%@page contentType="text/html" %>
If enabled, your router will periodically anonymously probe some of your peers to see what sort of throughput they can handle. This improves your router's ability to pick faster peers, but can cost substantial bandwidth. Relevant data from the load testing is fed into the profiles as well as the test.rtt and related stats.
If you can, please poke a hole in your NAT or firewall to allow unsolicited UDP packets to reach you on your external UDP address. If you can't, I2P now includes supports UDP hole punching with "SSU introductions" - peers who will relay a request from someone you don't know to your router for your router so that you can make an outbound connection to them. I2P will use these introductions automatically if it detects that the port is not forwarded (as shown by the Status: OK (NAT) line), or you can manually require them here. Users behind symmetric NATs, such as OpenBSD's pf, are not currently supported.
This setting causes your router identity to be regenerated every time your IP address changes. If you have a dynamic IP this option can speed up your reintegration into the network (since people will have shitlisted your old router identity), and, for very weak adversaries, help frustrate trivial intersection attacks against the NetDB. Your different router identities would only be 'hidden' among other I2P users at your ISP, and further analysis would link the router identities further.
Note that when I2P detects an IP address change, it will automatically initiate a restart in order to rekey and to disconnect from peers before they update their profiles - any long lasting client connections will be disconnected, though such would likely already be the case anyway, since the IP address changed.