Paul, my position regarding the publication of metrics (as well as for monitoring techniques) is that there are many, and more are going to be introduced in the future. I consider as a mistake to freze them inside a document for basic interoperability. Instead, the basic document must specifiy where and how the specific technique can be plugged in. The specific technique (in your case, push of monitoring results) can later be specified in another document: the basic remains valid while the more specific has room to evolve. Think for instance to the RTP protocol: one basic protocol, and many "profiles". The RTP protocol is still there, while multimedia formats have gone through a fast evolution. If the RTP had included the MPEG protocol, it should have been updated frequently... In the specific case of the OCCI monitoring, there is a "Publisher" entity that is the empty container for the specification of the publishing technique: there you can include the attributes that describe your specific technique as a Mixin. Of course, OCCI may later decide to publish a document to specify what are the conformant attributes for the pushing technique: would you like to write it? In attachment there are the slides of the seminar given at Zurich. I am currently elaborating a new draft, after a profitable discussion in Zurich, trying to simplify and improve consistency. I hope to release it by mid-december. I'd like to take this opportunity to announce you that in May there will be a Special Issue on Cloud Interoperability during the Closer conference in Aachen (deadline March 1st). More info: http://closer.scitevents.org/SpecialSessions.aspx Augusto 2012/11/28 Paul Christ <paul_christ@t-online.de>

** Thank you, very much and sorry, Augusto, for the delay.

I was specifically interested in 'push - issues' - which are not (yet) dealt with in your paper if my understanding is right.

Looking at papers of your colleagues (Sam Johnston, Thijs Metsch, Andrew Edmonds, Alexis Richardson etc.) from 2009 and recently in IEEE Internet Computing July/August 2012 (Vol. 16, No. 4) pp. 15-25 "Toward an Open Cloud Standard" "the OCCI working group is currently investigating asynchronous behaviors associated with service offerings. ..."; in the following also the Advanced Message Queuing Protocol (AMQP) is mentioned.

Was this on the agenda of your meeeting (last?) week? Unfortunately, the link http://tinyurl.com/1311-occi-ciuffoletti-pdf was not accessible for me. In summary I (still) have problems to envision a monitoring infrastructure which includes both pull and push elements in an OCCI context .

I will try to follow (and possibly comment on) your work and would be gratefull for any further information.

Yours

Paul

------------------------------ *Von:* augusto.ciuffoletti@gmail.com [mailto: augusto.ciuffoletti@gmail.com] *Im Auftrag von *Augusto Ciuffoletti *Gesendet:* Montag, 8. Oktober 2012 10:11 *An:* paul_christ@t-online.de *Betreff:* Fwd: OCCI-Monitoring paper

Hi Paul,

I forward you the mail I sent to the group a couple of weeks ago, with the document in attachment. Feel free to contact me if you need so. I do not plan any advancement on the topic during next weeks, so you have plenty of time...

Augusto

---------- Forwarded message ---------- From: Augusto Ciuffoletti <augusto@di.unipi.it> Date: 2012/9/20 Subject: OCCI-Monitoring paper To: "occi-wg@ogf.org" <occi-wg@ogf.org> Cc: "Metsch, ThijsX" <thijsx.metsch@intel.com>

Hi all,

enclosed you find the 7 pages paper about OCCI-Monitoring.

While working at the OCCI-SLA/Monitoring draft, I realized that the whole issue was too much for a single document. The reason to modularize it is the same as for modularizing software: in a word, monolithic is bad.

So I tried to describe monitoring in OCCI terms: there is a section that explains how this model interacts with the SLA, and with infrastructure resources, and the interface between them is sufficiently small to claim that the modularization was successful.

My suggestion for managing the collaboration on this paper (whose source is in LaTeX, but that I send you as a pdf) is by email and annotations: there are many free tools that enable to annotate a pdf. I suggest "xournal" or "okular", that are supported by many platforms. You can annotate and send me the file that contains the annotations (.xoj for xournal). If you send your annotations to the list others --- using the appropriate tool --- can see your annotations. There are LaTeX collaboration platforms, but I think they are not useful at this stage.

Augusto

2012/9/18 Augusto Ciuffoletti <augusto@di.unipi.it>

...

Before the end of this week I'll submit to the maillist a paper about SLA/monitoring. I considered that the whole SLA+monitoring was too much for a single document, and I concentrated on the monitoring part: the connection with SLA is explicit, but the paper covers a generic monitoring infrastructure, not specific for SLA. So the paper is not syntactically derived from the old one, but it is aligned with it and should hopefully give a sound basis for discussion.

Further, I tried to adhere to the revision of the "core" document that circulated in the maillist.

Bye

Augusto

2012/9/18 Metsch, ThijsX <thijsx.metsch@intel.com>

...

Hi all,****

****

There will be no call today as many of us are double and or tripple booked today.****

****

Next week we’ll pick up the calls again too hopefully finalize the JSON and SLA/Monitoring discussions.****

****

Cheers,****

****

-Thijs****

------------------------------------------------------------- Intel Ireland Limited (Branch) Collinstown Industrial Park, Leixlip, County Kildare, Ireland Registered Number: E902934

This e-mail and any attachments may contain confidential material for the sole use of the intended recipient(s). Any review or distribution by others is strictly prohibited. If you are not the intended recipient, please contact the sender and delete all copies.

_______________________________________________ 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)

-- Augusto Ciuffoletti Dipartimento di Informatica Università di Pisa 56100 - Pisa (Italy)

-- Augusto Ciuffoletti Dipartimento di Informatica Università di Pisa 56100 - Pisa (Italy)

-- Augusto Ciuffoletti Dipartimento di Informatica Università di Pisa 56100 - Pisa (Italy)