0
From the Bitcoin Lightning Network:
it is possible to find a path across the network similar to routing packets on the internet.
Is routing approach of Bitcoin Lighning Network performed in a "decentralized" and "trust-less" manner ?
Also, the fact that in the Bitcoin network there is no routing table does not matter ?
1
Is routing decentralized and trustless:
Generally, yes. It uses sphinx onion-routing to prevent any peer along the way to know origin/destination of payment.
Also note that Information of channel balances can leak if routing attempts are used to probe liquidity along a route.
Routing is not robust against sybil attacks, as compromised nodes can determine routes by reconstructing path via payment_hash in the htlc output scripts.
It is, however, incumbent upon the channel operator to ensure the lightning implementation is always synced with a full-node. Only this way can it detect a cheating channel counter-party and broadcast a penalty transaction. Watchtowers may provide vigilance as a service but they are not trustless
Also, note that the lightning network is a separate p2p network.
Discovering topology occurs with peers gossiping.
Thank you , Could you please say that if "separate p2p network" causes a kind of centralization or trust-based? If you could introduce a reference/paper/website on explanation of "separate p2p network" it would be perfect. Thanks – Questioner – 2018-11-30T10:14:46.530
1I cant really say that because there are no guarantees of centralization nor decentralization. Some argue liquidity for channels is expensive so it will aggregate in the hands of exchanges/individuals with high BTC accumulated. So nodes with high channel count activity will tend to have more payments touted though them, there may be a centralization pressure, yes. However, same goes with mining pooling and large merchants on the Bitcoin network. – James C. – 2018-11-30T10:19:09.143
You can say that Lightning is trustless in the sense that you can always penalize invalid channel closures (given you see them before the timeout). – James C. – 2018-11-30T10:20:33.327
Thank you , And is there a good reference explaining the term of "separate p2p network" ? Thanks again – Questioner – 2018-11-30T10:54:35.577
1
Sorry this is a little heavy but checkout the peer chapter in the lightning protocol rfc: https://github.com/lightningnetwork/lightning-rfc?files=1 Note that the lightning network “interfaces” with the Bitcoin network via Bitcoin transactions (open, commitments, close).
– James C. – 2018-11-30T10:59:52.423Do you mean this one? : https://github.com/lightningnetwork/lightning-rfc/blob/master/02-peer-protocol.md
– Questioner – 2018-11-30T11:06:16.830Thats the p2p messaging protocol, yup. – James C. – 2018-11-30T11:08:25.253
Thanks , In this paper: (https://bit.ly/2RpuavY) it's claimed that: "... belief is that routing tables are necessary for large operators as for current Internet routing algorithms as BGP or Cjdns protocols. " AND: "existence of a group of beacon nodes (even when selected randomly and renewed regularly) presents a threat to decentralization. Indeed, even the global knowledge of the geometry of the network can be a vector of attack." Do you agree with this assertion? Thanks
– Questioner – 2018-11-30T11:27:14.4101Thats always the trade-off - a routing table via distributed hash table is slow, but decentralized. Central Beacons (even if somehow selected deterministically) are more performant, yet centralized. – James C. – 2018-11-30T11:29:33.460