Dan,
Thanks for putting this discussion back on the standard track.
I do not think that we have done a satisfactory job in regards to presenting
our solution to David, John and the individuals from OASIS BCM TC. I want to
assure you that our FERA-based ebSOA architectural solution includes all
aspects of the SOA architecture on the level of the specification details
that can be used for enterprise SOA implementations:
- Collaborations (internal and external),
- Security,
- Messaging,
- Protocol profiles and agreements,
- Universal business documents formatting,
- Registry and repository-based management of meta-data and business
artifacts,
- Business rules,
- Agent frameworks
Using the Federated Enterprise Reference Architecture (FERA) reference model
we produced the set of ebSOA specifications needed to provide full
enterprise level SOA support from both a functional as well as
implementational point of view. We think that the missing factor in the
ebSOA architecture is the semantic integration which brings all of the above
aspects together. We have fully defined the FERA-based ebSOA collaboration
semantics that resolves this problem. Our ebSOA collaboration semantics is
also based on the formally defined information model that can be stored in a
standard registry like the OASIS ebXML Registry or OASIS UDDI Registry and
used to take care of both context and content related to any business
process. This is supported by a single open standard-based XML document we
call the Collaborative Process Information Document (CPID) that can be
generated from a business process definition in a visual modeling tool.
Submitting the content of this document to a standard registry will generate
the entire collaboration process information model that together with
FERA-based ebSOA collaboration semantics already implemented via FERA-based
ebSOA managers (remember all FERA-based managers in ebSOA architecture -
Federation Manager, Agent Interface Manager, Event Manager, Activity
Manager, etc.) enables full open standard-based support for the enterprise
business collaboration processes of any complexity.
Another important aspect of our proposed solution is standard convergence.
We very much believe that standard convergence is the future and in our
architectural solution we reference the best "standard brands" available
today. It does not matter which standard body they are coming from. What
matters is the standard's value in the ebSOA functional offerings. For
example, these are some of the standard components we reference in the
FERA-based ebSOA: OASIS ebXML BP, OASIS WSBPEL, OASIS UBL, OASIS ebXML
Registry, OASIS UDDI, W3C SOAP, OASIS ebXML Messaging, WS-Security, OASIS
XACML, OASIS SAML, WS-Reliability and others.
We also want to mention that two important components of the FERA-based
ebSOA architecture, Agent Framework and Business Rules, are not supported by
any standard today. OMG has two pending Requests For Proposals related to
business rules: Business Semantics of Business Rules and Production Rule
Representation. Our definition of a standard is a specification produced and
approved in a large recognizable standard body like OASIS, W3C, OMG, IEEE,
etc.. Some specification components are coming from groups and programs like
Java Community Process (JCP). While some of them achieved great success over
last few years, we believe that ebSOA specification should reference
standard components first. However, when the standard is not provided we
recommend using other groups' specifications, but on the implementation
level only. We believe that if some industry groups have valued
specification offerings they should be encouraged to submit their spec
proposals to standard bodies and speed up the process of developing very
much needed standards that we do not have today.
I hope that this clarifies key aspects related to our FERA-based ebSOA
proposal and our vision about enterprise SOA.
Regards,
Goran
----- Original Message -----
From: "Dan Pattyn" <>
To: <>; "'David Webber (XML)'" <>
Cc: <>; "'Goran Zugic'"
<>; "'Dale Castle'" <>; "'ebSOA
OASIS TC'" <>; <>;
<>
Sent: Wednesday, May 11, 2005 11:08 AM
Subject: RE: [ebsoa] events and agents as defined in an ebSOA
>I have been looking at JXTA as a middleware peering discovery management
> mechanism for Asterisk the open source IP PBX. JXTA is available from
> Verizon and I have heard that Telenor and some other northern European
> telcos have incorporated a security model on top of JXTA. Anybody know of
> these links.
>
> I don't know why we got off on the discussion of the implementation
> language. We should just be sending a declarative logic description of the
> service contract(policy and contract-- BPSS 3.0?) (XML? Or OWL?) to the
> agent ecosystem. I think this is where the FERA CPIM comes into play.
>
> Verizon JXTA service for $14.95 a month.
> http://newscenter.verizon.com/kit/iobi/network.vtml
> http://newscenter.verizon.com/kit/iobi/profeatures.vtml
> http://newscenter.verizon.com/kit/iobi/enterprise.vtml
>
> You can provide the same services and epr functionality on Asterisk IP PBX
> offering enterprise level email with collaboration and federation support
> over VOIP with team specific collaboration presence, discovery, and role
> based availability. Eventually they will figure out how to add services
> on
> top of this generic infrastructure service. Maybe they will look to
> ebSOA????
