RE: [ebxml-dev] ebMS vs RNIF

[Lars.E.Abrell@skanova.com]

Thanks all for your kind responses 

I have some more questions - specially regarding the reliable messaging capabilities of RNIF 2.0.
In several answers I have got (both online and offline) I'm told that RNIF does NOT provide reliable messaging.

On the other hand, in a paper (from WebMethods) comparing RNIF 1.1 and RNIF 2.0 I have read that: 

"RNIF specifies a reliable messaging mechanism based on Acknowledgements and supplies a set of standard choreography models that all PIPs must follow."

 <<RNIF 1.1 - 2.0.pdf>> 
By reading the answers, the above paper and also the RNIF 2.0 spec - and I'm now a bit confused.

Is it true that RNIF 2.0 only provide a reliable messaging mechanism when you also use the different RosettaNet PIP's? I.e RNIF 2.0 can not be considered as a "standalone" Message Service Handler (MSH) in the same way as ebMS? And that the reliable messaging mechanism in RosettaNet is achieved by sending "Receipt Acknowledgment" messages/documents from the "application layer" by executing the RosettaNet PIP's?

Also I understand that the future RNIF 3.0 will be layered on top of ebMS. Or even that RNIF 3.0 == ebMS? 
Is it true that ebMS will be used INSTEAD of RNIF (3.0) or what capabilities do you believe that RNIF 3.0 will put on top of ebMS?

When do you believe that this new RNIF 3.0 spec will be available?

My best regards

Lars Abrell
Telia / Skanova                  

> -----Ursprungligt meddelande-----
> Från:	spatil@iona.com [SMTP:spatil@iona.com]
> Skickat:	 den 11 april 2002 00:20
> Till:	Abrell, Lars E. /Skanova /075-152 07 75, 0705-61 90 80; ebxml-dev@lists.ebxml.org
> Ämne:	RE: [ebxml-dev] ebMS vs RNIF
> 
> 
> In terms of core functionality, they are pretty close to each other i.e. both of these
> protocols ...
> 1. support multiple transports (HTTP/S, SMTP), MIME packaging
> 2. define semantics and syntax for carrying over bulk of the messaging metadata such as from, to, timestamp, business process identifier,  correlation ID, description of contained business document, etc
> 3. support synchronous, asynchronous messaging, one-way/two-way interactions
> 4. support binary attachments
> ...
> 
> Where ebXML has upper hand is in ...
> 1. Specification defined reliable messaging (helps ensure interoperability)
> 2. Better technology for persistent signatures. In addition, specification includes outline of an array of  security services and various combinations in terms of profile (again helping interoperability as capabilities can be described in terms of precisely defined profiles)
> 3. Multi-hop messaging
> 4. Provisions for addressing partner agreements as part of the messages
> 5. Horizontal protocol, not limited to any vertical or types of business processes in terms of interaction pattern such as two-way, two-party, etc.
> 
> 
> RNIF on the  other hand has upper hand in
> 1. Relatively mature - multiple vendor solutions, testing tool kits available today
> 2. Less optional features, easier in terms of ensuring interoperability
> 3. Security solutions based on commonly available, mature technologies
> 
> Hopefully the efforts are merging (there are already solid signs of progress in this direction) and if I am not wrong, the common solution will be based on ebXML Messaging.
> 
> There are quite a few vendors providing solutions for both of these, recently.
> IONA has been a key participant in specification development and also early adoption in terms of providing solution, promoting interoperability, etc  for both RosettaNet and ebXML. RNIF 1.1, RNIF 2.0 and ebXML Messaging 2.0 solutions are part of IONA's offerings.
> 
> thanks,
> Sanjay Patil
> ----------------------------------------------------------------------------------------------------------
> IONA
> END 2 ANYWHERE 
> Phone: 408 350 9619                                 http://www.iona.com
> 
> 
	-----Ursprungligt meddelande-----
	Från:	arvola@tibco.com [SMTP:arvola@tibco.com]
	Skickat:	 den 10 april 2002 23:28
	Till:	Abrell, Lars E. /Skanova /075-152 07 75, 0705-61 90 80; ebxml-dev@lists.ebxml.org
	Ämne:	RE: [ebxml-dev] ebMS vs RNIF

	Please see my embedded comments.
	 
	-Arvola

	-----Original Message-----
	From: Lars.E.Abrell@skanova.com [<mailto:Lars.E.Abrell@skanova.com>]
	Sent: Wednesday, April 10, 2002 12:47 PM
	To: ebxml-dev@lists.ebxml.org
	Subject: [ebxml-dev] ebMS vs RNIF
	
	
	Hi,
	we have an internal discussion regarding ebMS vs. RNIF.
	
	If we are not using the RosettaNet PIP's (processes and documents) - is it best to use ebMS or RNIF as wire protocol?
	
	<arvola>

	RosettaNet is a vertical standard focusing on the needs of the Information Technology, Electronic Components, and Semi conductor Manufacturing industries, whereas ebMS is intended as a horizontal standard. If you are not using RosettaNet PIPs, it is preferrable to use ebMS as the wire protocol because:
	
	1. It is compatible with SOAP.
	
	2. It offers reliable messaging capabilities.
	
	3. It supports XML Digital Signature which is more flexible compared with S/MIME (supported in RNIF) in terms of allowing referenced objects (not necessarily part of the physical message payload) to be signed.

	4. It may have wider interoperability with third-party B2B implementations.

	</arvola>
	
	What are the pros & cons of using ebMS (2.0) vs. RNIF (2.0)?

	<arvola>
	
	If you want to use SMTP as the transport protocol, then you will probably need payload encryption. This is directly supported in RNIF 2.0 using S/MIME.
	
	Payload encryption is not specified in ebMS 2.0. TIBCO's ebXML offerring does support payload encryption. However, this is not guaranteed to inter-operate with ebXML software offerings from third parties.
	
	TIBCO's RNIF 2.0 implementation has been available since the summer of 2001. It is a more mature product compared with TIBCO's ebXML offering.

	</arvola>
	
	Does anyone have or are aware of any document comparing ebMS (ebXML Messaging Service) and RNIF (RosettaNet Implementation Framwork).
	
	<arvola>

	As part of the investigation for the feasibility of layering RNIF 3.0 on top of ebMS, the RosettaNet engineering team has looked into the mapping between RosettaNet message headers and ebXML message headers. However, that analysis does not really compare ebMS features agaist those of RNIF.

	</arvola>
	
	Best Regards
	
	Lars Abrell
	Telia / Skanova

Attachment: RNIF 1.1 - 2.0.pdf
Description: RNIF 1.1 - 2.0.pdf