Columbia University’s Pupin Physics Laboratories hosted a workshop October 30-31 for about 40 DUNE Far Detector DAQ Consortium members and collaborators — about half in person, half remotely — to review in more detail different architectures being considered for the DAQ system and to examine them under existing technical constraints. The workshop was organized as part of the DAQ Consortium’s effort to narrow down options to a baseline DAQ design.
The DUNE FD raw data rate of roughly a few TB/s presents a challenge, especially when compounded with several unknown and difficult-to-predict-a-priori conditions, such as expected noise levels and radiological background rates, which can significantly affect data processing performance.
The DAQ system must be designed with sufficient redundancy so as to be able to handle — and to transfer to offline computing — data rates significantly higher than expected under ideal operating conditions, and nominally with enough processing capability to be able to effectively and efficiently reduce data to rates that are manageable by offline computing, without jeopardizing DUNE’s physics sensitivity.
These conditions — and the short timescale between now and detector operations — require a system design that is built with redundancies, is modular and scalable, and that has been demonstrated to a large extent. The workshop participants examined existing technical solutions form other experiments, including MicroBooNE, protoDUNE SP and DP, SBND, and ICARUS.
The workshop discussions distilled into at least two reasonably thought-out architectures, concluding with lists of action items aimed at developing detailed cost estimates for each option, and understanding the data processing steps that the design must accommodate. The DAQ Simulations team will focus on the latter during the upcoming DUNE Physics Week at Fermilab, and, since supernova physics is a driving factor for the design, will do so in close collaboration with the SNB/LE physics group.
A follow-up workshop in the UK is planned for January 25-26, at which the DAQ Consortium members will aim to finalize architecture designs for the Technical Proposal.