Re: BoF Announcement for IETF-57, Vienna: (IP over MPEG-2)

[Dr G Fairhurst <gorry@erg.abdn.ac.uk>]

TAKEI jun wrote:
> 
> Gorry,
> 
> I believe this BOF is continuous work from the unofficial BOF in Salt
> lake city IETF and the situation has not been changed a lot.
> 

The one in Salt Lake City was a BAR-BoFD - this is an OFFICIAL BOF.

A few changes:

(i) The encapsulsation scheme(s) is/are more advanced.
(ii) The subject of address resolution now needs to be aired in the 
open to decide what people really think...
(iii) It is likely we'll now focus on key things first. and get some
action going.... things have been moving too slowly for my liking
when we had too broad a charter.

> Based on the information which I have and the situation in Japan,
> urgent issue related on MPEG/DVB is IPv6 encapsulation. In terms of
> IPv4, there are existing standard even though it was not defined by
> IETF and it is not efficient ;-). If the WG try to define new
> encapsulation mechanism of IPv4, it needs strong merit ,reason and a
> time to make it be RFC. But IPv6, there are few mechanism which can
> handle on MPEG/DVB-TS(I am not saying no standard). So my comment is
> WG should focus on IPv6 encapsulation. Hopefully it should be better
> that current MPE mechanism by using lessons and learned from IPv4
> experience.
> 

If I udenrstand correctly, I'd say that IPv6 is a key goal.

The new encpauslation will be oriented at the IPv6 service, and
I expect will be backwards compatible with IPv4.

> Second item which on the agenda, address resolution between MPEG PID
> and IP address. Those address space are totally different each other.
> And normally PID assignment has already done by service provider(TV
> broadcasting company, IP service provider, etc.) or satellite
> operator. So it is difficult to assign specific IP address to specific
> PID. Maybe it is better to define new service mapping table like NIT,
> PAT and PMT.
> 

OK, so this is an interesting topic.  I think one thing we can
***DEFINITELY*** look at is how to signal the IP/MAC to PID mapping.

It turns out, that since the Salt Lake City Bar-BOF, DVB has
published an update on their view of how to do this.  This uses
the INT - an MPEG-2 signalling table. This document describes the 
table, not tell one how to build the network service. An RFC
describing how this relates to IPv4/IPv6 would be good, and
identifying what can/can not be done using this mapping. 

Is the DVB/INT mapping good for dynamic multicast? 
What other options are there?
How should it work with UDLR? 
What shoyuld happen in IP-only networks that do not
send other MPEG-2 tables?

Thoughts.... from people on this list???

> Anyway this is the start point of the discussion.  I hope we can
> define something from IETF.
> 

Yes - we NEED to debate the charter and get to work to finish
the documents. I look forward to hearing from people!!!

Gorry

> Best regards,
> 
> Jun Takei
> 
> Dr G Fairhurst wrote:
> 
> > A BOF will be held in Vienna on the subject of building standards for IP
> > networks using MPEG-2 Transport Stream transmission. This is intended to
> > cover issues common to implementors/operators wishing to build efficient
> > IP transmission networks using DVB, ATSC, etc. Satellite systems are
> > currently a major user of this technology.  The description is below.
> >
<<snip>>
> >
> > The mailing list archives are at:
> > http://www.erg.abdn.ac.uk/ip-dvb/archive
> >