U.S. ATLAS TDAQ FY06 M&O Planning A.J. Lankford UC Irvine
ATLAS TDAQ Overview
Overview of FY06 TDAQ M&O Activity TDAQ Construction Project continues beyond CD-4A. Most hardware acquisition and installation in FY06-07 Much software development, esp. HLT, continues. TDAQ M&O activity overlaps with Construction in FY06-07. Construction Project ‘labor’ is ramping down quickly. FY06 is first year of ‘significant’ M&O activity. FY06 M&O activity focuses on DAQ & HLT Infrastructure s/w. TDAQ M&O activities can be classified as: Hardware commissioning Hardware maintenance Software commissioning Software maintenance Operations Detector support TDAQ support
HLT/DAQ Component View
Pre-series Hardware
FY06 Hardware Commissioning & Maintenance Complete commissioning of Pre-series system 10% scale of final system w/ components of final specs Commission final RoI Builder hardware Commission DAQ infrastructure Network infrastructure (large central switch) - connectivity to ROS’s DAQ processors - SFI’s, SFO’s, L2SV’s, pROS’s Racks & rack instrumentation Initial complement of HLT processors - late in year Hardware Maintenance Pre-series system RoI Builder DAQ infrastructure Test beds - particularly “performance testbed”
Software Development: Transition to M&O The Combined Test Beam Release of HLT/DAQ software was nearly complete with respect to DAQ and HLT Infrastructure functionality. This release was commissioned, operated, and maintained for the 2004 ATLAS CTB Run. Subsequent development of this HLT/DAQ software consists of: Added functionality (e.g. fault tolerance & error recovery) HLT selection algorithm development Refinement of existing code based upon CTB and subsequent experience and feedback Development of added functionality and algorithms is Construction Project activity. Refinement of existing code is M&O activity. Starting in FY05, software development is increasingly M&O. After first physics run, software development will be entirely M&O.
FY06 Software Commissioning & Maintenance Commissioning of new functionality Examples of U.S. activity: LVL2 Supervisor and RoI Builder software DAQ software Monitoring Inter-process communication and Information Services Message passing HLT Infrastructure software Run Control & Supervision LVL2 Steering EF Dataflow Software Maintenance Maintenance of above (bug fixes, releases, TDAQ/Offline consistency) Incorporate new functionality based upon experience
FY06 Operations Operating DAQ infrastructure in USA15 and SDX1 in support of Detector (and TDAQ) commissioning Network administration System administration (DAQ and HLT computer systems) TDAQ support teams will start rotations once level of activity at Point 1 demands
FY06 Detector & TDAQ Support Detector Support ROD Crate DAQ maintenance and support Support for: Detector-specific software development, e.g.: ROD software Online histogramming & monitoring Detector commissioning Detector calibrations (and test beam runs ?) TDAQ Support Support development & commissioning of other TDAQ subsystems Examples: SRB support for commissioning LVL1 trigger subsystems Integration of HLT vertical slices Network support for Dataflow performance tests
