LHCONE AUP Discussion - LHC[OPN/ONE] und SIG-NGN meeting #55 @ KIT

That leaves 4 categories of proposed changes to define

Rob E:

wasn't just SKA as a motivator to revisit the AUP, there was an issue of opportunistic use of LHCONE

Peter:

There may be two new sites where there is a new VO, and they share resources. It may be unavoidable to have that traffic on LHCONE, even if the VO is not joined. This is a site routing issue. Do sites have to force the VO to join? Or can this fall into incidental usage. It's a gray area.

Bruno:

The expectation from the working group is that scenario falls into the incidental usage category.

Enzo:

Remove HPC from incidental, clarify this is about hosts that would be allowed because they are a part of at least 1 joined experiment.

Is there a proposed new text document? Not yet

Bruno:

We want to agree on what text changes we would want.
The plan is to take the opinion on the proposed changes, figure 4 week for email discussion & other communication channels for feedback.
After that, Bruno would have the new text which would be distributed over email for discussion. We could have that "ratified" at the next LHCONE meeting.

Lars

HPC is just another computing resource and doesn't need to be treated differently.

Eduardo

If a cloud is using the site's IP addressing, this model should work.
Otherwise for cloud if prefixes can be dedicated and the resources emulate a site, then this would be permitted.

Eli:

If a site pledges cloud resources, it is the responsibility of the site to do the necessary integration work.
Add that as a separate cloud use case.

Tim:

Lars:

There are v6-only cloud providers, these could be brought onto LHCONE like any other site.

Next steps: Bruno will solicit more comments via email. Then, by next meeting new AUP text to approve.