discv5: additional text around NAT hole punching

[emhane]

Theory around the actual hole punching: keeping a hole punched, how a node discovers if it is behind a NAT and some on how the handling of holes maintains properties held by the p2p network.

[emhane]

Now, it is ready for review. If you may @fjl .

[fjl]

Right now, it isn't super clear from this text what the implementation requirements are. The text has both descriptions of implementation features and rationale-type content, where you explain 'common situations' and their outcomes.

Please add a section that clearly states the logic that implementations should apply. Specifically, please describe separately the rules for NAT traversal initiator, relay, and recipient.

[emhane]

Right now, it isn't super clear from this text what the implementation requirements are. The text has both descriptions of implementation features and rationale-type content, where you explain 'common situations' and their outcomes.

Please add a section that clearly states the logic that implementations should apply. Specifically, please describe separately the rules for NAT traversal initiator, relay, and recipient.

Ok, done, let me know what you think.