pull down to refresh

Ark looks really promising https://www.arkpill.me/
Interoperable with Lightning but seems much better privacy and scalability.
What do you guys think? Any downsides?
So from quick read and chat with Burak, here's what I got so far (but take with grain of salt, I'm not that smart):
  • You first send your onchain coins (UTXO) to one of the pools (Ark Service Providers) and it becomes virtual-txo (comparable to having sats on lightning).
  • The pool coinjoins all vtxos every 5 seconds generating new onchain transaction for every coinjoin (but it's using taproot smartness, so every coinjoin is represented only by a simple 1 input, 1 output transaction taking ~130vB).
  • So this means that each pool will be submitting ~15k vB to mempool every 10 minutes (so there's a limit for number of pools running at the same time - if my math is correct there can be AT MOST 256 pools). But also note here that the numbers are just a suggestion here, the "5 seconds" could change to e.g. 1 minute.
  • When you have your bitcoin in this layer, then you can freely transact with other folks, the transactions happen in 5 seconds and with solid privacy - as far as I can tell the cryptography should prevent the pool from actually being able to link your transactions.
  • Any time you want you can pull your coins out of the pool and send to onchain address.
  • This solves the "I don't have inbound liquidity" problem that lightning has.
It seems that it could be much easier to run Ark node compared to LN node, since you just need to fund it with BTC and the daemon can do its job from there, but also your funding only seems to be going in one direction, so your coins are locked for at least 2 weeks and you may need to be adding more gradually (?). If it works the way I understand it can be quite profitable to run Ark Service Provider node.
reply
"each pool will be submitting ~15k vB to mempool every 10 minutes (so there's a limit for number of pools running at the same time - if my math is correct there can be AT MOST 256 pools). But also note here that the numbers are just a suggestion here, the "5 seconds" could change to e.g. 1 minute."
256 pools will utilize all the bitcoin tx throughput... and fixing it only a bit by waiting 1 minute for transaction to happen instead of 5 sec - is like going back from L2 to L1 "slow motion" - unacceptable user experience, then
that's mad idea at first look, imo
reply
thanks for write up
who pays for constant stream of onchain tx?
it sounds like its good solution for privacy. swap into layer to transact, not store funds indefinitely
reply
The Ark Service Provider (ASP) has to pay the fees for those transactions and so that will be passed onto the users. But the trick is that the more users use it, the cheaper it becomes. i.e. users can make 1000 ark transactions and they split only a single onchain fee for 1in-1out tx.
reply
You wouldn't need to run an "ark node". Idea is that this improves UX that much that it is feasibe to have just a mobile Ark wallet.
reply
Great write up!
I think limits for number of pools that get into a block is huge improvement over LN.
What would be the limit on number of transactions within single coin join (which happens say each 5seconds)?
reply
Nice, attacking bitcoin strengthens it. In this case, attack is ordinals and BRC-20, and strengthening part is the effort to invent scalability solutions like Ark or Tbdxxx. Or perhaps those solutions just get more attention now, not sure.
edit: just saw another post claiming that Ark == Tbdxxx. Seems plausible, contact email is Burak's.
reply
Yeah I made that post because it needed to be said. TBDXXX was a stand in name until Burak came up with a real name. Making a post declaring that new name was something I felt was important for the future of discussion around it.
reply
Yep, its new name for Tbdxxx
reply
Sounds too good. I need to study more in order to have an opinion. But I do not reject it.
reply
I don't get the principle of it from the text. It was not easy to get LN either, even after reading the WP. I'm building a similarly complex but quite different protocol and it's also pretty hard to explain it simply at first, but after you run it around the block a few times you get better at explaining it. I see something about these virtual UTXOs but honestly every time I see any idea these days my eyes glaze over about something to do with how they are lying about the consensus or that there is no consensus and it's actually federated, which is cheap.
reply
"The only downside is that Ark require users to come online and "refresh" their coins every few weeks, otherwise the ASP can sweep the funds."
  • Directly from that link.
