Flash crowd mitigation via adaptive admission control based on application-level observations

  • Authors:
  • Xuan Chen;John Heidemann

  • Affiliations:
  • University of Southern California, Marina del Rey, CA;University of Southern California, Marina del Rey, CA

  • Venue:
  • ACM Transactions on Internet Technology (TOIT)
  • Year:
  • 2005

Quantified Score

Hi-index 0.01

Visualization

Abstract

We design an adaptive admission control mechanism, network early warning system (NEWS), to protect servers and networks from flash crowds and maintain high performance for end-users. NEWS detects flash crowds from performance degradation in responses and mitigates flash crowds by admitting incoming requests adaptively. We evaluate NEWS performance with both simulations and testbed experiments. We first investigate a network-limited scenarion in simulations. We find that NEWS detects flash crowds within 20 seconds. By discarding 32&percent; of incoming requests, NEWS protects the target server and networks from overloading, reducing the response packet drop rate from 25&percent; to 2&percent;. For admitted requests, NEWS increases their response rate by two times. This performance is similar to the best static rate limiter deployed in the same scenario. We also investigate the impact of detection intervals on NEWS performance, showing it affects both detection delay and false alarm rate. We further consider a server memory-limited scenario in testbed experiments, confirming that NEWS is also effective in this case. We also examine the runtime cost of NEWS traffic monitoring in practice and find that it consumes little CPU time and relatively small memory. Finally, we show NEWS effectively protects bystander traffic from flash crowds.