[one-ecosystem] Libcloud Update

Tue Nov 15 04:19:30 PST 2011

Updated.
Andy
andy.edmonds.be

On Tue, Nov 15, 2011 at 12:15,  <florian.feldhaus at tu-dortmund.de> wrote:
>
>
> Am 15.11.11 12:28 schrieb "Daniel Molina" unter <dmolina at opennebula.org>:
>
>>Hi Florian and Hutson,
>>
>>On 13 November 2011 19:24, Hutson Betts <hut101 at tamu.edu> wrote:
>>> Dear Florian,
>>>
>>> You may have to help me understand a little more about OpenNebula's
>>> implementation of the OCCI Draft Specification.
>>>
>>> You mention a version 1.0 and a version 1.1 of the OCCI Specification,
>>> however, all OpenNebula documentation refers to " latest draft of the
>>> OGF OCCI API specification", not a version number. Therefore, I've
>>> always assumed each version of OpenNebula simply updated the OCCI server
>>> to use the latest version of the draft specification.
>>
>>Essentially, as Florian points out, the OpenNebula distribution
>>includes a OCCI implementation based on one of the first drafts (0.8)
>>of this specification, including some extensions, requested by the
>>community, to support OpenNebula specific functionality.
>>
>>I have updated the documentation in order to clarify this:
>>http://www.opennebula.org/documentation:rel3.2:occicg
>>http://www.opennebula.org/documentation:rel3.2:occiug
>>http://www.opennebula.org/documentation:rel3.2:occidd
>>
>>The implementation of the latest draft (1.1) is being developed by
>>Florian's team in a ecosystem project:
>>http://www.opennebula.org/software:ecosystem:occi
>
> Thank you for updating the pages!
>
>
>>>
>>> Furthermore, the OCCI Working Group is a little vague about the current
>>> version of the specification. On their Specification page, they only
>>> list the latest version of their documents, as indicated by their GFD
>>> numbers. It was only a blog post which allowed me to realize that GFD
>>> 185 is actually OCCI 1.1.
>
> This could indeed be clarified by the OCCI WG. I set the Admin (Andy
> Edmonds) for the website on CC so he may update the page.
>
>>>
>>> DISCLAIMER: I don't speak for Libcloud's wonderful community. I speak
>>> only for myself, and only based on my experience to date based on my
>>> work with the greater Libcloud community.
>>>
>>> From the Libcloud perspective, we target only one interface of the
>>> OpenNebula toolkit, it's OCCI interface. However, we treat that
>>> interface as an OpenNebula-specific interface, and not as a generic
>>> interface specification for interacting with arbitrary cloud providers.
>>> Therefore, the OCCI specification, as implemented by OpenNebula, is
>>> packaged as an OpenNebula driver within Libcloud.
>>
>>As Hutson points out, the implementation of the Libcloud and
>>Deltacloud drivers were OpenNebula-specific due to the fact that the
>>OCCI specification was not finished in that moment and the OpenNebula
>>implementation was the only one.
>>
>>>
>>> As for your specific question, we do not target the OCCI 1.0
>>> specification, as denoted by the OCCI Working Group. The reason we do
>>> not target OCCI 1.0 is because OpenNebula does not have a consistent
>>> implementation of the OCCI specification across versions. Rather
>>> OpenNebula 1.4 and OpenNebula 2.0+ use the same semantics to describe
>>> networks but different semantics to describe compute nodes. Therefore,
>>> the OpenNebula driver had to be written per OpenNebula version and not
>>> per OCCI version.
>>>
>>> It would be wonderful in the future if Libcloud could support an OCCI
>>> driver built to a solid specification for use by OpenNebula, ..., and
>>> other cloud providers. If you would like to work on a joint effort to
>>> support an OCCI 1.1 specification-derived OpenNebula interface, and
>>> OCCI-derived Libcloud driver, I would be happy to participate.
>>
>>I also think that it would be interesting to develop a new driver
>>exposing the last version of the OCCI interface, in order to interact
>>with OCCI-compliant clouds, so that we can test interoperability among
>>different implementations
>
> I would be glad to assist in writing such a driver. Maybe we can find a
> timeslot in the next days where we go through the current implementation
> and identify what needs to be changed and how we could proceed. If so,
> what time would fit you best and who else wants to join?
>
>>
>>Kind regards
>>
>>>
>>> --
>>> Hutson Betts
>>> Computer Science and Engineering
>>> Texas A&M University
>>>
>>>
>>> On Sun, 2011-11-13 at 13:25 +0000, florian.feldhaus at tu-dortmund.de
>>> wrote:
>>>> Hi Hutson,
>>>>
>>>> great news! Do you use the old OCCI 1.0 API? If so, could we work
>>>>together
>>>> to support OCCI 1.1 using the OCCI 1.1 for OpenNebula Ecosystem project
>>>> (http://dev.opennebula.org/projects/ogf-occi)?
>>>>
>>>> Cheers,
>>>> Florian
>>>>
>>>> Am 09.11.11 19:46 schrieb "Hutson Betts" unter <hut101 at tamu.edu>:
>>>>
>>>> >Libcloud has released version 0.6.1 (Release announcement:
>>>> >http://mail-archives.apache.org/mod_mbox/libcloud-users/201111.mbox/%
>>>>
>>>>>3CCAJMHEm+8XX704mSY4qw4P0YSBjGK=0SWCKjzSHBe8sLD__2UnA at mail.gmail.com%3E
>>>>>)
>>>> >which supports compute node management for OpenNebula 1.4 through
>>>> >OpenNebula 3.0 using the OCCI API.
>>>> >
>>>> >Full changelog for the new OpenNebula support can be seen here:
>>>>
>>>>>https://svn.apache.org/viewvc/libcloud/trunk/CHANGES?revision=1198753&v
>>>>>iew
>>>> >=markup
>>>> >
>>>> >Furthermore, these changes have been tested in my own production
>>>> >environment.
>>>> >
>>>> >--
>>>> >Hutson Betts
>>>> >Computer Science and Engineering
>>>> >Texas A&M University
>>>> >
>>>> >
>>>> >_______________________________________________
>>>> >Ecosystem mailing list
>>>> >Ecosystem at lists.opennebula.org
>>>> >http://lists.opennebula.org/listinfo.cgi/ecosystem-opennebula.org
>>>>
>>>
>>> _______________________________________________
>>> Ecosystem mailing list
>>> Ecosystem at lists.opennebula.org
>>> http://lists.opennebula.org/listinfo.cgi/ecosystem-opennebula.org
>>>
>>>
>>
>>
>>
>>--
>>Daniel Molina
>>Project Engineer
>>OpenNebula - The Open Source Toolkit for Data Center Virtualization
>>www.OpenNebula.org | dmolina at opennebula.org | @OpenNebula
>
>