RChain founder Greg Meredith and blockchain scientist Atticbee joined Foresight Institute of Vision Weekend 2021 for a keynote presentation. Thought leaders from all over the world will gather together to share ideas of the latest developments of decentralized cooperation, biotech, vanguard art, molecular physics, neurology, artificial intelligence, etc and cross silos to brainstorm for more innovations. With its unique concurrent computational model and correct by construction approach, RChain has the answers to many of the questions of future intelligent cooperation.
Concurrency, Rho-Calculus and Blockchain Scalability
As a keynote speaker, Greg Meredith gave a talk about what concurrency means to the blockchain scalability.
Greg Meredith opened his talk on a question asked by others: what technology has not reached mainstreet yet, but it should be known to blockchain practitioners in order to achieve scalability and security? Concurrency tends to be the blindspot of our thinking. However it particularly showed up in the blockchain setting. As we know most of the financial transactions are isolated. Just think about it, someone buys empanadas from a street vendor in Santiago and another one buys grilled tofu in Shanghai, these two transactions are almost certainly unrelated.
The reason the current financial system can scale is they have this concurrency built in. If we funnel all transactions down into one lane, or a single thread, then the system can never scale.
Greg Meredith gave an example: if we have a 8-lane highway with not many lane crossings, then we can achieve 8 times the throughput of 1 lane. However if we funnel them down into 1 lane, then cars get back up, which is exactly what we see in Ethereum.
What computational model can provide a good, clean and simple way to deal with this isolation and concurrency? It turns out this family of computational models only came into being during the early 90s when the Pi-Calculus was first discovered. After that its related research exploded and many varieties of calculus were introduced. The founder of Pi-Calculus, Robin Milner, later won the Turing reward because of his finding of Pi-Calculus.
Pi-Calculus answers one question. The sequential Lambda calculus underlies functional programming languages Lisp, Ocaml etc. So the question is, Lambda calculus is to the sequential computation, is like what to the concurrent computation? Pi-Calculus is the answer. Greg Meredith however, found a tiny “bug” in the Pi-Calculus. When he fixed that bug, the calculus obtained a “massive explosion” of expressive power. And the upgraded version of Pi-Calculus is Rho-Calculus, which is the heart of Rholang, the smart contract language used in RChain.
One of the benefits of using Rho-Calculus is that it allows RChain to scale linearly. Unlike other systems where adding nodes lead to slower performance, adding nodes to RChain actually increases the throughput. Greg Meredith argues this style of scaling by adding more resources like RChain is how we have been scaling computing systems for the last 50 years. We just couldn’t just bet and pray for a software genius event to improve the performance. What we can rely on is to throw more resources in order to get more work down.
Currently, without much optimization in place, RChain already gets about 1,000 TPS per node per CPU. So if its network consists of 10 nodes and each node has 10 CPUs, the throughput will be 10x or 10,000 TPS. Similarly, 20 nodes with 20 CPUs each will produce 20,000 TPS. So on and so forth, until it hits about 1,000 nodes at which it has no resources to test, but at that time the network will reach 1 million TPS. That’s what concurrency can do to the scalability of blockchain.
More about RChain
RChain breaks the infamous blockchain trilemma of scalability, decentralization and security by introducing concurrency to the blockchain world for the first time. It is the only leaderless blockchain to allow safe and atomic concurrent execution of smart contracts, both inside and across shards, and therefore achieves linear scalability and seamless cross shard cooperation. This feature is the result of 30+ years of research in concurrency and process calculi.
Based on the revolutionary Rho-Calculus, RChain solves a series of problems preventing blockchain platforms from mainstream adoption.
Its fast and scalable conflict detection algorithm accompanied with Casper CBC consensus allows: all nodes to produce and verify blocks concurrently without global epochs so it becomes the first smart contract platform to achieve single-shard scalability; and – large data to be stored directly on chain, removing the dependency on other data storage solutions such as IPFS.
Built on Rho-Calculus, RChain is the first chain to allow any complicated cross-shard transactions to be verified and finalized atomically and concurrently by all involved shards. This means the cross shard transactions can be done seamlessly and safely just as within a single shard.
It will also be equipped with a Rho-Calculus based behavioural type system. This allows smart contracts to be formally verified quickly in a concurrent and sharded setting, which makes possible the orchestration of large quantities of smart contracts.
Its unique reactive smart contract system is more suitable than other chains’ active systems for time sensitive applications such as DeFi.
RChain’s unique technology makes it the best candidate to build a functional world computer and is a key technology to achieve the next level of intelligent cooperation of humanity.