FY05 M&O Activities Hardware Commissioning Hardware Maintenance Pre-series system, including pre-series RoIB DCS for TDAQ racks Planning for TDAQ infrastructure commissioning Hardware Maintenance Set up RoIB test bench TDAQ test bed maintenance & administration Software Commissioning Releases, release testing, release testing procedures Software Maintenance Maintain code authored by U.S. members, e.g.: Message passing, Information Services, Online Histogramming, HLT Steering, HLT Control, Monitoring Synchronization of releases with Offline Adapt/port code to new compiler and new CERN Linux
FY05 M&O Activities – p. 2 Operations Detector Support TDAQ Support Completion of CTB run System administration planning Detector Support TileCal beam test Monitoring Support of production and simulation of test beam data TDAQ Support Support integration with RoIB in CTB Support of HLT algorithm integration Test bed setup in UK, Dubna
M&O Planning Methodology/History Experience in FY04, particularly CTB, indicated prior estimates of M&O effort were low. New bottoms-up estimate of M&O effort in Jan. 2005 By “new kid on block”, w/ input from all involved Based on effort needed for commissioning and ongoing maintenance of hardware & software components delivered by U.S. Constructed matrix of effort per component vs. time Estimate larger than prior estimate => add Management Reserve MR added covers ~1/2 of increase in estimate Checked total estimate w/ 20% of top-down estimate for ATLAS Coordinated w/ TDAQ Management Team
3.6.1 Preoperations – FY06 3.6.1.1 Supervisor – RoI Builder 28k$ 3.6.1.1.1 ANL: 15k$; Dawson/Drake; TDAQ Support 3.6.1.1.2 MSU: 12k$; Ermoline; TDAQ Support 3.6.1.2 Communications & Travel (16k) 3.6.1.1.1 ANL: (8k) 3.6.1.1.2 MSU: (8k) 3.6.1.3 Programming Support 103k$ (201k) 3.6.1.3.1 ANL: 59k$; Schlereth/TBN; S/w Comm, Detector Supp 3.6.1.3.2 MSU: 44k$ (17k); Ermoline/TBN; H/w+S/w Comm, Det Supp 3.6.1.3.3 UCI: 0k$ (104k); Stancu/Kolos/Negri; H/w+S/w Comm, Det+TDAQ Supp 3.6.1.3.4 UWisc: 0k$ (80k); Wiedenmann/dosAnjos/Xu; S/w Comm, Det+TDAQ Supp
3.6.2 Operations – FY06 3.6.2.1 Supervisor – RoI Builder 137k$ 3.6.2.1.1 ANL: 97k$; Dawson/Drake/Schlereth; H/w+S/w Maint, TDAQ Support 3.6.2.1.2 MSU: 40k$; Ermoline; H/w+S/w Maint, Ops, TDAQ Support 3.6.2.2 Communications & Travel 51k$ 3.6.2.2.1 ANL: 13k$ 3.6.2.2.2 MSU: 13k$ 3.6.2.2.3 UCI: 13k$ 3.6.2.2.4 UW: 13k$ 3.6.2.4 Test Facilities 3.6.2.4.1 ANL: 10k$ 3.6.2.4.2 MSU: 11k$
3.6.2 Operations – FY06 (p. 2) 3.6.2.3 Programming Support 668k$ (241k) 3.6.2.3.1 ANL: 248k$ Schlereth/TBN; S/w Maintenance, Detector Support 3.6.2.3.2 MSU: 90k$ (22k) Ermoline/Hauser/TBN; S/w Maintenance, Ops 3.6.2.3.3 UCI: 205k$ (206k) Stancu/Ciobotaru/Unel/Kolos/Wheerler/Negri/Rogge H/w+S/w Maintenance, Ops, Detector+TDAQ Support 3.6.2.3.4 UW: 126k$ (13k) Wiedenmann/dosAnjos/Xu; S/w Maintenance, Det+TDAQ Support
FY06 Management Reserve FY06 MR priorities (in rough priority order): Hardware commissioning: 0.25 => 0.5 FTE Software commissioning: 0.25 => 0.5 FTE Hardware maintenance: 0.5 => 1 FTE TDAQ support: 0.4 => 1 FTE Operations: 0.3 => 1 FTE Detector support: 1 => 2 FTE SRB: 1 => 1 FTE; Software maintenance: 2.5 => 2.5 FTE Total: 6.2 => 9.5 FTE ~458k$ Impact of MR: Improved readiness of TDAQ for cosmics and beam Improved efficiency of detector commissioning Improved tools for detector monitoring More efficient HLT/DAQ operations Improved efficiency of operations due to s/w enhancements Without MR funds, either or both: Reduced DAQ support for detectors, with delays in detector commissioning Reduced effort on ongoing HLT/DAQ development, with delays in completing critical software