SCIC Video Conf Mtg 7/6/99
Rough notes by Les Cottrell
CERN/IN2P3
The US - CERN link is now at 12 Mbps, of which there is about 9-10Mbps available
and they are seeing 8Mbps utilization. t\The main users appear to be BaBar & D0
going to IN2P3. The CERN link will be upgraded to 20Mbps on Sept 1st 1999. At
the same time the IN2P3 link will be upgraded from 6Mbps to 34Mbps.
CERN is pushing to further upgrade the US link to 45Mbps in March 2000
and expect a
vendor RFP response at the end of July 1999.
There is work in progress to get connectivity to Internet2 and they are working
on an AUP for IN2P3. The Internet 2 board met recently, but the results are not
available.
There was a discussion on what would be reasonable for BaBar to plan on using
on the link from the US to IN2P3. BaBar wants 4Mbpsto transmit some of its data
to In2P3, i.e. about 1/3/ of what will be available. This may require some
traffic shaping to accomplish, maybe via an ATM PVC between CERN & Chicago.
Richard Mount & the IN2P3 folks will specify the requirements.
Harvey wants CERN to commit to
doubling the speed of the US link by a factor of 2 each year. He hopes
that meeting the LHC bandwidth requirements will
require only a gradual increase in funding
due to the dropping in cost/mbps of links.
It would be very valuable if major sites made their
link utilization data to experiments and also if the major sites could break
down the usage by experiment. Basically experiments need a better idea of what
is going on. IN2P3 has some tools to identify who is using a link. One part of
this would be to identify hosts that are to be used for large data transfers.
Working Group activities
There has been very little activity so far. They want to do a technology update and
provide trends, Harvey Newman, Richard Mount & Micahel Ernst are on this WG.
They want to provide
some expectations on how networks will evolve, there may be some interesting data
from the recent CERN tender.
Letters had been drafted prior to the meeting concerning Canada - Germany connectivity
and Japan - Europe connectivity. These were discussed and amended, and will be sent to
the appropriate people.
Japan
See Update on
Japan-EU.
The NACSIS to Europe connection will be upgraded from 2 to 30 Mbps 1, Oct 1999.
There will be a 3-4Mbps KEK - CERN ATM PVC. It will be funded with Japanese money,
they couldn't wait for the Europeans. The Japan/NACSIS link to Cupertino (US W.
Coast) will be
150Mbps, of which 10 Mbps goes to STAR-TAP for ESnet peering,
30Mbps is for Europe (10 Mbps (broken
off at NY) is for peering with Dante, 5 Mbps for DFN, and 15 Mbps for TEN-155).
There are VCs to CERN (3-4Mbps), UK, FR & BE. the costs are about $1200/Mbps/month
which is confirmed by a Dante Itailian link cost. Yukiio is working with Esnet (Jim
Leighton) to try and get a 10Mbps PVC into LBNL, this would reduce the RTT from
Japan to the W. Coast by about 80 msec. They also want to use this PVC for H.323 &
IPv6 pilots.
The KEK - BINP (Novosibirsk) fiber link at 128kbps is being soon
extended to Moscow to bring in MSU,
ITEP, and IHEP. BINP already has a 2Mbps land line for HEP use to Moscow.
Italy
See
http://www.ba.infn.it/~ruggieri/Presentazioni/ICFA-SCIC6Jul199/index.htm.
Canada
See Canada Report.
Dean carlen pointed out that the main Canadian HEP problem was the performance
between Canada & DESY. More on this can be seen at
http://www.slac.stanford.edu/cgi-wrap/table.pl?from=TLD&to=TLD&file=pingloss&data=1999-06.
Michael Ernst will draft a letter to the DFN about this.
Germany
The satellite provider from DESY to the FSU has been changed. they hope for better
performance, they are now using a commercial instead of a military satellite. They
now have a dedicated 2Mbps line for HEP for DESY into the Moscow area. The other 8
regions of the FSU are supplied by a new satellite.
Monitoring
See
Status of SCIC Working Group on Network Monitoring.
Representatives
There was discussion on the need for a UK representative, Davis Williams will find
someone to represent India, they also want representatives for China and Pakistan.
There is also a need to contact ICFA representatives to get responses from the
Russian and Korean SCIC representatives.
Lepton Photon Conference
The organizers have asked Matthias Kassemman to make a 15 minute presentation on the
SCIC activities. He will want 2-3 slides from the network monitoring people. he will
put together a summary of what he is going to talk about, a week before the
conference (9 Aug-99) and will circulate for comments.
The next SCIC meeting will be at CERN on Saturday November 13rd, 1999.
Feedback