What is Layer 3? \ stacker news ~bitcoin

pull down to refresh

844 sats \ 36 comments \ @l0k18 22 Jan 2023 bitcoin

There is a general consensus that sidechains and LN/sidechain type protocols are Layer 2 on top of the Bitcoin base layer, layer 1.

Many people, including our dear Jack think that the competition is still open to define THE layer 2.

The reality is that Bitcoin will work best when its blocks are filled with HTLCs and building on top of LN gives you a lot more options, you can anchor in the base layer, as needed, such as Fidelity Bonds, and/or you can anchor in the next layer, and provide spam resistance via the use of LN microtransactions.

The main achievement of Bitcoin is finding a way to both stop spammy transactions and distribute the privilege of publishing blocks as widely as possible. The block reward is just an inducement to compensate for the inevitable volatility in the early stages.

When you build on top of LN, you can create protocols that leverage the economics of bitcoin's perfect monetary policy, that can scale freely without any limit, stopping spammy activity, and anchoring your protocol's hierarchy of "stake" in whatever way suits your use case.

Unlike most Proof of Stake blockchains, where it has been proven that early adopters tend to be immovable from their top positions, the liquidity of LN allows a more fluid consensus based on unforgeable sats.

Bitcoin is the foundations. LN is the walls and supporting structures. Layer 3 is the roof which allows us to control the environment in the space between.

view all related items

231 sats \ 1 reply \ @Merlin 22 Jan 2023

I think stuff like fediments are layer 3

0 sats \ 0 replies \ @l0k18 OP 22 Jan 2023

Fedimint is depends directly on layer 1 so it is a type of layer 2. The guarantees are weaker, mints can abscond with your tokens, whereas it's a much tighter competition with LN channels and much harder to cheat the protocol.

I can see some of the reason behind it but one of the core assumptions of the design is that network connectivity need not be operational to validate a payment, only cached keys. I think there's something a little pessimistic in this view and unrealistic considering you can literally have an internet connection to anywhere on the planet without anything but a modem.

17 sats \ 3 replies \ @lightcoin 23 Jan 2023

There is a general consensus that sidechains and LN/sidechain type protocols are Layer 2 on top of the Bitcoin base layer, layer 1.

I would not say there is general consensus here e.g. https://twitter.com/gakonst/status/1146793685545304064

In the same way that L2 = new layer built on L1 that inherits full L1 security while moving tx execution off L1, I would apply more or less the same definition to L3. L3 = new layer built on L2 that inherits full L2 security while moving tx execution off L2.

Users can securely move coins from L1 to L2, from L2 to L3, back to L2, then back to L1, as the need arises.

10 sats \ 2 replies \ @l0k18 OP 23 Jan 2023

Yeah, I agree, it's hard to argue that a crummy cosmos sidechain bound to Bitcoin with a kludgey solution to finality agreement does not really inherit the security. It sems to me like it would compound rather than sum, as in, the weakness of both would be the result rather than inheriting the strength of the "lower" layer.

Since there isn't a stronger blockchain than Bitcoin, side chains only reduce security, in direct proportion with their real decentralisation.

LN, by contrast, is a p2p protocol rather than a global broadcast protocol, and it rides the optimum path in terms of game theory and cryptographic security. Which is why I say, there is no other layer 2, to ape Saylor's "there is no second best". LN thus does not have weaknesses in the same way that Bitcoin has, thus it can be said it adds security rather than compounds its weaknesses.

Really, I think an immediate disqualifier for a protocol to be called layer 3 when it rides on top of LN (and maybe tickles the main chain a bit for such as bonds), is that the network is aiming to achieve a similar goal.

A side chain is just another shitcoin chain, and another LN network is going to lag the main LN swarm by anonymity set and liquidity.

Thus, a layer 3 could not be merely a payment network, and cannot either be a global broadcast based network, since it is supposedly using the strengths of the two best at this type of task. It would have to be a data storage/retrieval or routing protocol or some combination of these two.

I think a bearer certificate based securities exchange protocol could also qualify as L3 since it has a substantially different goal, it is not intended to be a money, but a way to distribute profits back to investors.

Off the cuff I can't think of how to make such a system work, or what kind of consensus it would use, but it seems to me like this might actually be a case where the Proof of Stake style model could work, based on the fact that the stake is not issued by any party in the network, which is the root of Proof of Stake as a L1 platform.

I'll try to keep my mind inside the box of thinking about an L3 that provides anonymity though. Someone else will eventually figure out the same thing if they don't read my comment or have the idea relayed to them from it.

10 sats \ 1 reply \ @lightcoin 23 Jan 2023

it's hard to argue that a crummy cosmos sidechain bound to Bitcoin with a kludgey solution to finality agreement does not really inherit the security.

If you are referring to Nomic and Babylon chains, they are using bitcoin to improve their own security (clearly so, solving the weak subjectivity problem) but not really "inheriting security" in the same way that an L2 does.

LN thus does not have weaknesses in the same way that Bitcoin has

What do you mean by this? LN does have a different security model than L1, introducing the dependency of the fault proof ("penalty tx") to prevent double-spending, making LN payments vulnerable to short-term miner censorship or L1 congestion.

