[
Date Prev][
Date Next][
Thread Prev][
Thread Next][
Date Index][
Thread Index]
[
List Home]
RE: [cosmos-dev] Status of Datacenter SML example
|
The data center use of key was not a
good one. It was a "temporary" short cut. After all the example
was not completed... We had actually come up with an important technique
for identity handling as well as we discussed this with the CMDBf folk.
I can explain separately. For consistent queriability, I would actually
advocate considering an identity or key facet. Name and description are
perhaps properties of a generic/description facet. I think an interesting
point is that different usage patterns may in fact have different "names".
In the end, even a query/get of the name is still just another access pattern
;-)
acyclic is about not allowing loops
in instances.
I would be interested in talking about
your identity problem, or should I say your problem with data identity
;-)
The use of aliases in SML-IF is the
beginning of a relatively simple pattern that can be used to deal with
morphing and merging identities by basically always putting an identity
in a cluster of identities where the cluster is a proxy to the object.
Basically this makes all "unique" identities an "alias"
to the actual object. This makes interchange much more viable.
Thanks for your time.
________________
Harm Sluiman,
IBM DE / Technical Executive
phone:905-413-4032 fax: 4920 cell: 1-647-300-4758
mailto:sluiman@xxxxxxxxxx
Admin : Queenie Lam qlam@xxxxxxxxxx Tie: 313-5864 1-905-413-5864
"Hawkins, Joel"
<Joel.Hawkins@xxxxxxxxxxxxx>
Sent by: cosmos-dev-bounces@xxxxxxxxxxx
05/27/2008 02:10 PM
Please respond to
Cosmos Dev <cosmos-dev@xxxxxxxxxxx> |
|
To
| "Cosmos Dev" <cosmos-dev@xxxxxxxxxxx>
|
cc
|
|
Subject
| RE: [cosmos-dev] Status of Datacenter
SML example |
|
Harm,
Thanks for the background on
facets. Do you see any properties of a type as being ‘intrinsic’, i.e.
not belonging to a facet? Looking at the datacenter example, the
only one I see is item:name, which is also used as an sml:key for all of
the types (rack, cable, etc. all extends item). I’m assuming that sml:key
could be applied to complex types as well as attributes? Could a facet
(or some aspect of a facet) act as an sml:key?
Can you give an example of when
you would use sml:acyclic=false? Is this for the typical bill-of-materials
scenario, or do the cycles refer to instances rather than types?
Mark,
I’m working with some internal
folks to get up a use case, but principally it comes down to being able
to import subsets of external model definitions (initially as a type system,
and then as instances), and to associate instances in my internal representation
with external sensor data streams so that I can do live impact/root cause
analysis, SLA monitoring, etc. Some of the things I’m dealing with
are identity discrimination (basically identifying entities by sets of
properties and/or relationships to other entities), model mappings (we
attach to many technologies with different “world views”), time-series
identification, and concern separation (which is where the facet/capability
stuff comes into play). My life would be made easier by something
like CML, particularly if it had the following characteristics: I’d
like to see a mechanism for specifying identity equivalence in the model
(I think this is already do-able with SML). I’d also like to see a way
to add semantic hints (in an agreed upon fashion) to property definitions
in the model, so that I could use the model to drive sensor identification
and configuration and to decide how deal with the sensor data. I
think this could be easily done using schema annotations, similar to the
way the sml:key information is handled.
Cheers,
Joel
-----Original Message-----
From: cosmos-dev-bounces@xxxxxxxxxxx [mailto:cosmos-dev-bounces@xxxxxxxxxxx]
On Behalf Of Harm Sluiman
Sent: Tuesday, May 27, 2008 9:04 AM
To: Cosmos Dev
Subject: Re: [cosmos-dev] Status of Datacenter SML example
facets were intended to create what would in effect be change units of
work based on access patterns and roles. They were intended to allow overlapping
(however avoid it if at all possible), but basically no content could be
enabled without being in at least one facet. This is sort of, but not exactly
like capabilities.
In SML we discussed the concepts of requirements and capabilities as the
data driven constraint mapping to associated artifacts/resources. Capabilities
are basically the visible properties of a resource that can be consumed
or used as a match to a requirement. For example a "blade" may
surface the memory it has, and the operating system may declare the amount
of memory it requires. The genics constraint tests/maps the requirement
of 500 meg to a capability of 2 gig.
