Particpants: Edoardo Martelli, Phil Demar, Dale Carder, Pablo Collasto Soco, Asif Shan, Joe Mambretti, Bill Johnston, Marian Babil, Luca Bassi, Bruno Hoeft, Ettore Tamiro. Andy Lake, Chris Walker. Thomas Tam.
Next LHCONE meeting:
- LHCOPN 20 year celebration: We would like to form a commitee to organize a small celebration. Volounteers: Bruno, Bill, Joe, Edoardo
- Presentations for LHCONE meeting: send proposal to Edoardo.
MultiONE new tags.
discussion about the email sent by Edoardo to the mailing list.
There may be the need to for new tags for these cases:
A) LHCONE Site affiliated with a given experiment, but offering Opportunistic Resources to other experiments.
- Should we add one community for such case? Is it one community enough for all the experiments, or should there more?
B) LHCONE Site providing resources to other experiments which should not be connected to LHCONE.
We have seen in the past that LHCONE can be used by experiments outside the AUP. I got a question from a site worried that the day that MultiONE filtering is implemented, it could break connectivity for underground experiments it is serving.
- Should we add a generic community for Other-not-LHCONE-Experiments?
Discuusion follow: For A) the solution would be to apply all tags for which opportunistic resources are granted. For (B) it would be better not to filter.
Ettore presents a solution where every experiment has an additional tag for the opportunistic resources that can be offered to another experiment.
The problem is that the tagging must be symmetric to the filtering, otherwise the path will become asymmetric.
The matter will be discussed more in detail during the f2f meeting in Manchester
Change of AUP
During the MultiONE email discussion, a site commented that they "would certainly love the AUP language to be softened such that nodes on which LHC jobs could be scheduled dynamically at any time could use LHCONE to talk to any other LHCONE-connected networks"
For the people in the call there should not be a problem to add servers that are in the same datacentre network with the WLCG nodes. It seems the AUP it's not clear enough and seems to restrictive.
This also will be discussed in the meeting in Manchester
