[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: Open Mash consortium and Status of vic/vat/mb:



Hi -

We'd be glad to have you make the changes in the Mash source and send
them to us.  They will then be part of the standard release. You can get
a copy of the source code and make the modifications - see the following
web page for info on setting up a source tree for you to modify:
	http://www.openmash.org/distributions/index.html
You should also join the the developers mailing list by sending email to
	openmash-users-request@openmash.org 
with subscribe <your email address> in the body and you'll be put on the
list.  That's the best place to send questions.

Once you've setup the code, you should look though the talks at the
recent OpenMash Workshop at
	http://www.openmash.org/events/00workshop/prog.html
The slides for all talks are on-line and the videos+slides archives will
be up shortly. You should look at my talk on the second day and Paul
Huang's talk at the end of that day for info on getting started.

Once you setup your source environment, the multicast network code is in
the file
	mash-1/net/net-rtip.cc
That's where the multicast session is opened and the file
	mash-1/net/net.cc
is where the send/receive calls are located (actually, some rtp receives
are in net-rtip.cc).  But, the code is pretty easy to figure out and it
ought to be easy to add your abstractions.

I've cc'd Peter Pletcher, my sysadmin, who will create a developers
account for you on the Open Mash site so you can check-in the changes
you make to the CVS tree.  I've also cc'd the openmash-users so they can
see what you're up to.  

Tell Hui I said hi and that I'm excited that you are interested in doing
this.
	Larry
-- 
Professor Lawrence A. Rowe          Internet:  Rowe@BMRC.Berkeley.EDU
Computer Science Division - EECS       Phone: 510-642-5117
University of California, Berkeley       Fax: 510-642-5615
Berkeley, CA 94720-1776            URL: http://bmrc.berkeley.edu/~larry

-- BEGIN included message

Hi Lawrence,
	I am a graduate student working with Hui Zhang at Carnegie Mellon.
Currently, we are working on a project called End System Multicast, where
we take the stand that IP Multicast has several serious problems 
such as scalability, deployment, and support of higher level functionality 
(like reliability and congestion control). Instead, in our alternate 
architecture,
end systems actually participating in the multicast group self-organize into 
an overlay
structure *without* router level support. We have proposed a completely 
distributed
protocol called Narada for achieving this self-organization.

	This work recently appeared at Sigmetrics 2000. 
A pointer to the paper and to the project web page can be found at:
http://www.cs.cmu.edu/~sanjay/EndSystem/endsystem.html


	I am contacting you because we plan to release a toolkit containing
our implementation of the protocol. The plan is to distribute the toolkit
in such a way that vic, vat and mb can use it directly. The idea is to
make no (or extremely limited) modifications to source code of these 
(and other existing IP multicast based ) applications. The applications 
assume they are talking IP Multicast, while in reality, the packets are
intercepted by a proxy process that we implement; the packets are distributed
on the wide-area Internet by Narada as opposed to IP Multicast.

	In the process of supporting vic, vat and mb, we have some simple
suggestions that could be added to the tools that would be greatly useful.
We would prefer to have these changes incorporated in the official version
of these tools, rather than supply our own patches. We are wondering if you
could consider supporting this. If you can refer us to the appropriate people,
we could send details of what we require.

			Thanks a lot,
			Regards,
			Sanjay.
-- 
********************************************************************
Sanjay Gopinatha Rao
Ph.D student, School of Computer Science, CMU.

Ph: (412) 268-1425 
Email    : sanjay@cs.cmu.edu HomePage : http://www.cs.cmu.edu/~sanjay
*********************************************************************

-- END included message