>
>
> Dan
>
> -----Original Message-----
> From: john c hardin [mailto:]
> Sent: Monday, May 09, 2005 2:46 PM
> To: David Webber (XML)
> Cc: ; Dan Pattyn; ;
> Goran Zugic; Dale Castle; ebSOA OASIS TC; ;
>
> Subject: Re: [ebsoa] events and agents as defined in an ebSOA
>
> Understood David, thank you. I was suggesting that we use applicable
> universal
> "concepts" from the JXTA projects, not the actual implementation.
> Obviously,
>
> JXTA could be implemented as an agent framework if the implementer chose
> to
> do so.
>
> I just wanted to make sure that we get a well-rounded picture of agent
> concepts.
>
> David Webber (XML) wrote:
>> John,
>>
>> The caveat I have is simple - the design of ebSOA should be
>> implementation
>> language
>> neutral.
>>
>> That is the whole point behind the BCM linking and switching - its based
> on
>> classical
>> declarative logic processing - and can be expressed as XML structures -
> that
>> can
>> then be read and processed by any agent engine.
>>
>> There are literally 100 different agent engines out there - all with
>> their
>> own
>> proprietary syntan and API - (even if they are written in Java - or maybe
>> especially
>> if they are written in Java!).
>>
>> What the world does not have is a common method of integrating *all*
> these.
>> That I believe is where traditionally OASIS can look to make a
>> significant
>> contribution...
>>
>> DW
>>
>> ----- Original Message -----
>> From: "john c hardin" <>
>> To: <>
>> Cc: "Dan Pattyn" <>; "David RR Webber"
>> <>; <>; "Goran Zugic"
>> <>; "Dale Castle" <>; "ebSOA
> OASIS
>> TC" <>; <>;
>> <>
>> Sent: Monday, May 09, 2005 2:48 PM
>> Subject: Re: [ebsoa] events and agents as defined in an ebSOA
>>
>>
>>
>>>Richard Manning from Sun Microsystems presented on JXTA at the 2004
>>
>> Symposium.
>>
>>>Slides are found at:
>>>
>>>http://www.oasis-open.org/events/symposium/slides/manning.pdf
>>>
>>>john c hardin wrote:
>>>
>>>>I would like to start a conversation on the list re: the agent
>>>>architecture, as proposed in the Federated Enterprise Reference
>>>>Architecture approach.
>>>>
>>>>Thanks to Goran, Vasco and the ebXMLsoft team for creating the
>>>>presentation attached, and to David Webber and Dan Pattyn for pointing
>>>>me to the BCM material.
>>>>
>>>>I think that there are several inputs that we can include to round out
>>>>the ideas of a non-intrusive, event-driven agent model. Opinions are
>>>>requested from the group on these, and preferences for how to specify
>>>>the agent mechanisms using a combination of the approaches here:
>>>>
>>>>- The Agent Framework is described in the attached BCM / FERA
>>>>presentation. There is a lot of good material here, however, we need to
>>>>make sure that we are re-using stable OASIS specs where possible. I
>>>>would like to map the specifications mentioned in the presentation to
>>>>the existing specs.
>>>>
>>>>- The BCM Linking and Switching as a mechanism to bind to the agent
>>>>framework
>>>>
>>>>- Possibly including JXTA concepts, where possible, as they describe
>>>>peering between components and applications.
>>>>http://www.jxta.org/JXTAFAQ.html
>>>>
>>>>- Complex Event Processing, using the concepts outlined by David
>>>>Luckham: From http://www.complexevents.com/
>>>>
>>>>Complex Event Processing (CEP) is an emerging technology for building
>>>>and managing information systems including:
>>>>
>>>> * Business Activity Monitoring
>>>> * Business Process Management
>>>> * Enterprise Application Integration
>>>> * Event-Driven Architectures
>>>> * Application Servers and Middleware
>>>> * Network and Systems Security
>>>>
>>>>
>>>>
>>>>
>>>>
>>>>
>>>>
>>>>
>>>
>>>--
>>>~~~~~~~~~
>>>john c hardin
>>>Chair, OASIS ebSOA Technical Committee
>>>http://www.oasis-open.org/committees/tc_home.php?wg_abbrev=ebsoa
>>>313.279.1377 new *VONAGE* number
>>>mailto:
>>>
>>>"The new electronic interdependence recreates the world in the image of a
>>
>> global
>>
>>>village."
>>>
>>> Marshall McLuhan, "Gutenberg Galaxy", 1962
>>>
>>>
>>
>>
>>
>>
>
> --
> ~~~~~~~~~
> john c hardin
> Chair, OASIS ebSOA Technical Committee
> http://www.oasis-open.org/committees/tc_home.php?wg_abbrev=ebsoa
> 313.279.1377 new *VONAGE* number
> mailto:
>
> "The new electronic interdependence recreates the world in the image of a
> global
> village."
>
> Marshall McLuhan, "Gutenberg Galaxy", 1962
>
>
