Latest point2point meeting from Gerben

*Minutes LHCONE Point2PointExperiment meeting SC'14 - November 19, 2014*
Attendees: Joe Metzger, Artur Barczyk, Bruno Hoeft, Vlad Lapadatescu,
Vincenzo Capone, Joe Mambretti, Gerben van Malenstein
1. Timelines and Actions of Activity 1 [Connecting LHC Sites to the
Sites that are committed to connecting to the AutoGOLE fabric and adapt
their middleware are:
- Caltech (Artur Barczyk)
- FNAL committed (Phil deMar)
- SURFsara (Sander Boele, Freek Dijkstra)
- DE-KIT (Bruno Hoeft)
- AGL-Tier2 (Shawn McKee)
- Brookhaven (Michael Ernst) expressed interest, but they are in the
middle of a 100G upgrade, available after SC’14
We discussed that the LHC sites will each have at least 1 VLAN. The
AutoGOLE is capable of VLAN retagging, so at least one path per site can
be scheduled. Should sites need more than one path, they can have more
than one VLAN. LHC sites choose a VLAN and a prefix that will be routed
over the VLAN when the dynamic path is up. At the LHC sites, routers are
being connected to the AutoGOLE VLAN; we start with BGP sessions between
Action items:
- Gerben to write a one-page architectural draft. Please see this draft
attached. It's a first sketch after our SC meeting, all input is
welcome. I may ask the group to provide more input as well or to put
specific ideas into this document.
2. Timelines and actions Activity 2 [Adapting LHC site middleware]
Activity will be lead by Tangui Coulouarn, with support from John
MacAuley and Tony Wildish.
The first action is creating an brief inventory of site middleware for
ATLAS and CMS: what’s there? In which part of the software can dynamic
paths be scheduled and setup?
- Results of this activity should be available before CHEP (April/May)
3. Status small experiment DEKIT-SURFsara
The setup of this dynamic path:
SURFsara - NetherLight (SURFnet BoD) - GÉANT (AutoBAHN) - DFN
(statically) - DEKIT/Karlsruhe
The connection between KIT and SURFsara worked 10 minutes after the
meeting. Bruno is to perform a data transfer over this connection, via
his pre-production environment.
4. AOB
Our next call will be in 3 weeks: December 9 (today!), please see call
coordinates that I've sent on Dec 2.

Submitted by David Foster on Tue, 12/09/2014 - 09:46