MONARC References

• Computing Models
• Data Management
• Technology Tracking
• Object Databases and Storage
• Simulations
• Interesting Articles

Computing Models

• ATLAS Computing Technical Proposal,CERN/LHCC 96-43,19 Dec 1996
  
• CMS Computing Technical Proposal,CERN/LHCC 96-45.19 Dec 1996
  
• LHCb Computing Model, LHCb 98-046
• The analysis model and the optimisation of the geographical distribution of computing resources, M.Campanella,l.Perini,INFN/Milan,July 98, MONARC Note 98/1
  

Data Management

• SLAC Data Management Workshop
• RD45: Distributed Data Management Assumptions
• Joint D0/CDF/CD Run II Data Management Needs Assessment
  Contains some parameters of real physics analysis cases.
  See also http://www-d0.fnal.gov/cdrun2/dma/dma.html
• Joint CDF/D0/CD Needs Assessment Group
• Proposed CDF data model
• CDF data model, a longer note: CDF4346
• Flow and controls: CDF4493
• speadsheets

Technology Tracking

• ICFA Networking Task Force, Status report, July 1998, ICFA/98/671
  and their requirements report
• PASTA Technology Tracking Team for Processors, Memory, Storage and Architectures
  
• NT3 Network Technology Tracking Team have documents on the CS Group web pages
  

Object Databases and Storage

• Status report of the RD45 project, 8 April 1998
  
• Objectivity/DB official page
  
• GIOD Globally Interconnected Object Databases, Caltech, CERN, HP Joint Project
  
• HPSS High Performance Storage Systems project at CERN
  
• Objectivity/DB Read/WritePerf, Sep. 98
  Very detailed set of measurements of read and write performance vs. object size. Martin Schaller tells me that he may update this report in future.
• Clustering strategies, July 1997, RD45 workshop
  Slide #5 has a graph on read performance vs. object size with an empty OS cache. This supplements Martin's measurements above.
• AMS Performance Tests at SLAC
  Lots of measurements on AMS performance over a LAN

• Federations

  • Eva's CHEP98 talk

• Storage overheads

  • Objectivity/DB Storage Overhead
    Very detailed set of measurements, together with a mathematical model. Similar measurements can also be found in one of the RD45 reports but I believe these are the most detailed ones so far.
  • CDF/Tufts/CD Objectivity Working Group Status Report, CDF4522
    Includes some storage overhead measurements.

• Performance under various clustering arrangements

  • Clustering and Reclustering HEP Data in Object Databases
    My recent clustering 'overview paper' which can be used as a basis for estimating disk performance under different workloads and clustering arrangements. Includes a lot of performance measurements.

• Scalability

  • Scalability to Hundreds of Clients in HEP Object Databases
    Covers, among other things, the performance effects of having many clients use the same hardware. Also discusses potential bottlenecks in Objectivity/DB

• Useful Objectivity related software

  • TOPS -- Testbed for Objectivity Performance and Scalability (V1.0)
    My objectivity performance testbed. Currently compiles under v4. Includes documentation
  • Simple Objectivity I/O tracer (V1.1)
    Works on Sun, can be used to study page-level clustering details of running Objectivity clients. Includes documentation.
  • Trace of Objectivity/DB I/O Operations See section 9.
    This discusses a tool similar to the I/O tracer above, made by Marcin Nowak (I believe) using the AMS interface

Simulations

• Simulation of Distributed Architectures(SODA), C Von Praun (follow WWW links for several relevant documents)
  

Interesting Articles

• Why simulate?