Fair Real-Time Traffic Scheduling over a Wireless LAN

  • Authors:
  • Maria Adamou;Sanjeev Khanna;Insup Lee;Insik Shin;Shiyu Zhou

  • Affiliations:
  • -;-;-;-;-

  • Venue:
  • RTSS '01 Proceedings of the 22nd IEEE Real-Time Systems Symposium
  • Year:
  • 2001

Quantified Score

Hi-index 0.01

Visualization

Abstract

Unpredictable wireless channel errors may cause applications with real-time traffic to receive degradedquality of services due to pac ketlosses.In the presence of such errors, a challenging problem is how toschedule packets to achieve fairness among real-timeflows and to maximize the overall system throughputsimultaneously.We capture fairness by minimizingthe maximum degradation in service over all flows.In this paper, we show that no online algorithm canguarantee a bounded performance ratio with respect tothe optimal algorithm.We then compare four differentonline algorithms and evaluate them using simulations.The first two are EDF (Earliest Deadline First) andGDF (Greatest Degradation First) that consider onlyone aspect of our scheduling goal respectively.EDFis naturally suited for maximizing throughput whileGDF seeks to minimize the maximum degradation.Thenext two are algorithms, called EOG (EDF or GDF)and LFF (Lagging Flows First), that consider the twoaspects of our scheduling goal. EOG simply combinesEDF and GDF, whereas LFF tries to favor lagging flowsin a non-trivial manner. Our simulation results showthat LFF is almost as good as EDF in maximizing thethroughput and also is better than GDF in minimizingthe maximum degradation. Finally, we also show thatthere is an optimal polynomial time algorithm for theoffline version of the problem.