[AG-TECH] AG 2.0 and VRVS vic and rat and other ramblings

[Ivan R. Judson]

Hi Doug,

I think it's important to understand the layering of the software in the
overall architecture and then to understand where each project fits into
that layering.

The bottom layer is the streaming media, and almost all large scale systems
are using RTP/RTCP now for that. Including vrvs, sip, h323 and the AG.

Above that you can label the layer, session coordination. This is where
H.245, SIP and the AG differ. They each have a different solution to the
problem, SIP is the forthcoming replacement for H.323's H.245 signalling
protocol. SIP is a nice ietf standards based solution that's an open
protocol (unlike H.323 - you have to buy the spec). Rumour is that VRVS is
closely related to SIP, if so, the question is why have 3 solutions doing
*exactly* the same thing?

The AG is taking a completely different approach by leveraging the Web
Services technology for this, WSDL/SOAP. This allows us to seemlessly
integrate things like Grid Services into our collaboration system, you just
can't get that with the SIP/H.323/VRVS solution. Closely related to this
observation is the fact that as far as I can tell, we're the only ones
building a "collaboration platform/framework", the other solutions are
either commercial products (H323/SIP), or research turned production (VRVS).

One significant and often overlooked point is that the AG is the only Open
System in your list. The others are all closed to the end-user and
developer.

Media tool work is big and hairy problem and the AG effort is currently not
supported to do it. We'd be happy to get support to do some media tool
development, but there are better solutions available than continuing to use
ancient media tools. 

A nice set of platform enhanced tools for each of the major platforms (OS X,
Windows, Linux) would be a big benefit, and if built upon the same
underlying infrastrcture (use RTP, use it right) the tools can interoperate.
The advantate to this would be I could use Microsoft's msvideo codec or
Apples H.264 when it comes out. You know they are going to make it faster
than an opensource version. The only gotcha is that then you have to have
either one least common denominator among all platforms, or a solution to
transcode on the fly, otherwise the community fragments around platforms.
 
One key effort that is important to note is that the inSORS solution was
derived from the AG1.x software, but has been tracking (via plug-ins) the
AG2.x software. Although there has been divergence to a small degree, we're
working with inSORS to re-converge in our 3.X software so that we're
interoperable at the basic level (text, audio, video).

I don't believe we're going to see the AG community fracture, both the
research and commercial participants in the community understand the value
of a cohesive AG community. But I do think the differences between the AG
and the other systems you've listed are going to become very evident in the
next twelve months.

--Ivan

> -----Original Message-----
> From: owner-ag-tech at mcs.anl.gov 
> [mailto:owner-ag-tech at mcs.anl.gov] On Behalf Of Douglas Baggett
> Sent: Thursday, October 21, 2004 12:02 PM
> To: ag-tech at mcs.anl.gov
> Cc: Thompson, Kevin L.; jkoss at nsf.gov; dgatchel at nsf.gov
> Subject: [AG-TECH] AG 2.0 and VRVS vic and rat and other ramblings
> 
> I've been evaluating which software to use for my desktop 
> users and have been quite impressed with the work the VRVS 
> people have done with vic/rat and software installation and 
> ease of use. My question has to do with the differences and 
> advantages/disadvantages with going with VRVS for desktop use 
> to access grid meetings vs standard AG 2.x  vs Insors (any 
> suggestions or experiences greatly appreciated!). Also, have 
> there been any discussions with the VRVS folks about the 
> changes they've made to vic and rat and possibly taking 
> advantage of their work (and of course helping them as well 
> in the opposite direction) to improve the version of vic and 
> rat used in AG? Just a cursory search of the mailing list 
> reveals some other vic forks (other than VRVS) as well to add 
> MPEG4 and other stuff (OpenMash and others). 
> 
> The added fact that InSors develops their own video software 
> seems like we are headed down a road where the least common 
> denominator will start to influence what kind of advanced 
> video features are available. It looks like MANY people want 
> the same set of features (MPEG4, H.264, possibly HDTV....) 
> but people are going down their own road, which leads to 
> incompatible versions of vic using higher quality codecs and 
> resolutions.
> 
> Thanks for all the great work! 
> 
> -Doug
> 
> --
> Douglas Baggett
> CISE IT Support
> National Science Foundation      
> CISE/OAD                         
> 703-292-4551                      
> M-F 8-4pm EST                     
> 
> 
-------------- next part --------------
A non-text attachment was scrubbed...
Name: smime.p7s
Type: application/x-pkcs7-signature
Size: 3658 bytes
Desc: not available
URL: <http://lists.mcs.anl.gov/pipermail/ag-tech/attachments/20041021/049f0c22/attachment.bin>