pull down to refresh

I'm sure we could argue to the end of time about what constitutes a layer 2, but it's good to read through how other people see these things.

Here's Spark's taxonomy:

  • State Channels (Lightning)
  • Federated Sidechains (Liquid)
  • Merge-Mined Sidechains (Rootstock)
  • Statechains (Spark, Mercury Layer)
  • Virtual UTXO Protocols (Ark, Arkade)
  • Proof-of-Transfer Chains (Stacks)

It's interesting what they exclude: notably BitVM, Rollups, and ecash -- they do mention these at another point in the article, but it's a very brief mention.

Later on under the section "Trust Assumptions Compared" they present the following table:

ProtocolWhat You TrustWhat You Can VerifyFailure Mode
LightningNothing (trustless with watchtowers)All channel states on-chainMust monitor chain or use watchtower
LiquidFederation majority is honestL-BTC supply via auditFederation collusion can steal funds
RootstockPowPeg federation + minersMerge-mining attestationsFederation compromise risks peg
StackssBTC signer set (14 signers)All Stacks transactions on-chainSigner collusion risks sBTC peg
SparkAt least 1-of-n operators is honestPre-signed exit transactionsAll operators offline: liveness failure, not theft
ArkOperator does not double-signVTXO branch proofsOperator equivocation before on-chain confirmation

In particular, I think that the statement that Spark's failure mode is only "liveness failure, not theft" is a bit strong. It seems to me that there are scenarios where a Spark operator could take your sats, although I haven't thought about it in a while and can't produce any examples at the moment.

All in all, I think Bitcoin Layers did a better job of this comparison.

103 sats \ 0 replies \ @ynniv 28 May

with spark ...

you trust: 1-of-{ Lightspark }, because they're the only operator
you can verify: pre-signed exit transactions, but not whether they've deleted keys
failure modes: availability, censorship, privacy, griefing, and yes, potentially theft

reply

This is NOT scaling... this is enshitification !

reply

Spark is taking over sadly

reply
reply