Minutes of 4th EVN-NREN Meeting

Exchange Avenue, Schiphol Airport, Netherlands.

12th October 2005 (11:00-17:00)

 

 

 

 

Present:

 

Name                                      Organisation/Institution                   e-mail address

 

Walter Alef                              MPIfR                                                  alef@mpifr-bonn.mpg.de         

Paul Burgess                          Manchester University                        pb@jb.man.ac.uk

Simon Casey                          Manchester University                        scasey@jb.man.ac.uk

John Chevers                          DANTE                                                john.chevers@dante.org.uk

Michael Enrico                        DANTE                                                michael.enrico@dante.org.uk    

Tony Foley                              ASTRON                                             foley@astron.nl

Marcin Garstka                        PSNC                                                  marcinga@man.poznan.pl

Rüdiger Haas                          Onsala Space Observatory                haas@oso.chalmers.se

Giuseppe Maccaferri               CNR-IRA                                             g.maccaferri@ira.cnr.it

Marco Marletta                        GARR                                                  marco.marletta@garr.it

Per Nihlen                               NorduNET/SUNET                              per@nordu.net  

Niels den Otter                        SURFnet                                             otter@surfnet.nl

Eugeniusz Pazderski              Torun Centre for Astronomy               ep@astro.uni.torun.pl

Sergei Pogrebenko                 JIVE                                                     pogrebenko@jive.nl

Jouko Ritakari                         Metsähovi Radio Observatory            jr@kurp.hut.fi

Toby Rodwell                          DANTE                                                toby.rodwell@dante.org.uk

David Salmon                         UKERNA                                             D.Salmon@ukerna.ac.uk

Julianne Sansa                       JIVE/Kapteyn                                       sansa@astro.rug.nl

Ralph Spencer                        Manchester University                        res@jb.man.ac.uk

Martin Swany                          University of Delaware                        swany@UDel.Edu

Arpad Szomoru                       JIVE                                                     szomoru@jive.nl

Lingling Wang                         ASTRON                                             wang@astron.nl

Martin Wilhelm                        DFN                                                     wilhelm@dfn.de

 

Apologies:

 

Mike Garrett                             JIVE                                                     garrett@jive.nl

 

 

 

Agenda

 

11.00 Open Meeting, Introductions

11.20 EXPReS (Arpad Szomoru, JIVE)

11.50 GÉANT2 ((Michael Enrico, DANTE)

12.20 Status reports:

          Medicina (Giuseppe Maccaferri, CNR-IRA)

          Onsala (Rüdiger Haas, Onsala Space Observatory)

          Effelsberg (Walter Alef, MPIfR)

13.00 Lunch

13.40 SURFnet6 (Niels den Otter, SURFnet)

14.10 Data transfer challenges at JIVE ((Julianne Sansa, JIVE)

15.30 ESLEA (Ralph Spencer, University of Manchester)

15.30 LSL (Martin Swany, University of Delaware)

15.30 Future of correlator design (Sergei Pogrebenko, JIVE)

16.00 10Gbps and the future (Jouko Ritakari, Metsahovi Radio Observatory

16.30 AOB

17.00 Close

 

 

Minutes

The slides of the presentations themselves may be accessed through the links in the Agenda. Minutes taken by Tony Foley.

 

 

Status of Observatories

 

Medicina

  • The Medicina last mile was deployed this summer. Tests are planned for October. There is still no connection to the Bologna POP, but within the next few weeks a connection will be made via Faenza
  • The new telescope in Sardinia will be connected via Rome and Palermo
  • Connectivity to Noto remains problematic
  • The 20m telescope in Matera has a 32Mbps connection to GARR, but repairs are needed to fix a structural problem

 

Metsähovi

  • Dark fiber is being bought, FUNET will take care of the link to NORDUNET
  • Price should be a few thousand Euro/year, with an additional fee for the actual traffic
  • No decision has been made yet whether to install 1 or 10 Gb equipment, expected within three months

 

Onsala

  • A wavelength multiplexer has been installed to enable 4470 MTU
  • Participated successfully in iGRID2005, trans-Atlantic fringes between Onsala-Westford-GGAO at 512Mbps
  • Chalmers now has 2.5 Gbps infrastructure

 

Torun

  • Experienced problems during last e-VLBI test
  • Suggestion to test at night, and to try using both available 1 Gbps connections

 

Effelsberg

  • At this time limited to 2 Mbps
  • Will acquire 24 single mode fibers linking Effelsberg to Bonn
  • Aim is to connect to VIOLA network, from there on to GÉANT

 

 

Presentations

Remarks/questions

 

GÉANT2 (Michael Enrico)

 

  • RS: will we run out of wavelengths? ME: probably not within 5 yerars.
  • RS: are all-optical networks considered? ME: is being researched.
  • NdO: prices? ME: ~50000 per transponder, 2 per wavelength.
  • AS: possibility of chopping up 10Gbps through VPN? ME: can be done.
  • AS: is  network built on long-term contracts? ME: only flexible contracts, up to the end of  GÉANT2 funding

 

Data transfer challenges at JIVE (Julianne Sansa)

 

  • PB: is Linux kernel 2.6 being considered? AS: serious bug in 2.6 making it unsuitable (for the time being) for e-VLBI

 

ESLEA (Ralph Spencer)

 

  • SP: suggestion to use Straight Through Mode for data speed tests.
  • AS: both dedicated light path and production network from Jb to JIVE seem to perform equally well. PB: this may be related to the low  use and the limits imposed by the receiving machines. EP: where is the UDP policy limit? Answer: nobody seems to know.

 

LSL (Martin Swany)

 

  • TR: DANTE wants to test LSL between JIVE and one station (probably Torun).
  • EP: why are LSL machines placed at edges?
  • AS: why not run the program on Mk5 itself? MS: LSL should improve throughput even with machines at edges of network, and with more powerful processors the program could run on the hosts themselves.

 

Future of correlator design (Sergei Pogrebenko)

 

  • WA: why not use standard CPUs? SP: the CELL processor offers big advantages like double IO buses, acting like one big vector machine working coherently.
  • RS:data transmission for a 100-telescope SKA will ~10 billion euro. SP: prices will come down in the future, and it may be possible to generate income by for example charging NASA for finding space junk

 

AOB

 

PB: CERN video conferencing (VRVS) was used successfully by two stations during last e-VLBI run, worthwhile to test further.

 

Next meeting

 

AS noted that a lot of momentum has been generated in e-VLBI and with many projects funded and underway fast developments can be expected. This also means that the next meeting will be part of the EXPReS project and that the format of these meetings may change. Schiphol still seems to be the preferred venue. March 2006 was set as a tentative date for the next meeting. After the meeting, two participants suggested video conferencing as a possible alternative.