LHCOPN meeting Barcelona
*Rediris*
Building a dark fibre infrastructure
*David's reports*
Traffic for LHC is already bigger than expected and will grow as LHC improves.
Necessary to start planning the budget for the increase of capacity.
CPU is cheap, disk is expensive; maybe more remote storage should be use (thus more networking).
T2 networking: Experiments will provide network requirements. Kors Bos is in charge of the activity. The LHCOPN community is in charge of designing the architecture based on the requirements Kors will provide. This infrastructure should be ready for 2013 when the LHC will restart after the upgrade.
*MDM update*
MDM audit: some problem, but SD and E2ECU are working to fix them.
Two dashboards proposed.
*Monitoring WG*
Presented the requirements. Basically the first step is to have e2e connectivity check based on owamp and a dashboard that shows Normal/Degraded/Down status for any T1.
*IPv6*
No request to implement IPv6 so far. The GRID software is not ready to use IPv6 yet.
Not an urgent subject. But we can start working on the LHCOPN part.
*Operations*
Need a responsible for operations. John has been suggested.
GGUS integration with WLCG: Parent-child function will soon be implemented so to correlate tickets.
*Monitoring Discussion*
Keep it simple: by now OWD is the only parameter which will be used to define the status.
Architecture: Perfsonars - StatusLogic - Dashboards.
Dante will provide the interface to the data retrieved by the Perfsonar instances. It's there already.
SARA will investigate if they can implement the item which retrieves data from perfsonars and calculate the LHCOPN status.
CERN will investigate if there is any developers to implement a dashboard.
Statuses:
- UP: (OWAMP +/- 15% / 5 minutes) && (Loss<0.1% /5 min)
- DOWN: Loss=100% / 5min
- DEGRADED: what's in between.
*Actions*
- Edoardo: Invite Kors to report on T2s traffic needs.
- MonWG: Finalize document. Find developers and have a prototype for next lhcopn meeting.
- Next meeting: Tiers to give an update on what their countries and sites are doing for T2 connectivities.
- JohnS, MonWG: Finalize SLD
- JohnS: clean monitoring mailing lists
Next meeting: CERN (Geneva-CH) 7-8 October 2010