ICMP echo request/reply is handy for determining if a route exists at all ...
(reads article - I've got a five digit /. ID and that was after lurking for several years - respond first, ask questions/read article later)
Oh. You now fail to understand networks in Rust instead of C/Python/nicker elastic. sighs in policy based routing tables.
A modern mtr (traceroute is so 90's) should do things like run up and down the stack for each point along a route. It will still probably need to use the TTL field to find each point (IP) but then use ICMP/TCP/UDP/etc to measure that point in some way or perhaps interpolate it from points either side.
When I want to really get to grips with latency and stuff, I start off with a small dedicated box on a customer network and "smoke ping" with all points measurable on the path. I also have several running from our datacentre and a fair few RIPE Atlas probes too.
traceroute is handy but you must be able to decipher what it is telling you. Wearing a stethoscope does not make you a doctor.
(reads article - I've got a five digit /. ID and that was after lurking for several years - respond first, ask questions/read article later)
Oh. You now fail to understand networks in Rust instead of C/Python/nicker elastic. sighs in policy based routing tables.
A modern mtr (traceroute is so 90's) should do things like run up and down the stack for each point along a route. It will still probably need to use the TTL field to find each point (IP) but then use ICMP/TCP/UDP/etc to measure that point in some way or perhaps interpolate it from points either side.
When I want to really get to grips with latency and stuff, I start off with a small dedicated box on a customer network and "smoke ping" with all points measurable on the path. I also have several running from our datacentre and a fair few RIPE Atlas probes too.
traceroute is handy but you must be able to decipher what it is telling you. Wearing a stethoscope does not make you a doctor.