Page Comparison

1. Opening

Rémy

Discussion on /wiki/spaces/WIS2/pages/306970677

Thorsten: it is better to focus more on technical functions and let the political issues outside the discussion

Remy: 15 GISCs, running same cache/Catalogue, not working, in WIS2, we should not go for this path. Shared service among 1 GISCs.

Hassan: We all agreed about the shared services. the strategy of WIS is based on evolution not revolution. I think to be efficient, let’s focus more on technical functions and go, as we agreed in the previous meeting to reach TT-GISC.

Rémy:

Kai: avoid the supercenter that is running all shared services.

Henning: This is a scale-out problem. For modern architecture, it should be designed to be scaled with more flexibility, such as in future there would be more GISCs that can run the shared service.

Rémy: The fact is that not all the 15 GISCs are not working properly. The concern is not the number of 15 GISC , but to have the some of them pretending to fulfil the GISC role but in reality not. we should learn from the experience of WIS1 to success WIS2.

Summary:

• The team agreed that a shared service is the best solution;

• A small number of Centers to run the shared services is preferable for efficiency;

• Scalable architecture design is needed.

2.

ALL vote for global cache/broker?

• Do we need a global cache?

  • Yes (Jeremy, Henning, Tom, Thorsten, Dana, Kenji, Peter)

• Do we need a global broker?

  • Yes (Kai, Jeremy, Tom, Kenji, Henning, dana, Thorsten, Peter, Rémy)

• Do we need a Global Control Center?

• Enrico introduce the idea of the GCC, to complete an administrative role more than technical.

3

Jeremy

Jeremy presents the data flow between NC/DCPC