How We’re Scaling Ethereum Part 1: Replacing Proof of Work

To put it simply, Proof of Work cannot be fast. No network distributed between thousands of unknown entities, with unknown hardware, in a trustless fashion that requires an insane amount of compute power to process even just a few transactions can ever perform well. It’s just physically impossible.

This isn’t meant to bash Proof of Work, it’s done an amazing job of securing such a network, but if performance is what you are looking for, you won’t (and can’t) find it there.

GoChain’s primary goal is to increase performance of Ethereum by 100x (and more in the future), so how do we get there? Well, let’s start by replacing Proof of Work with something more sustainable, something that can’t be centralized based on who has more resources (or more money), and something that has the potential of handling a couple orders of magnitude increase in performance. Something like Proof of Reputation.

I won’t get into all the details of Proof of Reputation (PoR) here as we’re only talking about how it helps scaling in this article (you can read more about proof of reputation here). PoR has a relatively small set of nodes when compared to a Proof of Work network. Why is having a smaller set of nodes better than having thousands? Because you don’t have to keep thousands of nodes in sync. Do you know what the fastest environment is for any networked software? A single machine. Once you go beyond one machine, you have network and bandwidth issues to deal with. When you have thousands of them, these issues compound. If you’re only doing a few transactions per second, it may be OK, but when you’re doing thousands per second, it becomes very hard to keep a large network in sync.

Also, since we trust the participants in the network, we don’t need to waste compute cycles “mining” a useless hash value, we can instead use that compute power actually processing and verifying transactions and blocks.

Now it’s not all roses. Even with a more efficient consensus algorithm, there are still a lot of problems we’ve had to solve (which we’ll cover in the next parts of this series), but it does give us a good place to start from. With PoW we could never achieve our goals, with PoR we can.