BTM Meeting with AP

Meeting: 18h July 2017

Participants: Jutta Fitzek, Hanno Hüther, Raphael Müller, Vitaliy Rapp

Minutes: V.Rapp, state: Draft

Outcome

Discussion: BTM require a list of Devices (Trafos) which should be taken for measurements for particular BPCs. Were to get the list?

• BTM needs to know which devices to listen and aggregate for a particular BPC
  • The rules for device selection may be complicated e.g. use Trafo X if Trafo Y is not available (broken) etc.
  • BTM concept does not consider the system to be able to decide by its own which devices need to be aggregated for particular BPC

• LSA knows which Accelerator Zones are relevant for particular BPC
• LSA knows which Trafos are positioned inside of those Accelerator Zones, but only those, which require settings from LSA
• The full information about all Trafo devices inside of an Accelerator Zone, may be also found inside of the FESA-DB, but FESA DB may contain wrong, obsolete, test etc. devices. The information there is therefore not sufficient to pick right devices
• It should be possible to mask particular devices for the aggregated vector, for instance if those devices are broken, but. the beam are still executed. Those must be excluded from the beam quality evaluation and hence the rules should not be applied on the data from them

• The ability to select right devices for the aggregation is essential for the system. The rules how to do it, however are not clear and need a separate discussion
• TODO: Organize a discussion meeting with R.Steinhagen, D.Ondreka, S.Julicher, AP and SV to discuss it (possible from 17.08)

Discussion: BTM requires a list of currently executed BPCs

• The aggregation of values should be based on the BPCs currently executed in the facility. LSA is the primary source of this information

• LSA will not distribute the actual state, but do allow collection of it using the RMI Interface
• The Idea is, that LSA only informs clients, that the state was changed via messaging broker. The usage of the Kafka broker here could be considered

Discussion: Subscription for DevAccess devices

• BTM should be able to subscribe for all events, similar to wildcard subscriptions in FESA
• BTM must be able to map the Virtual Accelerator to the BPC

• The mapping between VirtAcc and BPC exists in LSA, and therefore may be transferred to BTM
• To aggregate the values for 1 BPC execution it may be required to take in account the BPC start timestamp, which is provided by timing
• Values for particular BPCs may also be accessed using the event subscription, however need to clarify with L.Hechler if it is possible to use the event subscription
• The subscription to all VirtAcc is possible (VirtAcc = -1). A check is required if JAPC is supporting this feature

-- VitaliyRapp - 19 Jul 2017