Hello All, one tiny comment, I would recommend making the deviceid attribute optional by setting multiplicity to '0..1'. This would allow us to think of it as resizing (of the OS image) or attaching a new block device. This way it wouldn't be dependent on a particular implementation of the concept. What do you think? Cheers, Boris On Wed, Apr 1, 2015 at 3:34 PM, Boris Parak <xparak@mail.muni.cz> wrote:

On behalf of Michel Drescher <michel.drescher@egi.eu>: ---------- Forwarded message ----------

Dear OCCI WG,

Further to our conversations please find attached a re-formatted profile document suitable for the recommendations track.

I have added the two co-authors in cc:, as well as the CloudWatch Technical Coordinator since the work has been partially funded by said project (see also acknowledgements in the document).

Although I have registered at redmine.ogf.org as OCCI WG member I am not sure that that also implies membership in the mailing list (the respective tab in redmine is empty). Boris, in case this Email gets bounced, may I ask you to submit on my behalf until the ML membership is sorted out?

In terms of contents, I have taken Boris' comments into account regarding ephemeral storage size and deviceid to make implementing it easier (see sections 7.1 and 7.2 for further details).

Cheers, Michel

-- Stichting European Grid Initiative (EGI.eu) http://www.egi.eu Mobile:+31 (0)6 303 726 55 Skype: michel.drescher.egi

Entropy isn't what it used to be...