable to
>>    any compute noun, or to otherwise differentiate between drives that
>>    are mountable and drives that are not mountable.
>>
>>
>> We plan to expose a basic set of well-defined nouns which implementors are
>> able to use in an interoperable fashion. This will include basic
>> functionality like clone (possibly with parameters like "full" vs "cow") but
>> things like "format", "check", "backup", etc. may or may not be specified.
>>
>> Note that a "mount" is currently done by creating an association between a
>> compute and a storage resource (specifying the local identifier - eg "sda").
>> I can't think of a sensible way to advertise "mountability" (short of
>> throwing errors when it fails), and having an "attach" verb on storage
>> devices which accepts a compute device as a parameter feels fugly.
>>
>> Sam
>>
>> ------------------------------------------------------------------------
>>
>> _______________________________________________
>> occi-wg mailing list
>> occi-wg at ogf.org
>> http://www.ogf.org/mailman/listinfo/occi-wg
>>
>>
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://www.ogf.org/pipermail/occi-wg/attachments/20090625/a4528082/attachment.html