For anyone who has not been living under a rock, scaling is the focal point. It has become the theme of the year and will be focused on the next couple protocol proposals on Tezos. We’ve already witnessed the introduction of Tenderbake
, Transaction Optimistic Rollups (TORUs
), and more recently Smart Contract Optimistic Rollups (SCORUs
In Nomadic Labs’ initial blog post
on the scaling roadmap for Tezos, these features were hinted at. There was also a snippet hinting at further avenues being researched and explored beyond strictly rollups.
While optimistic rollups are a good scaling solution for Tezos, they do currently have limitations.
One issue is that all transaction data for the rollup inbox must be included in blocks on the main chain, but there is a limit to how much data can be stored in each block.
Fast forward to today, we have been introduced to the first sidechain live
on the Tezos network. Developed by Marigold
, Deku offers a WASM VM (“webassembly virtual machine), faster throughput, and a native bridge to Tezos (made possible thanks to Michelson Tickets). This is a big leap forward for enthusiastic and serious projects, partners, and businesses that are building on Tezos.
If you’re a project on Tezos looking for high throughput, low fees, be able to develop with your own language and more… then Deku could be a solution for you.
Key Insight from Marigold
I had the chance to sit down and chat with Daniel Hines of Marigold and discuss all things Deku and scaling the Tezos network.
Can you explain the motivation behind working on Deku?
Sure, so Tezos is really cool because it’s this fully decentralized network and uses some pretty novel Proof of Stake design. Especially as of Jakartanet where Tenderbake was introduced, Tezos is quite fast compared to many Layer 1’s. But, even though Tezos is fast; it’s still nowhere near as fast as traditional Web2 centralized technologies.
So, imagine if you were to try to run YouTube on a blockchain. There’s just no way to run internet scale applications when you’re on a maximum transactions per second threshold of about 50. The VISA network is supporting 20–30 thousand transactions per second and until we get there, we won’t be able to handle internet scale applications on blockchains.
Layer 2 applications are addressing this problem of scalability. Now that we have the Layer 1 forming consensus in this fully transparent way, there are interesting ways that we can use that to achieve bigger scale without being bottlenecked by Layer 1.
What are the differences between scaling applications like sidechains and rollups?
The way that is on everyone’s minds right now is rollups. The idea behind that is to essentially use the Layer 1 to form consensus on a set of transactions in a particular order. But, you use the Layer 2 to form consensus on what the results of those transactions are.
The great thing about this setup with optimistic rollups is you only need a single honest node. Essentially, your rollup network can be secured by a single honest participant. This in turn, allows one to scale the compute power of the network. Certain things that are very computational intensive can be moved to a rollup and benefit.
However, there’s one thing rollups do not solve and that’s bandwidth. The order of operations is still being settled on the Layer 1 and there’s still limited bandwidth there. Sidechains on the other hand, take a different tradeoff than rollups. They run a full blockchain and that blockchain can have a wide array of design options. It can have a different consensus algorithm, bridge technologies, virtual machines, etc.
At the core, a side chain is any one of those blockchains with a light client embedded in the smart contract of a Layer 1 blockchain. What this forms is a bridge between the two blockchains. In the case of Deku, we have a consensus algorithm that is forming consensus around blocks but then periodically those blocks get posted to Tezos. Tezos has a witness of our consensus embedded in it so it can check the validity of these blocks.
That’s kinda the rough idea of sidechains. Unlike rollups which are secured by one honest participant, sidechains are secured by whatever the parameters for your consensus are. So, if we wrote Bitcoin as a sidechain it would only be secured if 51% of the participants were acting in an honest way. In our case, we’ve chosen the tendermint protocol. So, a Deku network is secure as long as 2/3 + 1 of the participants are acting correctly.
What prompted you to choose Proof of Authority for Deku?
Unlike Tezos which uses Proof of Stake; we’ve chosen Proof of Authority for Deku right now. Proof of Authority is a lot easier to implement and thus a lot easier to make sure the implementation is correct. It’s also very fast unlike say that of Proof of Work. So, that’s kind of the tradeoff. This is way faster than Proof of Work and way easier than Proof of Stake but the thing it trades is centralization.
You essentially have to trust that 2/3 + 1 of these participants are being honest. But, that’s not as bad as it sounds. What people don’t take into account here is the value of identity. Suppose that Marigold runs a Deku node in a Deku network, we’re effectively staking our name and reputation. So, you trust Marigold’s name and trust us then if we mess up; we have something at stake.
It’s not as immediately transparent as our Tezos tokens at stake but it’s our identity that is at stake and that’s very valuable to us. The idea of Deku is to run this network and distribute the validation keys across a network of authority that you trust. One would have to collude across organizations to damage the network. Those organizations have a lot at stake and one would be able to prove that there was collusion after the fact.
So, that’s the idea of Deku’s consensus and with this bridge between them; Deku’s consensus is being witnessed by Tezos and that provides some extra security. It basically provides a schelling point for people to be able to check the state of consensus. For Tezos to lie about the state of Deku’s consensus, the entire network would have to be compromised and that would be quite hard with billions of dollars at stake securing the Tezos network.
While the tides have gone to ease within traditional and crypto markets, the level of adoption and development has done nothing but accelerate. For Tezos, the focus on Layer 2 applications like Deku will remain the focal point for the rest of this year as we progress further along to the next protocol amendment following Kathmandu’s activation.