Charles Hoskinson Cardano Update - Ouroboros Hydra - 9/3/2020
(Written by @Eric_Czuleger)
On March 9th 2020, Charles Hoskinson sat down for an update where he discussed key information about Ouroboros Hydra.
-The Ouroboros Hydra paper has been published on eprint, this concludes the research schedule of Ouroboros.
-Simulations prove that each Hydra head can perform 1000 transactions a second while each stake pool will be able to host a head.
-This fits with the extended UTXO model and gives us a natural way to shard the stake space without sharding the ledger itself.
-We will now translate the science into code and bring it into the system.
-Rob Cohen will act as product manager for Hydra and ensure that monthly updates are given.
From 2015 on we have been following the same research agenda. In this, the Cardano project sought to make proof-of-stake practical for all users. This required us to look at the entire science of proof-of-stake and consensus protocols. We needed to come up with a practical solution which would allow us to gradually scale as we gained users. Essentially, as the system got more decentralized it would have to perform at the same rate or faster.
This was a large task given all of the other considerations such as adaptive security, semi-synchrony, fast finality, the ability to bootstrap from genesis without a checkpoint, decoupling the clock, and the ability to recover from spikes of dishonest majority. We also needed to build delegation and stake pool systems. Meanwhile we had to be able to orchestrate and curate all of this while handling a way of networking our for our entire industry.
The road to research
This has been a colossal task which has consumed 5 years of research from a team of two dozen scientists and engineers. This has resulted in over 1,000 citations and a corpus of knowledge. The capstone of this research has been Ouroboros Hydra. We finished the initial build of this in November. We submitted the conference built on February 15th. Today it has been published to eprint. This means that for the first time ever, the Hydra paper is available to the public.
This is an isomorphic state channel paper. It is a layer two solution which can sit on top of the Cardano protocol and it fits with our stake pool model. Simulations show that each Hydra head can perform around 1000 transactions per second and there is a lot of room for optimization of this. Furthermore there is evidence which shows that as we add heads to the protocol we should achieve near-linear scaling. Perfect linear scaling is impossible due to overhead.
Fitting into Cardano
This model fits well with the extended UTXO model because it gives us natural ways to shard the stake space without sharding the ledger itself. This means that if stake pools were to create a head, then as we add more stake pools we will get more heads. The result will be performance upscaling. If we have 1000 TPS and 1000 heads there could be a maximum theoretical performance rate of 1 million TPS. This would also allow fast finality like in traditional BFT protocols. This also allows for microtransactions. In addition, we benefit from years of research in state channels and payment channels for routing and interoperability.
We are excited to bring this to market because it is the first interdisciplinary paper where we had network, MPC (Multi-Party Computation), programming language, consensus people working together to figure out the theory. We are thrilled that the smart contracts which can run on the base layer are isomorphic and they can run on the state channels. This means that if you have a Plutus smart contract in our model it can run in our head one-for-one. It is exciting to show that this is possible.
There is clean up and prototyping which needs to be done. But it is exciting that this can be implemented in parallel with all of our work on Shelley, Goguen, and Voltaire. This is a solution which plugs on top of our model likely adopted first by the stake pools, we can have an independent team to work exclusively on the Basho side scaling. Secondly, it is complementary to ledger based sharding solutions.
The main team will be working on implementing the Shelley and Goguen features and continue to upgrade the network stack. We believe that the max theoretical performance unsharded on the base ledger will be around 1000 TPS on the base ledger. Each stakepool will then be able to run heads and add an additional 1000 TPS according to the design. This can be seen in the paper. More heads equals more performance. This brings us to Visa scale at the base layer, and at the head level we can scale to a global system which is natural and easy to shard. Those heads can be used for interoperability with other networks like the Bitcoin network and Lightning network. We will also be able to layer payment systems on top as well as a decentralized exchange mechanism.
To get to this capstone we have seen our language group prove isomorphic properties. We have also seen our extended UTXO model demonstrate successful space sharding, as well as development in the theory of consensus. The theory of Ouroboros matured to where it is comparable to bitcoin after recovering from a dishonest majority, to bootstrapping from genesis to operating in a semisynchronous environment to adaptive security. This is a major scientific milestone for the industry as a whole. We are proud to publish this paper.
The first research line and our first five years of effort is now coming to an end. We solved the problems we set out to solve for Cardano. This makes us a truly scalable system to bring it into the future. Now, we have to translate that science into real code and bring it into the system. Fortunately there is an enormous amount of work that has been done thanks to lightning and state channel research. We can also now engage a parallel team to start that implementation work in tandem with Shelley in Goguen.
When this is pulled together Cardano will truly be the fastest blockchain in the world. Also, it will contain all the capabilities of our competitors from fast finality, to micropayments and easy interoperability. We thank our supporters in their patience for this publication. We encountered some difficulty translating our paper for conferences into the one which appears on eprint and we had to go through a significant editorial process.
The product manager who will begin updating us on this will be Rob Cohen. Once the Shelley node is finished cleaning Rob will begin working full time with the implementation team building Hydra and bringing it in. He will be providing monthly updates. As always we look forward to adding this as a value proposition and revenue generator for stake pool operators. We believe that this will accelerate Cardano into the dawn of the Basho Era. This is the light at the end of the tunnel for the entire roadmap.