The intent is that a minimal number of basic Schematron constraints could
be provided that dynamically matched requirements to capabilities. This
in turn implies that they be specified in a rigorous fashion and be relatively
strongly typed. Grouping of capabilities into facets would greatly aid
this scenario.
So facets basically were intended to provide a grouping mechanism to optimize
access patterns and reduce searching. This did not make it into the SML
specification, and is now proposed for CML.
I hope this helps.
Thanks for your time.
________________
Harm Sluiman,
IBM DE / Technical Executive
phone:905-413-4032 fax: 4920 cell: 1-647-300-4758
mailto:sluiman@xxxxxxxxxx
Admin : Queenie Lam qlam@xxxxxxxxxx Tie: 313-5864 1-905-413-5864
Mark D Weitzel <weitzelm@xxxxxxxxxx>
Sent by: cosmos-dev-bounces@xxxxxxxxxxx
05/27/2008 08:05 AM
Please respond to
Cosmos Dev <cosmos-dev@xxxxxxxxxxx> |
|
To
| Cosmos Dev <cosmos-dev@xxxxxxxxxxx>
|
cc
| cosmos-dev@xxxxxxxxxxx, cosmos-dev-bounces@xxxxxxxxxxx
|
Subject
| Re: [cosmos-dev] Status of Datacenter
SML example |
|
joel,
would you be kind enough to provide an example of what you are looking
for?
from reading your note, it sounds to me that you are making the same leap
as me, e.g. facets are roughly equal to a capability. and you want
to have somehting like the RMD where you can "decorate" a property
with additional info, a la metrics. it would be good to get something
like this, and a use case in front of the RM people.
as for a use case, we would want to tie this together with a CMDBf scenario,
e.g. i need to query a CMDB for information about a resource so that i
can understand which properties i want to monitor or something like that.
if you could structure an example with the data center, then ali and i
could review it with the folks from IBM that are working on data model
activities.
-mw
_______________________________________________________________________________________________________________
Mark Weitzel | STSM | IBM Software Group | Tivoli | Autonomic Computing
| (919) 543 0625 | weitzelm@xxxxxxxxxx
|
Re: [cosmos-dev] Status
of Datacenter SML example |
Ali Mehregani
| to:
| cosmos-dev
|
05/23/08 09:50 AM |
Sent by:
| cosmos-dev-bounces@xxxxxxxxxxx |
Please respond to Cosmos Dev <cosmos-dev@xxxxxxxxxxx>
| |
|
Joel,
I'm not aware how the CML workgroup is progressing. You'll have to
look around to see if the workgroup is taking input from the general public.
The example you note below is a language detail that belongs more
in SML than CML. The latter is just instances of SML used to model
resources.
Thanks,
Ali Mehregani
Phone Number: (905) 413-3712
Service Modeling Language - COSMOS
http://www.eclipse.org/cosmos/
Thanks Ali. I really like the
facet concept, both as an organizing principle and as a means of capturing
and retaining separate concerns. I hope it makes it into CML.
Are any other conventions going
to go into CML (that you can talk about)? For instance, XMLSchema and SML
give a model designer the ability to denote identity requirements – would
it be possible to standardize on some additional annotations to decorate
models to distinguish between descriptive attributes and time-sensitive
attributes (for example)?
Cheers,
Joel
-----Original Message-----
From: cosmos-dev-bounces@xxxxxxxxxxx
[mailto:cosmos-dev-bounces@xxxxxxxxxxx]
On Behalf Of Ali Mehregani
Sent: Thursday, May 22, 2008 9:37 PM
To: cosmos-dev@xxxxxxxxxxx
Subject: Re: [cosmos-dev] Status of Datacenter SML example
Hi Joel,
The concept of facets is purely COSMOS. There is a possibility for
a similar concept to be incorporated in Common Model Language (CML) but
no such thing exists in SML.
Thanks,
Ali Mehregani
Phone Number: (905) 413-3712
Service Modeling Language - COSMOS
http://www.eclipse.org/cosmos/
Joel,
To get some examples for all the various SML syntactical elements, you
should check out our test suite. Go to testing/org.eclipse.cosmos.rm.validation.tests
and look in the test resources folder.
