Experimental Evaluation of an Adaptive Flash Crowd Protection System

Chen, Xuan and Heidemann, John
USC/Information Sciences Institute


Xuan Chen and John Heidemann 2003. Experimental Evaluation of an Adaptive Flash Crowd Protection System. Technical Report IIS-TR-2003-573. USC/Information Sciences Institute. [PDF]


Network early warning system (NEWS) is an adaptive admission control scheme that protects server and networks from overloading during flash crowds, and maintains high performance for accepted requests. Unlike other admission control systems, NEWS regulates requests by observing response performance, automatically adapting to changing traffic mixes. We have previously studied NEWS through simulation; this paper presents an implementation of NEWS on a Linux-based router and evaluates that implementation in testbed experiments with HTTP server log recorded during a flash crowd. This paper has three contributions. First, we use the implementation to evaluate scenarios not considered in simulation. In addition to validating our previous simulation results in network-limited scenario quantitatively, we further consider server memory-limited scenario, confirming that NEWS is effective in both cases. Second, we evaluate the run-time cost of NEWS traffic monitoring in practice, and find that it consumes little CPU time and relatively small memory. Finally, we extend core NEWS algorithms to include hot-spot identification function to protect bystander traffic from flash crowds efficiently.


  author = {Chen, Xuan and Heidemann, John},
  title = {Experimental Evaluation of an Adaptive Flash Crowd Protection System},
  institution = {USC/Information Sciences Institute},
  year = {2003},
  month = jul,
  sortdate = {2003-07-01},
  project = {ant, saman, conser},
  jsubject = {www},
  number = {IIS-TR-2003-573},
  jlocation = {johnh: pafile},
  keywords = {NEWS, implementation, admission control},
  url = {https://ant.isi.edu/%7ejohnh/PAPERS/Chen03b.html},
  pdfurl = {https://ant.isi.edu/%7ejohnh/PAPERS/Chen03b.pdf},
  copyrightholder = {authors},
  myorganization = {USC/Information Sciences Institute}