How many ⚡ channels could we open in a single block? \ stacker news ~bitcoin

pull down to refresh

How many ⚡ channels could we open in a single block?

1999 sats \ 6 comments \ @bitcoinbumbleB 23 May 2023 bitcoin

I mean theoretically, in terms of data usage, assuming ideal UTXO management, etc.

And also—are there any operations (rebalancing for example) that require writing to the chain, besides closing the channels if needed?

(This is a 'scaling' question)

Thanks

view all related items

152 sats \ 2 replies \ @DarthCoin 23 May 2023

Is hard to estimate and even know. Even if all txs are yours, opening a bunch of channels, you never be sure if all your txs are included in the same block.

Rebalancing do not require writing on the chain, are circular rebalancing not swaps onchain <-> LN.

This is a 'scaling' question

You will not scale Bitcoin with everybody opening a channel. There are many ways to onboard a lot of people. Not everybody will run a public routing LN node. Is not really necessary for that.

As I explained here, with "uncle Jim" community/family banks:

0 sats \ 1 reply \ @bitcoinbumbleB OP 23 May 2023

Thanks!

How much does one channel open take then? How heavy is that type of TX with minimum numbers of inputs?

In terms of scaling, I'm just interested in sovereign use. I know that there are solutions with custody trade-offs too.

0 sats \ 0 replies \ @DarthCoin 23 May 2023

How much does one channel open take then?

3 block confirmations. Some other solutions offer even 0conf (Phoenix, Blixt).

How heavy is that type of TX with minimum numbers of inputs?

Here you can calculate yourself: https://coinb.in/#fees

I know that there are solutions with custody trade-offs too.

I think you don't know about many self-custody LN nodes, here are explained:

10 sats \ 0 replies \ @k00b 23 May 2023

I know protocol devs discuss this quite a bit. If you know enough about the protocols you can get the theoretical bounds pretty easily. Tadge Dryja touches on this some on this excellent podcast from chaincode: https://podcast.chaincode.com/2023/03/20/tadge-dryja.html ...

At some recent austin bit devs there was a transcript shared about core devs discussing "onboarding the world in a week to bitcoin." I'm struggling to find it.

20 sats \ 0 replies \ @shibe 23 May 2023

Assume that all inputs are Taproot inputs (the cheapest input to spend) and since Lightning channel outputs are 43 bytes (Taproot won't change this) with each participant having an output channel, then the maximum channels you could open in one mega-transaction is around 9942 (just under the 1MvB limit, the transaction is 999181.5 vBytes). Such a transaction would be non-standard of course (current node rules don't relay transactions this large, but you can still mine them)

Now assume there is only 1 input. Then the maximum is around 23,000 outputs/channels (999435 vBytes)

All this is assuming all the transactions are aggregated though. Assuming each channel open was its own transaction then the limit is probably around a few thousand or so.

In the future I envision large Taproot conjoins like the first example into outputs that could be multisigs, singlesigs or Lightning channels (nobody can distinguish these with Taproot). Maybe someday we'll have cross-input signature aggregation as well, increasing this limit even higher. Hopefully we'll also have either hierarchical channels (google these, exciting stuff) or eltoo to move even more of Lightning off-chain

0 sats \ 0 replies \ @joda 23 May 2023

average : 2000 max theoretical : 5000 taproot max : 9000 one-to-many taproot : 23,000 (or they argue 23,000 channel factories)