I think an immediate disqualifier for a protocol to be called layer 3 when it rides on top of LN (and maybe tickles the main chain a bit for such as bonds), is that the network is aiming to achieve a similar goal.

Does LN not aim to achieve a similar goal as bitcoin L1 (p2p payments) only in a more scalable way? In any case, it seems arbitrary to me to say "this layer has the same use-case, so it's not really a new layer". imo the layer terminology should be applied independently of the use-case of the layer, and look moreso at the technical architecture and relationships to other protocols to determine whether it's really a higher layer or something else.

10 sats \ 0 replies \ @l0k18 OP 24 Jan 2023

Yes, sidechains benefit from bitcoin but that doesn't mean anything if a small group of the userbase has outsized control over token issuance.

Bitcoin's weakness relates to reorgs and its slow, broadcast propagation pattern. Lightning's weakness is griefers and expiring/stuck channels.

I don't think what bitcoin does is at all similar to LN. There is a much weaker consensus in LN, because it doesn't have to share the state changes, only the p2p information for generating paths and a summary at the end of each channel, the security problems for LN come from mischief with channel partners and channel liquidity.

17 sats \ 9 replies \ @angusp 23 Jan 2023

I'd argue things like LNURL and Cashu are layer 3

10 sats \ 8 replies \ @l0k18 OP 23 Jan 2023

Cashu is definitely a layer 3 as its protocol depends on being connected to a LN node with channels to the swarm. I'm dubious about its real use value due to the problem of disappearing mints, but yes it's a layer 3.

LNURL? Isn't this part of most LN node implementations? Isn't it really just a messaging system that uses the p2p network of LN? It interfaces to LN but it's really just an extension of the peer communication layer.

0 sats \ 7 replies \ @angusp 23 Jan 2023

LNURL is a separate spec from Lightning-proper, even if many/all nodes are using parts of it. Lightning has BOLTs, whereas LNURL has LUDs.

Because LNURL depends on HTTP stuff not the Lightning peer protocol, I'd say that's enough to call it a layer 3 -- you're using it as a separate protocol to make it easier to use and interoperate with lightning, which as a metaphor is like using DNS to make doing things over IP easier

0 sats \ 6 replies \ @l0k18 OP 23 Jan 2023

Ah, makes sense. And literally part of that relates to DNS, since LNURLs usually are tied to DNS addresses, though on the LN layer they don't have to be literally answering to that address to advertise it. I presume there is some kind of verification to stop false advertising in the protocol that probably involves a DNS record.

I'm not all that interested in name resolution systems, per se, except as far as to say that it would be great if there was a full layer 3 DNS registry built on bitcoin!

If things don't settle down between NATO and Russia there could even be a root split in DNS, Russia has got its own, separate system ready to go for this case. With a Bitcoin/LN powered name service, the entire "web5" ecosystem could be human friendly and not subject to the whims of thus far light touch on domain registration, which many governments can use to force DNS servers in their jurisdiction to force redirect and essentially steal a domain from a user for whatever incomprehensible reasons they do this for, DMCA being a common example in the past.

A DNS type registry running on Bitcoin/LN would also be able to provide more name types, and include hidden service addresses for anonymity networks with this capability, which would be pretty amazing.

17 sats \ 5 replies \ @astonvilla 23 Jan 2023

Don't you think a Namecoin sidechain on bitcoin could be easier and more censorship resistant than LN to address such hard task as permissionless global DNS ?

10 sats \ 4 replies \ @l0k18 OP 23 Jan 2023

With the relatively high latency that a name service requires it could be implemented as a sidechain, using anchor transactions in a similar way as the Cosmos IBC but adapted to cope with the probabalistic finality. Really, a bitcoin transaction is beyond practical reversal within the span of a day, which matches up with the existing latency of DNS database changes.

But hell, why can't we make it faster than that? I dunno about your experience doing web development and hosting but that delay is a pretty dang annoying thing.

0 sats \ 3 replies \ @astonvilla 23 Jan 2023

You could have some kind of self hosted electrum server for domains, that could be faster than requesting a remote peer

0 sats \ 2 replies \ @l0k18 OP 23 Jan 2023

Of course, and an SPV like Neutrino also.

Also just want to remark, that the 10-60 minute latency of on-chain confirmations of such data is still better than the 3-24 hours that you get with DNS. But why be satisfied with that when updates to the state can happen at the speed of Lightning???

view replies

17 sats \ 9 replies \ @notgeld 22 Jan 2023

An example of layer 3

https://notgeld.medium.com/lightning-network-computation-layer-27c7ba81a214

10 sats \ 6 replies \ @ama 22 Jan 2023

Correct me if I'm wrong, but isn't that exactly what RGB does/is?

17 sats \ 2 replies \ @notgeld 22 Jan 2023

No, because LN computation layer shouldn't be touching chain at all.

0 sats \ 1 reply \ @ama 22 Jan 2023

True. RGB is defined by their authors are layer 2/3.

