Adaptive Offloading for Pervasive Computing

  • Authors:

  • Xiaohui Gu;Alan Messer;Ira Greenberg;Dejan Milojicic;Klara Nahrstedt

  • Affiliations:

  • University of Illinois at Urbana-Champaign;Hewlett-Packard Laboratories;Hewlett-Packard Laboratories;Hewlett-Packard Laboratories;University of Illinois at Urbana-Champaign

  • Venue:
  • IEEE Pervasive Computing
  • Year:
  • 2004

Quantified Score

Hi-index 0.00

Visualization

Abstract

Pervasive computing lets users continuously and consistently access an application on heterogeneous devices. However, delivering complex applications on resource-constrained mobile devices such as cell phones is challenging. Application- or system-based adaptations attempt to address the problem, but often at the cost of considerable degradation to application fidelity. The solution is to dynamically partition the application and offload part of the application execution data to a powerful nearby surrogate. This allows delivery of the application in a pervasive computing environment without significant fidelity degradation or expensive application rewriting. Runtime offloading must adapt to different application execution patterns and resource fluctuations in the pervasive computing environment. This offloading inference engine adaptively solves two key decision-making problems in runtime offloading: timely triggering of offloading and efficient partitioning of applications. Both trace-driven simulations and prototype experiments confirm the effectiveness of this adaptive offloading system.