Re: Network Latency Measurements

2012-12-05 Thread Harsha V. Madhyastha
As part of the iPlane project, we have been gathering traceroutes daily for the last 6.5 years from a couple of hundred PlanetLab sites to over 100K prefixes. Data for the last six months is available at http://iplane.cs.washington.edu/data/iplane_logs/2012/ and all older data is at http://ipla

Re: Traceroute explanation

2011-12-08 Thread Harsha V. Madhyastha
Another explanation could be load balancing. As Fred mentioned, traceroute sends out different packets for different hops on the path and since these packets have different headers, load balancers on the path may hash packets with different TTL values on to different paths. Check out http://www