Steve, Thank you very much for making this tool. It's very encouraging from our (ie. the Atlas team's) point of view to see people making useful tools based on the network.
I encourage you to continue your work, perhaps even collaborate with others in an open source fashion to scale up, and also let us -- the Atlas team -- know if we can be of assistance. Regards, Robert On 2018-07-16 14:15, Steve Gibbard wrote: > I wrote front-end to traceroute from the RIPE Atlas probes. It looks like a > standard looking glass — you select a probe by location and AS number, enter > a destination, and it does a traceroute. It’s on the web at > https://www.globaltraceroute.com. If this looks useful, please check it out > and tell me what you think. > > One of the things I've found frustrating when troubleshooting routing > problems was the lack of information about inbound paths. Various > measurement systems would tell me when performance was bad. Traceroutes from > my own network would tell me what path traffic to a destination was taking > outbound. Flow systems would tell me what interface inbound traffic was > coming in on, and sometimes what peer it was coming through. But determining > the full path inbound traffic was taking — why users of some ISP in Asia were > having their traffic show up at one of my POPs in Europe, for instance, was > much more difficult. > > I’ve been using looking glasses and commercial performance monitoring systems > that allow traceroutes from their probes, but those often weren’t where the > end users were. RIPE Atlas did have probes where a lot of my end users were, > so I started configuring one time measurements on RIPE Atlas whenever I > needed a traceroute from a simulated end user. But finding a suitable probe > and configuring the measurement was too cumbersome to do when I wasn’t pretty > desperate. This is my attempt to solve that problem. > > Thanks, > Steve >
