All things related to multicast

[Leonard Giuliano <>]

Dan, Bill, et al- interdomain multicast ain't dead yet!

The previous efforts from 12-14 years ago to get everyone in the world to 
enable mcast natively have beyond a doubt failed miserably.  Multicast has 
a number of problems, and I'm sure the folks on these mailing lists can 
each come up with plenty of their own favorites.  But the biggest problem 
has always been that multicast is an all-or-nothing solution, and getting 
every layer 3 hop, every router and firewall on the Internet to just turn 
on a few new protocols was simply something that wasn't going to happen.  
However AMT solves that problem.  With AMT, multicast works gloriously in 
natively-enabled pockets of the Internet, but also can be received in the 
other 98% of the Internet that doesn't support multicast.

What makes the Internet so powerful is ubiquity- just connect to the 
network, and you can get just about all the content that is available.  
But without AMT, only 2-3% of the Internet could receive multicast.  
Because of this, there was little inertia to add content or attempt to 
expand the audience.  You had the audience/content chicken-and-egg.  But 
AMT fixes that by theoretically expanding the audience to the entire 
Internet immediately.  Instead of needing everyone to support mcast, you 
need only one guy with AMT to support it.  And if it does succeed, and 
traffic does grow, those non-native networks finally have a real reason to 
turn mcast on.  So AMT is really just a first step.  

At IETF in Vancouver, AT&T announced that with their Octoshape engagement 
for such things as the French Open, the Eurovision Song Contest and CNN 
live, they had seen up to 50k simultaneous AMT tunnels.  This is 
interdomain multicast, and AMT is really only in it's infancy (after an 
admittedly long gestation period). 

As for v6 mcast, I think you'll need to see v6 unicast and v4 mcast each 
reach critical levels first.

-Lenny

On Mon, 17 Dec 2012, Dan Pritts wrote:

-) 
-) > I'm not about to suggest that we drop support for wide-area IPv4 
multicast.
-) > There are legitimate uses still being made of it, even though there 
aren't
-) > any in evidence on our network this Monday morning. I'm really beginning 
to
-) > wonder whether IPv6 multicast has a future, though. It seems to me that
-) > we're now seeing the other side of the adoption curve for wide-area
-) > multicast in general, and as a result we aren't going to have even the 
low
-) > volume of interesting applications for IPv6 multicast that drove the
-) > original adoption for IPv4.
-) > 
-) > I won't even go so far as to suggest that we 'early adopters' turn off
-) > interdomain IPv6 multicast, although nobody would notice. But if anyone 
is
-) > actively using interdomain IPv4 multicast, they really should get to work
-) > moving their application over to IPv6 before the tiny number of networks
-) > who support it decide that it's just another source of complication, bugs
-) > and potential security holes, and dump it.
-) 
-) I project-managed the network build for 12 Internet2 member meetings 
between
-) 2006 and 2012.  Some of these were easier than others to make happen, but 
one
-) consistent problem was getting multicast working.  Even in DC, where MAX 
just
-) turned on and off the same config year after year, I'm pretty sure it got
-) crufty and broke at least once during that period.
-) 
-) We never really saw a lot of application adoption.  We had a few of the 
same
-) folks every year, fighting the good fight, but never any critical mass.
-) 
-) When I arrived in the R&E networking world in 2002, multicast was still
-) viewed as an advanced technology.  It was clear to me, early on, that it 
was
-) really difficult to make it work properly, but there was at least the hope
-) that deployment of IGMPv3 would be the magic bullet.
-) 
-) When I left Internet2 this past , we had not turned on IGMPv3 on our office
-) LANs, because none of the applications worked with it (to be clear, our
-) equipment was capable, and we'd tested it).  We were still running MSDP and
-) having to bug Merit or MAX occasionally when something broke.
-) 
-) I never felt comfortable saying so while I was an Internet2 employee, but
-) inter-domain multicast is dead.  For the record, there was no explicit gag
-) rule.  The general feeling was that if members wanted it, we should do it,
-) and it surely wasn't my place to tell the membership otherwise.
-) 
-) Now I'm a member, so I say, shoot it in the head, and put it out of our
-) misery.
-) 
-) Disable IPv6 multicast immediately.  No one will notice.
-) 
-) Phase out inter-domain IPv4 multicast over the next year or two. A few 
people
-) will have to change their applications.
-) 
-) Most importantly...spend everyone's engineering effort on something more
-) useful.
-) -- 
-) Dan Pritts
-) ICPSR Computing & Network Services
-) University of Michigan
-) +1 (734)615-7362
-)