Skip to content

Per Node Failure Table

Stephen Oliver edited this page Mar 30, 2017 · 1 revision

We track, for a limited period (up to an hour), which peers have requested a particular key, and which peers we have requested it from. This enables us to limit the load from applications repeatedly polling the same keys (particularly chat clients), and to propagate that key quickly when it is found.

See ULPR for a detailed description.

Clone this wiki locally