A Systematic Simulation-based Study of Adverse Impact

Best effort applications over non-TCP protocols, e.g., UDP, can be used in attacks that capture unfairly large share of bandwidth compared to TCP flows. While earlier studies may have pointed out that short-lived TCP flows (mice) may hurt long-lived TCP flows (elephants) in the long term, no insight was given as to developing scenarios leading to drastic decrease in throughputs of long-lived TCP flows. We have systematically developed TCP attack scenarios that differ from all prior research in that we use short-lived TCP flows to attack long-lived TCP flows. Our attacks are interesting since, (a) they are more difficult to detect, and (b) they point out the increased vulnerabilities of recently proposed scheduling, AQM and routing techniques that further favor shortlived flows. We systematically exploit the ability of TCP flows in slow-start t