0

This matters a lot for people who want to host a Lightning Network node and will influence LN-transaction fees, because a node-host has to justify the miner fees of setting up channels for others with the income by charging fees themselves.

If a channel-funding transaction is for example 500 bytes in size, it will often cost more than 10$. Sub-one-cent transactions through 5 channels can now only be achieved if more than 5000 transactions pass through each channel before it closes which is quite a lot.

If transactions are that funding transactions are that expensive, centralization is a major concern, because smaller users can't justify paying for 5-10 or even more channels to sustain true decentralization and will rather have one channel to a big hub.

1

not sure what you mean by different size... amount of money/fees? Tx size in bytes is less relevant opening a channel. For this particular use case your use case seems valid. If you "insist" on opening 10$ channels for sub-cent channels (and another 10$ to close), you naturally come to the conclusion, that the network gets centralized. And the options are nearly unlimited to create assumptions/ideas to centralize with Lightning. However you can also avoid centralization. In your example, you have the assumption that 5 channels are opened at the same time, and the 50$ are "illiquid", and eventually lost to miners. Keeping one channel only, is a 10$ loss (+ maybe some LN fees for hopping through the network). So the question is, how many tx you push through the network, compared to "normal" bitcoin tx. You can have a channel opened for a year, and order every month a book. In lightning it looks like a one off cost for opening the channel. In bitcoin this would be 12 fee payments. There was a picure in the net, that I allowed myself to extend a bit:

enter image description here

The three left pictures are limited in a way, that everyone is connected with everyone else. The possibility that you and me just open a channel, close it afterwards, and can even decide when to close the channel (when fees are down) is often not mentioned. Instead assumption is always a short time view. I think a general topic is also the assumption, that fees will stay high forever. There is no proof for this, but as an underlying assumption it (only) could lead to centralization. Taking the assumption, that many tx go into the lightning network, the main blockchain is freed up from 10.000s of low fee tx (see also Jochen's page). This would reduce fees dramatically. We would then have segwit for "more capacity", and lightning for repeating low level tx - future only will show. Same applies for other crypto currencies :-) Really exiting times ahead.

btw: first tx are on mainnet, want to join the revolution?

https://bitcointalk.org/index.php?topic=2815332.0 https://www.reddit.com/r/Bitcoin/comments/7pwna9/lightning_network_megathread/ https://lightning.network/lightning-network-paper.pdf

Bitcointalk discussion here: https://bitcointalk.org/index.php?topic=2854596.0

  • I really hope that it will happen like this. I think though, that it is very unlikely that people will open channels that link them to one other node. It is far more appealing to open a channel to a hub so that you can basically access all of the network using only one channel. When two people want to transfer money daily for a year, they could either form a channel between them for this single purpose or both (for the same cost) link to a giant hub so they can do those daily transactions and also everything else. – Grunzwanzling Feb 1 '18 at 10:15
  • yup indeed. I think it boils down to the question, what "we" as community do with it. And how we understand, what centralization in this sense means. Usually we have some negative concerns with this word "centralization" (most certainly loosing funds or similar). btw: nice nickname, in which forests of southern Germany would I find this animal? :-) – pebwindkraft Feb 1 '18 at 11:39
1

The size of a transaction is no longer relevant. What matters is the weight (which is based on size), especially because LN uses segwit which has different weight values for segwit spends.

The expected weights for the commitment transactions and HTLC success and timeout transactions can be found here: https://github.com/lightningnetwork/lightning-rfc/blob/master/03-transactions.md#appendix-a-expected-weights.

The weight for the funding transaction is harder to know as that entirely depends on the inputs being spent and that is dependent on the sender. Since the inputs and their signatures make up the vast majority of a transaction, it is impossible to say what weight funding transactions will have as they will basically all be different.

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.