Polkadot’s recent spammening is very cool and it shows that Polkadot’s scalability competes with other high throughput chains. Nonetheless, I’m curious how parachains and solochains can push their performance to this limit, so that they alone also compete with other high throughput chains. How can solochains become real-time of the likes of new blockchain platforms being built today?
As a founder of a solochain Substrate project, Tangle (https://tangle.tools), we are thinking about how we can take our Substrate chain to the next level, accelerate finality and tps so that our chain functions as real-time as possible. We plan to support a variety of complex services on Tangle, such as MPC, and being able to rapidly post data onchain and finalize is highly attractive for multi-party protocols. I’d be curious for anyone’s insights and feedback on what we might explore to get to this point.
Of course there will be tradeoffs, and MPC requirements likely introduce some additional factors/requirements - I’m not a cryptographer just an intrested spectator.
One MPC technique that is intriguing are garbled circuits, my idea here was the 2-parties in the computation are the host provider and the guest provider.
Specifically, a GC RISC-V VM. A SodaLabs person indicated such a variant makes sense (conceptually).
It isn’t clear in the following what is being compared to what, but there is at least a claim that a gcEVM can be faster than a ZK approach, again likely not a apples-apples comparison (is it a zkEVM? a VVM? a zkVM?), but intriguing nonetheless:
Garbling protocols already allow for a very efficient client, however, with breakthroughs made by COTI V2, the technology is up to ten times lighter and performs ten times faster than ZK based solutions without negatively impacting the user experience.
Source:
I’m not sure if I intuit where Tangle is heading… is it something like: “with a faster consensus layer we can offer a gcPVM with (sub?) 6-second block times” ?
Yea perhaps it’s simply worth exploring alternative consensus here to speed up our blocktimes and tx processing.
Tangle is mainly an infrastructure for running general compute services, where instantiating these services and reporting results happens onchain. For MPC, being able to quickly finalize what the state of the MPC is (such as everyone posting their public key shares onchain) could be very useful instead of relying on offchain aggregation protocols. You can use the blockchain as a verifiable log to both track good and bad behavior quickly, such as people posting intermediate and non-secret steps in an MPC or other computation (similar to truebit style fraud proofs).