We allocate such a large address space per user because the source address itself can serve as a measurement ID. This gives users flexibility in how you can structure their experiments: a distinct source address can mark an entire campaign for easier grouping and analysis; separate addresses per vantage point make it straightforward to identify the origin of anycast measurements; and varying the source per destination prefix, or even per individual probe, enables studies of load balancing, path variability, or the behavior of middleboxes.
0 commit comments