I've just pushed a revision of the doc trying to meet your comment: attached is the pdf. You find added content from line 112 to line 129, and in fig. 2. How is it? 2014-12-20 22:17 GMT+01:00 Gary Mazz <garymazzaferro@gmail.com>:
Hi Guys,
Its been a while..
The notification work looks great !!
Is there a capability to place notifications in a workflow/deployment topology ?
cheers, gary mazzaferro
On 9/16/2014 4:59 AM, Andy Edmonds wrote:
Guys, here a summary of the KIS approaches as discussed at OGF42
https://docs.google.com/presentation/d/1x3_exe7RD6orOhwRajFNJwXxgV1zXHIZR0WU...
Andy
Andy Edmonds Æ Senior Researcher, ICCLab Institute of Information Technology Zürich University of Applied Sciences http://www.cloudcomp.ch, @dizz
On 16 September 2014 12:17, Augusto Ciuffoletti <augusto@di.unipi.it> wrote:
I agree with you that notification should be considered as a model, as an extension that introduces new resources and links, just like IaaS Infrastructure. I guess that this gives us all the tools to produce an effective model, without the side effects of introducing new core concepts.
I do not agree about the limitations you address: this depends on how we model the relevant concepts. If we'd use mixins, you could associate them virtually to anything, but for the same reason they are weak in formulating recommendations. So it is really relevant to use the right tool for a given task.
I hope to be able to propose my idea about that by the beginning of the next week. Suggestions are always welcome.
Augusto
PS: thanks to all organizers and participants to the OGF sessions, for a very productive and exciting meeting. In these times it is hard to find one :-/
2014-09-16 10:10 GMT+02:00 Jean Parpaillon <jean.parpaillon@free.fr>:
Hi Augusto, all, Thanks for all the interesting discussions in OGF, first.
Regarding the notifications, we discussed 2 approaches: - the first include the notion of notification into the meta-model, with versions of it in each rendering/protocol - the second one add categories for a notification system, hence a model
IMO, the second approach, hence simpler, has the problem of being limited to notifications on resources. I don't see how to plug notifications on collections, links or user mixins, for instance.
Have I missed somthing ?
Another possible approach should be to consider notifications only at the level of protocol, not touching meta-model.
My opinion is we should first think about at which level we want to introduce this notion of notifications.
Regards, Jean
Le jeudi 11 septembre 2014 à 10:05 +0100, Augusto Ciuffoletti a écrit :
Dear all,
talking and listening to people here during OGF42, I perceived that there is a strong demand for a notification mechanism inside OCCI, that is able to carry an asynchronous notification about a relevant event. For instance, "this job is now running" from an hypothetical job resource, or "this VM is going to shut down in 5 seconds" from a Compute.
This demand is in fact more compelling than that about monitoring, and may be in fact integrated in OCCI-monitoring at a later time. So I'll give precedence to OCCI-Notification.
The specification is very simple: one single link ("Notification"), and two types of mixins, one for the triggering event, one for the triggered action.
Any "thumb up" or input from you is relevant, and I will submit soon a proposal about that.
-- Augusto Ciuffoletti Dipartimento di Informatica Università di Pisa 56100 - Pisa (Italy) _______________________________________________ occi-wg mailing list occi-wg@ogf.org https://www.ogf.org/mailman/listinfo/occi-wg
-- Jean Parpaillon Open Source Consultant Phone: +33 6 30 10 92 86 im: jean.parpaillon@gmail.com skype: jean.parpaillon linkedin: http://www.linkedin.com/in/jeanparpaillon/en
-- Augusto Ciuffoletti Dipartimento di Informatica Università di Pisa 56100 - Pisa (Italy)
_______________________________________________ occi-wg mailing listocci-wg@ogf.org
_______________________________________________ occi-wg mailing list occi-wg@ogf.org https://www.ogf.org/mailman/listinfo/occi-wg
-- Augusto Ciuffoletti Dipartimento di Informatica Università di Pisa 56100 - Pisa (Italy)