On Lightning Nodes Centrality
Lightning Network is one of the key upcoming technologies that will change Bitcoin for the better, it’s a network of nodes with channels between these nodes that process payments and transactions “off-chain”, to keep only the few important transactions on-chain and decrease the load on the blockchain and the time it takes to send a transaction.
However there is a potential for node centrality in the lightning network, the main issue being that some nodes might end up functioning like banks, processing the bulk of these transactions. Developer Kulpreet Singh set out to measure the centrality of nodes in the Lightning network.
Treating Lightning as an undirected graph, he focused on three metrics. He used Centrality which is a measure of how often a node is in the shortest path between two nodes and Central Point Dominance which is the node with highest centrality, both defined in the seminal paper by Lindon C. Freeman 1. He also measured Articulation points which are nodes whose failure would partition the network, as per Robert E. Tarjan 2.
The Analysis was done on the 24th of August and the results were quite
outstanding. The Central Point Dominance was low at
0.1668 on a scale of 0 to
1. 0 being that it is not included in any shortest path and 1 being that it is
included in every shortest path in the network, the top twenty nodes ranged
0.1668 to 0.017.
There were quite a few articulation points, the metric used was that an articulation point must connect two components of at least five nodes each, there were about 20 nodes found ranging from 24 components to two components.
These are quite satisfying results as it turns out the Lightning network is currently not as centralized as people think, Lightning is still very new to the main net and hasn’t been used by the majority of Bitcoin users, just a few days ago the capacity of the network surpassed 100,000 BTC for the first time, we expect centralization to further decrease once it gains more adoption and more nodes and channels are created.
- “A Set of Measures of Centrality Based on Betweenness”, Sociometry 40, pp. 35-41, 1977. [return]
- “Depth first search and linear graph algorithms” SIAM Journal on Computing, 1(2):146-160, 1972. [return]
- https://twitter.com/renepickhardt/status/1034066602273193985 [return]
- https://github.com/ElementsProject/lightning/pull/1888 [return]
Support us and the authors of this article by donating to the following address:3MtoUDbbkS2SnpPrFXrKFAtABJom6NgZSp