Open Source Contributions

Issues and pull requests upstreamed to projects maintained by others — cloud infrastructure, storage engines, and autoscaling research. For projects I built and maintain myself, see Projects.

Java

Java simulation platform for cloud autoscaling research, used for the exponential-smoothing predictor work behind the IJGUC 2023 publication.

An event-driven and lightweight proactive autoscaling architecture for cloud applications (IJGUC, 2023)
  • PR · MergedFeb 2021
    #4Single exponential fix, double and triple exponential smoothing introduced

    Corrected the single exponential smoothing predictor and added double and triple exponential smoothing with modified upper bounds — the forecasting layer the published autoscaling architecture is built on.

  • PR · MergedJan 2021
    #3Adaptive platform

    Introduced the adaptive scaling platform that adjusts predictor parameters in response to observed demand instead of holding them fixed.

  • PR · MergedJan 2021
    #2TES constants: more controlled const selection

    Tightened how triple exponential smoothing constants are selected, narrowing the search to parameter ranges that converge.

  • PR · MergedJan 2021
    #1TES constants: controlled randomization

    Replaced unconstrained randomisation of triple exponential smoothing constants with a controlled scheme, making simulation runs reproducible and comparable.

Cloud Autoscaling Research

aslanpour/AutoScaleSim

Java17

CloudSim-based simulation toolkit for auto-scaling web applications in the cloud.

  • Issue · OpenNov 2020
    #2WeightedMovingAverage: recent parameters should carry more weight

    Reported that WeightedMovingAverage and WeightedMovingAverageFibonacci apply weights in increasing order against a parameter list ordered oldest-last, so the weighting favours stale history rather than recent observations.