Hi everyone First let me apologize for letting things slide for a while; I've been up to my ears in the project I work on. We need to start work on designing the survey to gather experiences with the UR 1.0 format. I'm not quite sure how to go about this, but I suppose a good way to start would be to provide some kind of template for people to fill in. So I'm going to write some random ideas I have for what we should be asking; chime in if you think I'm right or wrong. Name & Contact Email Project/Product Prose Description of Use Structured Use Characterization: Which UR1.0 Elements do you support? (Table, one row per spec element) Element Name | Generate | Parse -------------+----------+------- UsageRecord | yes | yes .... What extensions do you put in a Usage Record? e.g. executable name, working directory, arguments When do you generate usage records? At job start. At job end. At points during the job. What triggers the generation? After the job has run from other captured data. Corresponding to whole job? Corresponding to events within job? What do you do with Usage Records after generating them? Is there anything else that we ought to ask about? I should note here that the OGSA HPC Profile WG is looking into the collection of usage data from simple jobs, and that they are minded to use UR1.0 as the basis for this. Donal.
Hello all: The attachment summarises JISC review results in terms of UR 1.0. I hope the document would help us establish the initial idea on UR 2.0 profile and survey design in order to obtain detailed information. Remember the survey design aims at collection of comments on Usage Record 1.0 rather than JISC's scope in Grid accounting and Usage monitoring. Therefore the summary attachment does not give enough information for UR 2.0 profile. We need our own survey!!! But the summary information does share same issue we might enounter during interview. Here are my comments on Donal's email:
Hi everyone
First let me apologize for letting things slide for a while; I've been up to my ears in the project I work on.
We need to start work on designing the survey to gather experiences with the UR 1.0 format. I'm not quite sure how to go about this, but I suppose a good way to start would be to provide some kind of template for people to fill in. So I'm going to write some random ideas I have for what we should be asking; chime in if you think I'm right or wrong. Name & Contact Email Project/Product Prose Description of Use
I suggest that we need establish agreement on survey scope. There are many Grid or cluster projects that require usage logging and statistics, which may use Usage Record 1.0 with extensions or custom data model. The objective of the survey is then accordingly need to be clarity on "experiences on usage of UR 1.0" or "usage representation requirements". Things can be even more complex when coming to data service providers (see attchment section 1.6). The usage metrics are application/service-specific and totally differentiates from UR 1.0.
Structured Use Characterization: Which UR1.0 Elements do you support? (Table, one row per spec element)
Element Name | Generate | Parse -------------+----------+------- UsageRecord | yes | yes ....
It is a good idea to use above table to degree importances of various usage properties defined in UR 1.0 and highlight extensions on custom properties. Again the problem appears when a project does not use UR 1.0 at all for other usage logging purpose, e.g. WLCG/EGEE storage accounting schema, which does fall into the scope of UR 2.0. Besides, the UR 1.0 concentrates on Batch job level usage representation. Therefore using a single table does not give us information of other-level usage representation (e.g. a complex job, summary, and service). How do you think put one more option here allowing interviewees to put their custom schema or desired usage metrics if they are not using UR 1.0 at all. In this case, the WLCG/EGEE could put their storage schema and summary schema as well.
What extensions do you put in a Usage Record?
e.g. executable name, working directory, arguments
Good point! or we could put extensions into above question (Which UR1.0 Elements do you support? ) with UR 1.0 extension properties?
When do you generate usage records?
At job start. At job end. At points during the job. What triggers the generation? After the job has run from other captured data. Corresponding to whole job? Corresponding to events within job?
What do you do with Usage Records after generating them?
Is there anything else that we ought to ask about? I should note here that the OGSA HPC Profile WG is looking into the collection of usage data from simple jobs, and that they are minded to use UR1.0 as the basis for this.
could you please forward us the link (if any) to us ? X. Chen BITLab School of Engineering and Design Brunel University, Uxbridge Middlesex, UB8 3PH London mobile: ++44(0)7871876894
Donal K. Fellows wrote:
We need to start work on designing the survey to gather experiences with the UR 1.0 format.
Thanks for everyone's feedback. Here's my first attempt at writing the survey form properly. (If you want to edit it properly as opposed to just filling it in, there's no password set on the protection.) Feedback welcome. Donal.
Hi Donal! Do you plan to start the survey at OGF22? I won't be able to come, but I would send you a compiled version of the survey to let you know what we do in case of the DGAS accounting system. We have developed a prototype RUS interface within the OMII-Europe project, that of course is based on the UR. And we'd like to feedback our experiences to the UR-WG. I know SGAS and Unicore have also implemented RUS interfaces and hence worked with the UR. We might forward the survey to them as well (in case they're not ion this ML or don't come to OGF22). Let me know. Cheers, Rosario. Donal K. Fellows wrote:
Donal K. Fellows wrote:
We need to start work on designing the survey to gather experiences with the UR 1.0 format.
Thanks for everyone's feedback. Here's my first attempt at writing the survey form properly. (If you want to edit it properly as opposed to just filling it in, there's no password set on the protection.)
Feedback welcome.
Donal.
------------------------------------------------------------------------
-- ur-wg mailing list ur-wg@ogf.org http://www.ogf.org/mailman/listinfo/ur-wg
Rosario Michael Piro wrote:
Do you plan to start the survey at OGF22?
I think so. Seems as good a time as any. :-)
I won't be able to come, but I would send you a compiled version of the survey to let you know what we do in case of the DGAS accounting system. We have developed a prototype RUS interface within the OMII-Europe project, that of course is based on the UR. And we'd like to feedback our experiences to the UR-WG.
I know SGAS and Unicore have also implemented RUS interfaces and hence worked with the UR. We might forward the survey to them as well (in case they're not ion this ML or don't come to OGF22).
I'm not sure what the best method of dealing with survey documents is; it really depends on how many there are. If there's a handful, emailing them to the list will work just fine. If there's dozens, we'll have to invent something else (OTOH, we'd also have a really good case for UR1 being a standard. :-)) Donal.
participants (3)
-
Donal K. Fellows -
Rosario Michael Piro -
Xiaoyu Chen