13

As pointed out by many lightning developers in the past (e.g. Rusty's 1 million channel challenge) the gossip protocol is quite noisy and a part of the BOLTs causing many engineering challenges.

Today, I learned that Ethereum's Raiden network does not have a gossip protocol. Since all channels in Raiden are smart contracts that live in the Ethereum blockchain, Raiden nodes are supposed to learn about the network topology by looking at the Ethereum blockchain.

I believe funding transactions could have been designed to store information about Lightning nodes. Why was that not done?

Reasons I see:

  • With our design we can have private channels
  • blockspace is valuable, so we did not want to store more info on chain than necessary
  • we needed a transport and communication protocol between nodes anyway, so why not add the Gossip protocol?

My feeling is that there must have been other reasons for that design decision. So what am I missing?

  • 12
    You seem to have answered your own question! Lighting implementations could work a different way (and even be compatible with existing ones...) but the current way has advantages. Sticking the same gossip data into a global consensus protocol would make it massively less scalable, not more. – G. Maxwell Mar 12 '19 at 22:25
  • 3
    Oh that is a good point! Gossip data also has information about routing fees and cltv data so it would even be more data to be stored on chain. In particular this data can change while the channel is operational. Totally oversaw that when asking the question. I thought there might have been other reasons. If noone provides them I will edit the question and answer it myself including your hints. Thanks for opening my eyes. – Rene Pickhardt Mar 13 '19 at 0:14
  • 5
    On Etherium, it is almost impossible for a regular user to fully sync the Etherium blockchain to begin with, so I imagine their Raiden network is effectively unusable without relying on third parties for information. – Mark H Mar 13 '19 at 14:48
  • 2
    I think this is an interesting question, but it would be better if you separated your question from your partial answer. By giving a partial answer in the question, you discourage other answers and force potential answerers to address your theories. This question could be improved by posting the partial answer as an answer instead which would open up the scope for other answerers to respond more freely. – Murch Aug 11 '19 at 18:43
1

The main idea of Lightening Network is to make the Transactions as fast as possible and to solve blockchain Network scalability issues. Whereas the main essence of the Gossip protocol is that each participant randomly selects a peer and exchanges state. As it can be seen that Gossip approach is converging in O(logN) gossip rounds where N is the number of participants. Whereas communication on off-chain Network is only O(1) hence reducing the space on the blockchain. Applying Gossip protocol to the lightening Network on defeats the property of 'Lightening' and makes the network massively less scalable. Hope you like the answer.

| improve this answer | |
-1

I have no idea why the decision was made, but I can say why I would have made the same decision.

Gossip protocols are a well established way to find peers in p2p networks. Peer discovery is not very suitable for a permanent store like a blockchain, since peers frequently appear and disappear from the network and the network topology constantly changes. It makes sense to keep balances on-chain for settlement, but it's not an appropriate store for ephemeral data. Entirely too expensive and not scalable.

btw, sorry if this isn't a proper answer, but this new account doesn't have enough rep to comment. :shrug:

| improve this answer | |

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.