Open Roasbeef opened 3 years ago
is bt centrality a good metric on the lightning network? Do I even understand the metric correctly in that the more peers are connected to a node the higher its bt centrality score goes up?
This then leads to superhubs, no? But perhaps I don't yet understand bt centrality enough...
What I described above would be about iteratively maximizing your own centrality. So essentially, you want your node to be part of more of the possible set of shortest paths between two nodes. To do this you'd compute your current centrality, then compute the relative gain by adding a new channel to a given peer, then repeat this until you've allocated your funds to your liking.
Do I even understand the metric correctly in that the more peers are connected to a node the higher its bt centrality score goes up?
No, if a peer has 10 peers connect to it but those peers themselves say only have that one connection to that peer, this doesn't affect their betweenness centrality.
It would be useful to add a new order matching perdicate, called
MustMatch
(as an examlpe), that takes a list of public keys and only allows an order to be matched if the other side of the match is amongst that set of public keys. This allows users to attempt to purchase/sell channels to/from an specific node. Such a feature can be used as an example to create a market driven autopilot agent that attempts to greedily maximize a metric such as betweenness centrality by purchasing inbound channels from the set of N nodes that will serve to maximize its personal bt centrality gain.