Minutes of NEXPReS WP6 MS601 telecon on 2010-11-02
The telecon was held on 2010-11-02 and started at 11:00 UTC.
Participants:
Simon Casey (OSO), Neal Jackson (UMAN), Gerben van Malensteijn (SURFnet), Ari Mujinen (Metsahovi), Paul Boven (JIVE)
Agenda:
- Progress on MS601
- Any Other Business
- Scheduling for a bi-weekly telecon
Minutes:
Gerben gave a short overview of the demo that was performed by SURFnet and others at the GLIF Workshop at Cern in October, where dynamic circuits amongst several NRENs were created, tested and torn down again under software control. This included use of the Fenius protocol on OpenDRAC for controlling SURFnet equipment and a Force-10 switch. Fenius is an interim protocol for talking to Domain Controllers, which will eventually be supplanted by the upcoming NSI protocol.
SURFnet (Gerben and Hans) will be adding text to the MS601 draft and asked for an account for the WIKI.
Paul brought up the issue of IP assignment for dynamic paths. As the end-point of the lightpaths at JIVE will be the central JIVE e-VLBI switch/router, we should be able to assign VLANs (with virtual IP addresses) to the ports that connect to the dynamic lightpath equipment. At this moment, OpenDRAC only supports management of Force-10 switches, but the JIVE switch/router is an HP ProCurve 5412zl. We should investigate whether ProCurve switches can also be added to the supported base of OpenDRAC.
Gerben mentioned that the next generation of SURFnet (SURFnet 7) is currently being planned, and it is forseen that the current SDH based backbone will be replaced by Metro Ethernet in this version. As the rollout of SURFnet 7 will not happen before Q4 2011, we should plan on using the current SDH based system first, and will have to adapt to the new network somewhere during the NEXPReS project. As SURFnet 7 will also be controlled through OpenDRAC, that should not be a major change from the NEXPReS perspective.
Neal said he sees a distinction between 'hard real-time' requirements as for e-VLBI, and the possibly less stringent requirements for LTA data transfer, where it might be possible to adapt to available bandwidth, or change scheduling to match bandwidth availability. Unfortunately the scheduling of this telecon prevented Hanno Holties (ASTRON) from participating and commenting on LOFAR's requirements. Gerben replied that OpenDRAC has a 'find a timeslot' feature where it will find the earliest timeslot that can accommodate a particular request, although at the moment that feature isn't working.
Finally, Paul Boven (JIVE) proposed that, in view of the difficulties of scheduling these telecoms, that it might be helpful to have a biweekly telecon scheduled on a fixed time and day of the week. The participants of the telecon preferred Wednesday mornings for such a bi-weekly telecon, but of course the other participants in the WP will be asked for their preferences too.
Actions:
Create Wiki accounts for SURFnet (done already).