I'll have to look into your facet question and get back to you. Valentina
and Harm created the data center example, and to be honest, I never understood
the purpose of the facet elements we have in there.
David
---
David Whiteman | Tivoli Autonomic Computing
Eclipse COSMOS project committer | http://www.eclipse.org/cosmos/
david_whiteman@xxxxxxxxxx
| 919-254-8224 | T/L 444-8224
Hi David,
Thanks for the reply. I'm working
on an internal project (that has taken me away from COSMOS and appears
to be bringing me back) that involves ontology extraction from things like
SML models – so I figured hey, why not start with our COSMOS example?
A couple of other things I noticed
in there – we're not making use of any of the identity capabilities in
the SML schema (or from XSD, for that matter). Any reason for that? Can
I get a xsd:unique up in here? ;-)
Also, is the facet concept purely
a COSMOS thing? I was under the impression that it was part of the SML
effort, but I don't see it anywhere in the spec. The concept seemed to
align well with WSDM's capability model, and fits well the whole separation
of concerns thing, but it appears to be a best-practices recommendation.
Is that the case?
Thanks,
Joel
-----Original Message-----
From: cosmos-dev-bounces@xxxxxxxxxxx
[mailto:cosmos-dev-bounces@xxxxxxxxxxx]
On Behalf Of David Whiteman
Sent: Thursday, May 22, 2008 4:17 PM
To: Cosmos Dev
Subject: Re: [cosmos-dev] Status of Datacenter SML example
Hi Joel,
Yes, they are supposed to validate. The project itself should contain
correct SML code, and should be up to date for the 1.1 level of the SML
spec. That said, it appears our validator has some problems currently
with that sample. We have an open defect targeted for i11 that will
hopefully address some or all of these issues:
http://bugs.eclipse.org/228223
Add yourself to the cc: list if you want to keep informed on progress.
Might I ask what you're looking at SML for? I didn't know you had
an interest in it.
Regards,
David
---
David Whiteman | Tivoli Autonomic Computing
Eclipse COSMOS project committer | http://www.eclipse.org/cosmos/
david_whiteman@xxxxxxxxxx
| 919-254-8224 | T/L 444-8224
Howdy y'all.
Can someone tell me what the status of the
org.eclipse.cosmos.rm.example.datacenter project is? I'm (finally) trying
to get back into COSMOS, and I need some sample SML documents to play with.
When I run the datacenter documents through the Resource Modeling validator
I get a number of errors (occasional missing type constraints, doesn't
like xml:lang, can't find it:facetType(!)). Should these documents validate?
Thanks,
Joel
The contents of this e-mail are intended for the named addressee only.
It contains information that may be confidential. Unless you are the named
addressee or an authorized designee, you may not copy or use it, or disclose
it to anyone else. If you received it in error please notify us immediately
and then destroy it.
_______________________________________________
cosmos-dev mailing list
cosmos-dev@xxxxxxxxxxx
https://dev.eclipse.org/mailman/listinfo/cosmos-dev_______________________________________________
cosmos-dev mailing list
cosmos-dev@xxxxxxxxxxx
https://dev.eclipse.org/mailman/listinfo/cosmos-dev
_______________________________________________
cosmos-dev mailing list
cosmos-dev@xxxxxxxxxxx
https://dev.eclipse.org/mailman/listinfo/cosmos-dev
--
Ali Mehregani
COSMOS Project
http://www.eclipse.org/cosmos/
_______________________________________________
cosmos-dev mailing list
cosmos-dev@xxxxxxxxxxx
https://dev.eclipse.org/mailman/listinfo/cosmos-dev
--
Ali Mehregani
COSMOS Project
http://www.eclipse.org/cosmos/
_______________________________________________
cosmos-dev mailing list
cosmos-dev@xxxxxxxxxxx
https://dev.eclipse.org/mailman/listinfo/cosmos-dev_______________________________________________
cosmos-dev mailing list
cosmos-dev@xxxxxxxxxxx
https://dev.eclipse.org/mailman/listinfo/cosmos-dev_______________________________________________
cosmos-dev mailing list
cosmos-dev@xxxxxxxxxxx
https://dev.eclipse.org/mailman/listinfo/cosmos-dev