Citation
Marco Caccamo. "Toward the Predictable Integration of Real-Time COTS Based Systems". Talk or presentation, 28, October, 2008.

Abstract
Integration of COTS components in critical real-time systems is challenging. In particular, we show that the interference between cache activity and I/O traffic generated by COTS peripherals can unpredictably slow down a real-time task by up to 44%. To solve this issue, we propose a framework comprised of three main components: 1) an analytical technique that computes safe bounds on the I/O-induced task delay; 2) COTS-compatible devices, the peripheral gate and hardware server, that control peripheral access to the system; 3) a coscheduling algorithm that maximizes the amount of allowed peripheral traffic while guaranteeing all real-time task constraints. We implemented the complete framework on a COTS-based system using PCI peripherals, and we performed extensive experiments to show its feasibility.

Electronic downloads

• Berkeley2_Oct_08.ppt · application/vnd.ms-powerpoint · 4525 kbytes

• HTML

  Marco Caccamo. <a
  href="http://chess.eecs.berkeley.edu/pubs/505.html"
  ><i>Toward the Predictable Integration of Real-Time
  COTS Based Systems</i></a>, Talk or
  presentation,  28, October, 2008.

• Plain text

  Marco Caccamo. "Toward the Predictable Integration of
  Real-Time COTS Based Systems". Talk or presentation, 
  28, October, 2008.

• BibTeX

  @presentation{Caccamo08_TowardPredictableIntegrationOfRealTimeCOTSBasedSystems,
      author = {Marco Caccamo},
      title = {Toward the Predictable Integration of Real-Time
                COTS Based Systems},
      day = {28},
      month = {October},
      year = {2008},
      abstract = {Integration of COTS components in critical
                real-time systems is challenging. In particular,
                we show that the interference between cache
                activity and I/O traffic generated by COTS
                peripherals can unpredictably slow down a
                real-time task by up to 44%. To solve this issue,
                we propose a framework comprised of three main
                components: 1) an analytical technique that
                computes safe bounds on the I/O-induced task
                delay; 2) COTS-compatible devices, the peripheral
                gate and hardware server, that control peripheral
                access to the system; 3) a coscheduling algorithm
                that maximizes the amount of allowed peripheral
                traffic while guaranteeing all real-time task
                constraints. We implemented the complete framework
                on a COTS-based system using PCI peripherals, and
                we performed extensive experiments to show its
                feasibility. },
      URL = {http://chess.eecs.berkeley.edu/pubs/505.html}
  }
  

Posted by Hiren Patel on 29 Oct 2008.
For additional information, see the Publications FAQ or contact webmaster at chess eecs berkeley edu.