This website is no longer maintained. Its content may be obsolete. Please visit http://home.cern for current CERN information.
Summary of discussion with P.Ribeiro/T.Wijnands SL-CO/FE courtesy Michel Jonker for the LHC-CP workshop.
The LHC related activity in CO/FE started a long time ago in 93 by the 'dream-team' (Mugnai, Keyser, Parker) when KHK was group leader. They envisaged a system for replicating shared memory based in the Scalable Coherent Interface that was developed by other groups at Cern (H.Muller/EcP). Prototypes were made, reports were published and a chipset was made, but eventually the interest was lost by the experiments and the accelerators because the chip set was not generally supported with only one Norwegian firm producing it. The dream team continued in the investigation of the ATM technology for generic networks. End 96 the activity was stopped.
After the Dynamic Effects workshop in February 97, people realized the need for a real time control systems. SL-CO/FE started to investigate RT aspects of current state of the art frond-ends and networks. It was realized that the RT requirements by Frond-ends could be met, but this was not the case for the network and that a real-time network was required. This leaded to pick up the earlier ATM investigations. This was also done in collaboration with some experimental groups (J.P. Dufey, M.Letheren).
It was a slow start, investigating several systems. They studied point to point communications with transport through a switch. The conclusions was that the technology exists. Sl-CO/FE was asked to develop a strategy to set up the infrastructure for a distributed real-time control system for LHC.
A fellow was assigned - to set up a full-scale system based on: - WorldFip field bus (SL-PO choice) - FIP running LynxOs - ATM for intersystem communication.
- To clarify the requirements of the RT environment Requirements from LHC-CP A coherent view of the infrastructure should be provided. If not, the infrastructure will be dictated by the choices of the power converter group. This has already lead to strange effects, i.e. BI develops an orbit data acquisition systems capable of measuring at 100 Hz. 100 Hz is the refresh rate of the Power Converters, this is not the required band width from the physics and machine point of view.
Clear decisions & guidelines are needed to decide if and what real time control requirements are needed. With the frequent changing requirements and ideas, prototyping is very difficult. SL/CO-FE can evaluate a few prototypes, but not more. Current activities SL-CO/FE is building a prototype using ATM for a real time control in the SPS. Some new hardware is needed, needs the collaboration of PO and BI (S.Jackson, A.Burns, J.Pett, Q.King). From a development for a RT control with beam, one can learn things that one cannot learn from RT systems for the string test. Development for String-II, two independent world-fip systems, for PO and for magnet protection.