Quantifying Middleware Adaptation with Uncertainty Analysis for Streaming Power Grid Applications
Ilge Akkaya, Yan Liu, Ian Gorton

Citation
Ilge Akkaya, Yan Liu, Ian Gorton. "Quantifying Middleware Adaptation with Uncertainty Analysis for Streaming Power Grid Applications". Unpublished article, January, 2013.

Abstract
The power grid is incorporating high throughput sensor devices into power distribution networks. The future power grid needs to guarantee accuracy and responsiveness of applications that consume multiple sensor streams. It is essential for software infrastructure to adapt to situations with large volume and high frequency data streams. In particular, measurement errors in real-time data streams affect the quality of power applications. When the applications are distributed, the power grid middleware needs to coordinate a global knowledge of errors with time stamps across all the distributed parts. It remains a challenge to design scalable and adaptive middleware for this purpose. In this paper, we present a modeling based approach to quantifying the middleware design attributes and its effect on distributed algorithm runtimes. The models represent the entire data flow from sensor devices, through network and middleware to distributed application nodes. We specify uncertain parameters in middleware adaptation as random variables. Using the Ptolemy II integrated simulation tool, we generate Monte Carlo samples of these variables and execute the distributed power application models using the generated sample sets. The simulation results are further analyzed using regression to reveal the most influential parameters to achieve temporal requirements of the power applications.

Electronic downloads

Internal. This publication has been marked by the author for TerraSwarm-only distribution, so electronic downloads are not available without logging in.

Citation formats

• HTML

  Ilge Akkaya, Yan Liu, Ian Gorton. <a
  href="http://www.terraswarm.org/pubs/3.html"
  ><i>Quantifying Middleware Adaptation with
  Uncertainty Analysis for Streaming Power Grid
  Applications</i></a>, Unpublished article, 
  January, 2013.

• Plain text

  Ilge Akkaya, Yan Liu, Ian Gorton. "Quantifying
  Middleware Adaptation with Uncertainty Analysis for
  Streaming Power Grid Applications". Unpublished
  article,  January, 2013.

• BibTeX

  @unpublished{AkkayaLiuGorton13_QuantifyingMiddlewareAdaptationWithUncertaintyAnalysis,
      author = {Ilge Akkaya and Yan Liu and Ian Gorton},
      title = {Quantifying Middleware Adaptation with Uncertainty
                Analysis for Streaming Power Grid Applications},
      month = {January},
      year = {2013},
      abstract = {The power grid is incorporating high throughput
                sensor devices into power distribution networks.
                The future power grid needs to guarantee accuracy
                and responsiveness of applications that consume
                multiple sensor streams. It is essential for
                software infrastructure to adapt to situations
                with large volume and high frequency data streams.
                In particular, measurement errors in real-time
                data streams affect the quality of power
                applications. When the applications are
                distributed, the power grid middleware needs to
                coordinate a global knowledge of errors with time
                stamps across all the distributed parts. It
                remains a challenge to design scalable and
                adaptive middleware for this purpose. In this
                paper, we present a modeling based approach to
                quantifying the middleware design attributes and
                its effect on distributed algorithm runtimes. The
                models represent the entire data flow from sensor
                devices, through network and middleware to
                distributed application nodes. We specify
                uncertain parameters in middleware adaptation as
                random variables. Using the Ptolemy II integrated
                simulation tool, we generate Monte Carlo samples
                of these variables and execute the distributed
                power application models using the generated
                sample sets. The simulation results are further
                analyzed using regression to reveal the most
                influential parameters to achieve temporal
                requirements of the power applications.},
      URL = {http://terraswarm.org/pubs/3.html}
  }
  

Posted by Christopher Brooks on 25 Jan 2013.