Understanding and Implementing Traceroute in Rust

I've embarked on a journey to understand traceroute, a tool I've used but never fully grasped. Traceroute is a network diagnostic tool that traces the path packets take from a source to a destination, using a clever trick with the Time To Live (TTL) field in IP packets. By sending packets with incrementing TTL values, traceroute receives ICMP 'Time Exceeded' messages from routers along the path, revealing each hop.

To delve deeper, I decided to rewrite traceroute in Rust. The core idea is simple: send UDP packets with a low TTL so they expire at each hop, then listen for ICMP replies. I started by creating a function to send a UDP packet with a specified TTL and wait for an ICMP reply. Using Rust's socket2 crate, I set up a UDP socket for sending and a raw ICMP socket for receiving replies.

The process involves sending packets to a high port (33434) and listening for ICMP 'Time Exceeded' messages. The code reads the source IP from the ICMP reply to identify the router that dropped the packet. I implemented a loop to increment the TTL from 1 to 15, printing each hop's IP address and stopping when the destination is reached.

I added timing to measure round-trip time for each probe, enhancing the output with latency information. Real traceroute sends three probes per hop to account for network variability, reliability, and load balancer detection. I mirrored this by wrapping the probe call in a loop, tracking the last IP seen to maintain clean output.

Despite my implementation's success, it lacks some features of real traceroute, like DNS reverse lookup and TCP mode. Traceroute's output can be misleading due to asymmetric return paths, MPLS tunnels, and ICMP rate limiting, which cause the infamous '* * *' hops.

Running the code requires sudo due to the use of raw sockets, a privileged operation. This exploration has deepened my understanding of network diagnostics and prepared me for future challenges, like dealing with network restrictions at work. The code is available on GitHub for those interested in exploring further.