[e2e] Protocols for tightly synchronised multicast?

[John Lazzaro]

> My challenge problem for networks: nationally distributed musicians
> playing in concert.

I did a Berkeley to Chicago session over the Internet this weekend,
using the system we talked about at NOSSDAV, so we're almost there :-). 
Although, at 100ms latency, we were doing more call-and-response type
playing than truly interlocking music. Note that 100ms is pretty 
extreme for this distance though, one end was on the UCB campus, the
other was a cable modem via RCN in Chicago (I believe). Pings from
the UCB campus to www.uchicago.edu during the same time period were
yielding about 50ms. For the curious, here's the two traceroutes:

UCB-UC

1  e0-5.inr-180-soda.Berkeley.EDU (128.32.37.1)          1.093 ms
2  gig10-cnr1.EECS.Berkeley.EDU (169.229.3.65)           0.408 ms
3  gigE5-0-0.inr-210-cory.Berkeley.EDU (169.229.1.45)    0.776 ms
4  128.32.255.125 (128.32.255.125)                       1.071 ms
5  gigE3-0.inr-000-eva.Berkeley.EDU (128.32.0.201)       1.176 ms
6  pos3-0.c2-berk-gsr.Berkeley.EDU (128.32.0.90)         1.204 ms
7  SUNV--BERK.POS.calren2.net (198.32.249.14)            2.096 ms
8  Abilene--QSV.POS.calren2.net (198.32.249.162)         2.417 ms
9  dnvr-scrm.abilene.ucaid.edu (198.32.8.2)             28.905 ms
10  kscy-dnvr.abilene.ucaid.edu (198.32.8.14)           37.675 ms
11  ipls-kscy.abilene.ucaid.edu (198.32.8.6)            46.550 ms
12  vipernet-gw-a30-11537.uchicago.edu (128.135.249.89) 51.324 ms
13  msfc-1155-v909.uchicago.edu (128.135.247.85)        60.793 ms
14  krypton.uchicago.edu (128.135.12.53)                54.970 ms

UCB-RCNcablemodem


1  cm-gw0-atm-1-0.apm.enteract.com       6.296 ms
2  core0-fe-1-0-0.apm.enteract.com       6.550 ms
3  Serial4-1-1.GW1.CHI1.ALTER.NET        8.197 ms
4  502.at-2-0-0.XR2.CHI4.ALTER.NET       8.189 ms
5  194.at-2-0-0.XR2.CHI2.ALTER.NET      10.289 ms
6  0.so-2-1-0.XL2.CHI2.ALTER.NET         9.509 ms
7  POS7-0.BR3.CHI2.ALTER.NET )           9.446 ms
8  acr2-atm3-0-0-0.Chicagochd.cw.net    10.196 ms
9  acr1-loopback.SanFranciscosfd.cw.net 80.317 ms
10 cenic.SanFranciscosfd.cw.net         98.243 ms
11 BERK-7507--BERK.POS.calren2.net      95.749 ms
12 pos1-0.inr-000-eva.Berkeley.EDU      96.640 ms
13 vlan198.inr-201-eva.Berkeley.EDU     97.631 ms
14 fast8-0-0.inr-210-cory.Berkeley.EDU  97.164 ms
15 GE.cory-gw.EECS.Berkeley.EDU         98.171 ms
16 gig8-1.snr1.CS.Berkeley.EDU              *
17 snap.CS.Berkeley.EDU                 97.796 ms

What surprised me here was the "last mile" into the cable modem was
quite respectable -- the latency culprits were in the core ...

                                                                --jl

P.S. Obligatory URLs

http://www.cs.berkeley.edu/~lazzaro/sa/pubs/pdf/nossdav01.pdf
http://www.cs.berkeley.edu/~lazzaro/sa/sfman/user/network/index.html
http://www.cs.berkeley.edu/~lazzaro/sa/

and check out these other research projects taking different approaches
and getting interesting results too

http://www-ccrma.stanford.edu/groups/soundwire/
http://cnmat.cnmat.berkeley.edu/OSC/

-------------------------------------------------------------------------
John Lazzaro -- Research Specialist -- CS Division -- EECS -- UC Berkeley
lazzaro [at] cs [dot] berkeley [dot] edu     www.cs.berkeley.edu/~lazzaro
-------------------------------------------------------------------------