[
Date Prev][
Date Next][
Thread Prev][
Thread Next][
Date Index][
Thread Index]
[
List Home]
|
Re: [cosmos-dev] defining SML resources for data collection and manageable agents
|
It is interesting that the GLA is a
target. I say this because the GLA is really just a mutlithreaded job handler
which is used to parse data. The scheduler and the components that plug
into it are all controlled by a simple management api (start stop config...)
and each component has a xml based config file to drive it.
It seems like a perfect experiment to
do abstractions and customization with. You may want a generic way to get
to it etc. but you should be able to allow direct usage of it's own configurability
as well as perhaps some generic wrapping. This opens up all the issues
of having multiple control interfaces to a single "agent" with
a relatively simple use case.
From a SML point of view the scheduler
is an entity you can configure, but the function of each thread and the
components that make it up are all dynamic facets. Something that is not
yet heavily discussed in the SML world
Thanks for your time.
--------------------------------------------------------------------------
Harm Sluiman
IBM Distinguished Engineer
phone:905-413-4032 fax: 4920
cell: 1-647-300-4758
mailto:sluiman@xxxxxxxxxx
Admin : Queenie Lam qlam@xxxxxxxxxx Tie: 969-5864 1-905-413-5864
Valentina Popescu/Toronto/IBM@IBMCA
Sent by: cosmos-dev-bounces@xxxxxxxxxxx
04/05/2007 01:54 PM
|
Please respond to
Cosmos Dev <cosmos-dev@xxxxxxxxxxx> |
|
|
To
| cosmos-dev@xxxxxxxxxxx
|
|
cc
|
|
|
Subject
| [cosmos-dev] defining SML resources
for data collection and manageable agents |
|
Hi everybody
I have created a first draft of the SML definitions for describing data
controllers and data collection agents. Comments and suggestions are appreciated.
The following information is logged into the bugzilla enhancement below.
https://bugs.eclipse.org/bugs/show_bug.cgi?id=178590
I have created definitions for a data controller and a data collection
agent; see attached zip.
The zip file contains:
1. dataController.xsd
A common definition for a controller routine; the controller has references
to a set of agents, defined by the managedAgent.xsd. It also contains a
configuration facet and a definition facet.
Since the controller extends by restriction a generic item type, any controller
can only have these three type of facets : definition , configuration,
agents list.
The controller can be customized though through the predefined facets.
So although the content of acontroller is static, the configuration facet
can be extended to contain ANY type of properties. In a similar way,
the definition facet can be extended to contain more then just the predefined
properties.
A sample of the controller usage is under the tptpDataController.xsd/xml
2. managedAgent.xsd
A managed agent is a routine that can be controlled by a data controller
and can collect data.
The agent type is extended by restriction from the basic item type and
has a predefined set of facets.
Similar with the controller type, it has
- a definition facet ( extendable to add new properties
).
- A configuration facet – extendable to add new properties
- A source data facet; it has the following properties
- extendable:
A source type ( file, db, etc )
A source location property – describe how to get the agent input data
A description
A sample of the agent usage is under the apacheErrorLogAgent.xml and statisticalAgent.xml
I am in the process of validating these types by defining more agents/controllers
so more than likely this is not the final version. I would appreciate if
you review and comment on the current version.
A short introduction of the terms I am using, which are defined in the
Data Center SML model under the item.xsd file :
1. item - is the base type for any SML root object.
An item can have a
collection of facets
2. facet – is a generic grouping that can contain
collections of
properties, facets or facet references
3. property – a base type for a generic property
Thank you,
Valentina Popescu
Senior Advisory Analyst
IBM Toronto Labs
Phone: (905)413-2412 (tie-line 969)
Fax: (905) 413-4850_______________________________________________
cosmos-dev mailing list
cosmos-dev@xxxxxxxxxxx
https://dev.eclipse.org/mailman/listinfo/cosmos-dev
Attachment:
dataCollection.jar
Description: Binary data