Yikes. But at least they're honest about that downside. I'm intrigued by what I've read so far.
reply
Everything comes with it's own set of trade-offs, on-chain BTC is best security in the world but with great cost and low scalability, lightning is great security, infinite scalability, but still big cost (funding and liquidity management), ark is giving up some security for bringing the cost even lower for allowing even smaller tx's bypassing liquidity management problems. The way I see it is on-chain BTC is for very large tx's, Lightning is for medium tx's, Ark is for small tx's. But you can use or not use anything you want, that's the beauty of open and permissionless systems like bitcoin.
reply
a lot of people have a lot of time and capital invested in lightning, so I can imagine this not getting the attention and adoption it might deserve
I don't know enough about it but do love the fact that it demonstrates the beauty of open protocols and anyone can build on top of them.
look forward to seeing more
reply
Agreed that it might not get a lot of attention. There's an awful lot of momentum on Lightning's side.
It's a very common trope that the best technology often doesn't win.
reply
thanks to @nout:
"So this means that each pool will be submitting ~15k vB to mempool every 10 minutes (so there's a limit for number of pools running at the same time - if my math is correct there can be AT MOST 256 pools)"
do you really think that's "best technology" ? I'm not sure... 256 pools? and they said long time ago: "LN will be centralized"...
256 pools and we have no space for today's transactions in the block, which is fully occupied right now...
reply
I think 256 pools competing is plenty. Right now we have Wallet of Satoshi and Alby:-)
As long as this is open and new pools can join without any approvals (just plebs using it), I think its ok.
Also, smaller pools might find a way to join together to get in the block and compete with large pools (and increase privacy).
reply
implying that lightning may not be the best, it doesn't follow that this new protocol ark is
reply
I don't mean to imply anything, I don't know anything about ark. Just saying there's a possibility something better comes along and nobody uses it.
reply
yes I agree with you
reply
I see it less about something that competes with lightning and more something that competes with fedimint.
That said, if it requires 4 week respending then it really would be trivial for a single service provider to steal enough user funds IMO. At that point not much better trade off than fedimint, though maybe worse because fedimint requires multiple parties in a federation collude.
I'm also concerned about the heavy on chain footprint that Ark requires meanwhile fedimint requires none if just dealing with lightning.
reply
Very good point!
I think another difference is that Ark seems to be trustless. Ie at any time I can move my funds to on-chain, without anyones permission. That's not the case with Fedimint.
reply
Yeah that's true. It really is an interesting solution with different trade offs compared to LN/Fedimint. However, I think when it comes to LN, there's going to have to be LN integrations needed with both Ark/Fedimint given that LN is the L2 that has gotten the most adoption while maintaining trustlessness. A considerate amount of effort will be required to have a "flippening" moment, not simply "this protocol might be better than LN". There's an entire ecosystem of use cases and developers to consider.
reply
Since this is another ARK thread I can shill my https://github.com/fiksn/awesome-ark here too I guess :)
reply
Ark isn't "better" than Lightning
It is complementary
Ark is like subnet on Lightning
reply
If Ark works, why would anyone use Lightning? What would be the use cases where Lightning is better?
I think a channel with my friend that we often pay each other would be better served by Lighting. What else?
reply
With Ark fast off-chain transactions are limited to your ASP "island". Lightning then connects those "islands" together so you can also pay your friend who is using a different ASP offchain. I think there will never be the situation with just one Ark Service Provider. And if there was, such a centralization is bad...
reply
You saying with Ark I can only pay within ASP island? I don't think that's true. Can someone confirm this? I thought ASP is basically a coinjoin coordinator but you can switch the coordinator at anytime.
reply
I mean you can pay anyone with the vTXO, but in my understanding that recipient implicitly also becomes a "user" of that specific ASP. You trust that ASP to not do something nasty (like double-spend). If it does you don't lose your funds but it still gets expensive since you need to unilaterly exit to base layer. The better way to sweep if you have funds with ASP1, ASP2 and ASP3 whom you trust is to do an AMP lightning payment.
reply
I would use Lightning, 'cause already was fully tested and have so much capital invested in her. Accept fast payments without need create a new IOU, isn't necessary soft fork, etc.
reply
This requires a softfork, right?
reply
Nope. With softfork you can do transactions when a peer is offline, but the basics work on current bitcoin.
reply