0 sats \ 0 replies \ @l0k18 OP 23 Jan 2023

I think you can say that a simple on chain element to a protocol doesn't make it, itself, into a layer 2, such as fidelity bonds on nokyc markets, or any type of protocol that needs objective timestamps, but rather whether it routinely reads and writes data to the chain, read only most of the time doesn't really make it a layer 2.

I'd say the distinction is more about where the bulk of payment activity occurs in the protocol. If it's primarily transacting on LN then it's layer 3.

0 sats \ 2 replies \ @tlindi 22 Jan 2023

Seems yes. And etleneum IS DEAD PROJECT! Do not refer! Not belive, go their TG channel to see owner comments 22.1.2023@22:30GMT ish

0 sats \ 1 reply \ @ama 22 Jan 2023

I think that's good news. I don't remember the exact details, but I somehow remember I didn't like it at all when I found it first. I'm still hanging around in their TG group, though, hopping to never see the smallest sight of it coming back to life. LOL

0 sats \ 0 replies \ @notgeld 22 Jan 2023

Etleneum was great. There was nice Simple Auction Contract and Swap contract.

0 sats \ 0 replies \ @tlindi 22 Jan 2023

Has etleneum done Fenix?

0 sats \ 0 replies \ @l0k18 OP 22 Jan 2023

Now that's what I'm talkin about!

Indranet is also layer 3, its service is anonymity. I would argue that it's a fundamental service since many attacks in other systems can be evaded this way.

1 sat \ 7 replies \ @kr 22 Jan 2023

i get the idea of building on bitcoin in layers, but i don’t understand why we call them layer 1, 2, and 3.

this naming system gives newcomers no clues as to how the bitcoin ecosystem actually works.

sometimes people call lightning a payments layer and the base blockchain a settlement layer which makes sense to me, are there any useful terms for describing the activity that happens on layer 3?

27 sats \ 3 replies \ @roythomp68 23 Jan 2023 freebie

The highest "layer" built on Bitcoin will be defined as the "application" layer. However, whether it ends up being the 3rd, 4th, or 5th layer etc. is still TBD. All layers sitting below the application layer will be defined as infrastructure layers. This is similar to the current internet stack, which consists of a link layer, a network layer, a transport layer and followed by an application layer (the highest layer).

Prior to Lightning and other 'layer-2s', all that existed was the Bitcoin basechain. At that point in time, the layer-1 blockchain was the application layer, enabling participants to transact with one another. However, the creation of the Lightning network brought another infrastructure layer to Bitcoin, adding additional functionality, beyond just transacting value.

Each incremental infrastructure layer should improve the functionality and potential number of applications that can be built on Bitcoin. Once the the marginal utility from adding infrastucture layers approaches zero, no more layers will be added. At this point, we'll be able to define what the ultimate "application layer" on Bitcoin will be.

Maybe it will be layer 3 or 4, however, there is no reason to think that it couldn't be layer 6, 7 or 8. Only time will tell...

0 sats \ 2 replies \ @l0k18 OP 23 Jan 2023

HAHA. I'm sure that OSI will invent at least 7 when it goes fully mainstream. I think you can say that the same logic applies even to internet OSI layers, in that first there was the link (point to point) then there was IP, then TCP, and at first HTTP was an application layer.

Really, whatever is implemented by an application is the application layer. Applications can be built on top of other applications, so the distinction becomes a bit vague.

17 sats \ 1 reply \ @roythomp68 23 Jan 2023

Agreed, distinction will always be vague.

If we are forced to put structure around it though, I think that framework is reasonable, despite its flaws.

0 sats \ 0 replies \ @l0k18 OP 23 Jan 2023

Yeah, it just gets out of date as people keep stacking more things on top of what was previously called "application" as they interop them and the upper layers of the cake get more and more indistinguishable.

I think that really, even TCP/IP, as a network layer, is an application. How can you say that netcat is not an application? Or telnet, both of which sit directly above TCP/IP. Indeed, TCP is on top of IP, in fact, it's an application for reliable message transport. The network itself is an application!

Anyhow, la la la la so boring the grandstanding of academics who didn't invent the protocols to make pronouncements about other people's work.

17 sats \ 0 replies \ @notgeld 22 Jan 2023

Hosted and fiat channels are good candidates for L3. They provide verifiable proofs although they are custodial.

0 sats \ 0 replies \ @Kaffi 23 Jan 2023

is layer 2 because in order to enter the Lightning ecosystem you need to acquire bitcoin first to fund a channel. Layer 1 comes before layer 2. if there is a layer that then requires layer 2 like stablecoins then in my opinion thats a layer 3 protocol.

you can't get to the second layer of an onion without removing the first.

0 sats \ 0 replies \ @notgeld 22 Jan 2023

Something happening inside LN payment settlement? (up to 15 mins - - hodl invoice)

0 sats \ 0 replies \ @nullama 23 Jan 2023

Hype

0 sats \ 0 replies \ @tbonesteaks 22 Jan 2023

Butterscotch chips or condensed milk depending on how you count it.