The co-founder of Ethereum, Vitalik Buterin, believes that resilience and long-term scalability of blockchain depend on simplicity, such as Bitcoin. In a blog article on May 3, he described how “Ethereum in 5 years can become close to Bitcoin”. Buterin wrote:
“One of the best things about Bitcoin is how beautiful the protocol is.”
According to Buterin, the minimalist design and simplicity of Bitcoin make it accessible, so that even a secondary student can grasp the concept and architecture of the protocol. Simplicity, said Buterin, also brings other advantages, such as reducing the cost of creating new infrastructure and maintenance of existing infrastructure, as well as reducing the risk of bugs.
Recent upgrades such as proof of evidence (POS) and succinct integration not interactive with knowledge (ZK-SNARK) made Ethereum more robust. However, the negligence of the simplicity of design has added to the costs of Ethereum. Buterin explained:
“Historically, Ethereum has often not done this (sometimes because of my own decisions), which has contributed to a large part of our excessive development expenses, to all kinds of safety risks and the insularity of R&D culture, often in search of services that have proven to be illusory.”
Simplification of the Ethereum consensus layer
In November, the researcher of the Ethereum Justin Drake Foundation proposed an upgrade of the consensus layer called “beam chain”. Buterin believes that the beam chain is “well positioned to be much simpler” than its obsolete predecessor, the current chain of beacons.
Indeed, the beam chain will allow a revival of finality to 3 locations, which will eliminate complex concepts such as slots, eras and separate synchronization committees, noted Buterine. He also pointed out that a basic implementation of the finality at 3 locations can be carried out through around 200 lines of code, which makes it much simpler.
The beam chain will also reduce the number of validators active at the same time, which would make it “safer to use simpler implementations of the rule of choice of fork,” wrote Buterin.
The beam chain will also incorporate aggregation protocols based on Stark, which means that anyone can be an aggregator. Buterin noted:
“The complexity of aggregation cryptography itself is important, but it is at least a highly encapsulated complexity, which presents a much lower systemic risk towards the protocol.”
Buterin added that the reduction of active validators and the incorporation of aggregators based on a Stark “will likely allow a simpler and more robust P2P architecture”. He continued by saying that there was an opportunity to rethink and simplify several facets, the entrance to the validator and the exit to the inactivity flight. And this can be done both by reducing the number of code lines (loc) and creating “more readable guarantees”.
Buterin stressed that the consensus layer is “relatively disconnected” from the executions of the Ethereum virtual machine (EVM), which provides a “relatively large latitude” to make improvements compared to the execution layer.
Simplification of the Ethereum execution layer
Last month, Buterin proposed to replace the language of the EVM contract with RISC-V to increase efficiency up to 100x. Buterin argued that the adoption of RISC-V would also increase simplicity, because the “RISC-V specification is absurdly simple compared to the EVM”.
However, this would mean ensuring that compatibility behind for existing applications is preserved. Buterin wrote:
“The first thing that is important to understand is: there is not a single way to delimit what is the” base of Ethereum code “(even within a single customer).”
According to Buterin, the orange area cannot be reduced. The objective, said Buterin, is to minimize the green area, by moving the code to the yellow area, which indicates “the very precious code to understand and interpret the chain today, or for the construction of optimal blocks, but which is not part of the consensus”. Buterin compared this process to the way Apple reaches long -term compatibility backwards through translation layers. He wrote:
“Above all, the orange and yellow zones are a encapsulated complexity, whoever seeks to understand the protocol can jump them, the implementations of Ethereum are free to jump them and that all the bugs of these areas do not present consensus risks.”
This is why the complexity of the code in orange and yellow areas has “much less drawbacks” compared to the complexity of the code in the green zone.
To reduce the green area, Buterin proposed the following steps:
Phase 1: The new precomits will be written in RISC-V.
Phase 2: Developers will have the opportunity to write contracts in RISC-V.
Phase 3: All precompililes will be replaced by RISC-V implementations via a hard fork.
Phase 4: implement an EVM interpreter in Risc-V and push it as an intelligent contract.
The above steps would guarantee that the Ethereum consensus would not understand “natively” that RISC-V, said Buterin.
Protocol standards for simplification
Buterin proposed to share “a standard on different parts of the battery” as a path to simplification.
For example, Buterin suggested using a single erasure code for data availability sampling, P2P dissemination and distributed history storage. This would minimize the total code lines, would increase efficiency and ensure verifiability, he argued.
Likewise, he proposed to have a unique shared serialization format on the three Ethereum layers: execution layer, consensus layer and binary interface of intelligent contract application (ABI). Buterin suggested using SSZ, which is easy to decode and widely used.
Finally, once the EVM has been replaced by RISC-V or another simple language, Buterin offers to move to a binary tree of the Hexary Merkle Patricia tree, both for the layers of consensus and execution. This transition could improve efficiency and reduce costs while ensuring that all Ethereum strata can be accessible and interpreted using the same code, wrote Buterin.
A change of ethics
Buterin concluded by proposing that Ethereum, following the example of Tinygrad, adopts an explicit maximum code target. The objective, reiterated Buterin is to make “Ethereum Consensus-Code Critique close to Bitcoin as simple as Bitcoin”.
But more importantly, Ethereum must adopt an ethics where the simpler option is chosen as far as possible. This would mean promoting the encapsulated complexity compared to systemic complexity.
Buterin reassured that the code which deals with the treatment of the historic rules of Ethereum will continue to exist with its last proposal. However, this code must be kept outside the consensual critical code or the green zone.
