A Modest Proposal': Vitalik Unveils Multi-Year Vision for Ethereum

What do you do after you've created a multibillion-dollar cryptocurrency?

A skinny, 23-year-old hacker in a green "Doge" t-shirt gave us an answer today. At ethereum's flagship conference, Devcon, project creator Vitalik Buterin revealed he has been quietly working on a new long-term plan for the future of the blockchain network. What he called a "modest proposal," it's perhaps better described as a three-to-four-year roadmap for ethereum's technical development.

Notably at the heart of the vision is a long-in-the-making technical change to ethereum called "sharding," and while always expected to be included in the protocol's plans, today Buterin proposed what might be his most solidified strategy for the technique to date.

As such, the roadmap hints at problems yet to be solved on the platform, and as the emphasis on scalability for project developers. As ethereum nodes need to store everything that ever happened on the network, Buterin stressed that there's a need for solutions that mitigate expensive storage costs that could escalate exponentially as the system expands.

It's a topic that's long been top-of-mind for the developer, as Buterin recently released new research into alleviating this problem.

Still, the talk was evidence of his emphasis on finding solutions, and of his efforts to galvanize ethereum developers more broadly to be thinking about the effort.

"The amount of activity on the blockchain is orders of magnitude larger than it was just a couple of years ago," he said, pointing to daily transaction rates and the more than 20,000 nodes now part of the network.

With this, he suggested ethereum is running up against its limits.

Buterin told the crowd:

"Scalability is probably problem number one [...] There’s a graveyard of systems that claim to solve the scalability problem but don't. It's a very significant and hard challenge.
These are just known facts."

High-level details

And Buterin believes sharding is the "likely" solution to this problem.

A way of partitioning data into subsets that takes its inspiration from traditional databases, the idea is that each node will only have to store a small chunk of the total network. Yet, the vision is that the underlying math would hold the system accountable, and if they need it, nodes could rely on other nodes for data.

How to execute on this in practice, and securely – without nodes sending other nodes false information – is another question that researchers have been looking into.

But Buterin is proposing a new type of sharding infrastructure that would solve both scalability and governance – ensuring the eventual system is well maintained and that it stays in check.

The proposal revealed today is for ethereum to be split into different types of shards. There will be the main shard, which would comprise today's ethereum network. Then there would be other shards, which Buterin calls other "universes."

Crucially, though, Buterin believes the partitioning would allow for more aggressive changes on the smaller shards, and more cautious changes on the main blockchain. That way, ethereum still has platform stability, while developers still have room to test new changes and to experiment and move fast on the other shards.

Or as Buterin put it:

"Other universes where all this stuff we’ve been working on these last few years can be rolled out much much faster."

Looking forward

Buterin's roadmap includes other changes too, though they were less prominent in his talk.

These included planned upgrades to the ethereum virtual machine (EVM), the technology that today compiles smart contract code and communicates it to the network. He also addressed another long-in-the-making tech project, eWASM, for running ethereum in a web browser, one that hints at the necessity of ensuring this system given that the EVM has been implemented in other blockchain projects as well.

Another idea proposed was for so-called "stateless clients," a proposal for how clients could sync with the network more quickly.

"You'll be hearing about this idea more and more," he said. He invited developers to contribute to the effort, much of the research of which is housed on GitHub.

But, all in all, sharding looks to be the biggest change over the next three to four years, and Buterin ended by adding there's already developer work going on in these exploratory areas.

Notably, he hinted work might be further along than widely thought.

Buterin concluded: