It would be handy to have a standard way for an instance to offer to act as a relay for another instance, indicating how long it's willing to do so, and how much extra bandwidth it's prepared to commit to the task.
Revisions and Commits
|R1:d2919d895979 Allow latency reports for a subset of requested destinations|
|R1:e9a879b35cfc Defined null relay request|
|R1:f9c321a11c0a Defined relay offers|
|R1:a814531eb8a2 Invitations use pre-defined bandwidth limits|
|R1:31c5fbef2a50 Bandwidth limits apply proportionately to bundles|
|R1:e389559e455d Estimating bandwidth usage is sufficient to respect a limit|
|R1:56bb31f9af26 Defined bandwidth limit|
|Open||None||T1 Cheap, private payments via a decentralized network|
|Open||tim||T2 Finalize a first draft of a specification for a decentralized payments protocol|
|Open||tim||T12 Revise introduction to the specification|
|Resolved||tim||T5 Finish Messages section of the specification|
|Resolved||tim||T38 Relay offers|
Additionally, in order to decide whether to accept a relay offer, the recipient might want to know the latencies from the sender to the recipient's peers' relays (and whether the sender is able to communicate with those relays at all). A special relay request could be used for this, with the initiator's public key set to 0.