An idiot’s guide to the Ethereum Update

A breakdown of the roadmap with no cryptography (promise)

The folks working at Ethereum are smart… really smart. So much so that they seem incapable of dumbing down their ideas. And it doesn’t seem intentional. It’s just that their tech is so complex that even the best efforts to simplify it results in explanations that are seemingly comprehensible only by those with advanced degrees in math and computer science.

There are a slew of podcasts, blog posts and medium articles that do a great but rather convoluted job of breaking down Ethereum’s plan to upgrade to a Proof of Stake consensus mechanism and improve the network’s rather poor scalability. Which is all well and good, but what if, like me, you’re an idiot who wants to know just what the plan is without learning the minutiae of every single step in the roadmap? Let’s see if I can help.

The main goal of Ethereum’s upgrade is this: Increase scalability without sacrificing the network’s decentralization and security. And how does it plan to do this? By increasing the network’s throughput while making it easy for anyone to run a node (by reducing the hardware requirement for doing so). Everything I’m going to talk about from here on out is geared entirely towards this goal. But before we get to the crux of the matter, let me give you a breakdown of what’s going to happen in the coming months and years:

Vitalik has broken down the planned update into 5 stages, although the last 3 of these stages are non-sequential and will be implemented as and when they are ready. Also, these 5 stages happen to rhyme:

But before I get into all of this, let’s get the fundamentals sorted out or you’re going to feel really lost (trust me, I’ve been there). So let’s start at the bottom of the pyramid.

The Blockchain Trilemma

There are 3 core requirements for a public blockchain: Decentralization, Scalability, and Security. The trilemma faced by blockchains today is that any attempt to improve two of these aspects results in compromising the third.

Ethereum today is highly decentralized and secure, but it is in no way scalable. Since every node on the network has to verify every transaction, the blockchain’s throughput is low, only around 15–40 TPS which is paltry compared to rival layer 1 blockchains like Solana and Avalanche (which have their own problems!)

Recognizing the need to improve Ethereum’s throughput, the core dev team have identified a path forward, one whose foundations are built on Rollups. As we delve into Rollups and all the stuff that will go into making them work better, remember this one quote by Vitalik:

“True scaling isn’t just TPS. It is throughput relative to cost to validate.”

Low cost to validate here means that running validator nodes doesn’t cost a bomb in terms of hardware and financial resource requirements. This is turn increases decentralization. Now, back to the matter at hand.

Rollups

Ethereum transactions are expensive. Prohibitively so. Each transaction requires its own gas, whose fees can skyrocket due to high demand. But there’s another thing that Ethereum can process: data.

What Rollups essentially do is they take a bunch of transactions from the Ethereum, push them off-chain, bundle them together, and send the data back to the Ethereum mainchain. This way, the gas fee is only paid on a bundle of transactions, instead of each individual one.

This process is carried out by a single party known as a sequencer. Now if you’re wondering how this is decentralized, the short answer is that it is not. But you don’t really need the sequencer to be decentralized if: a) There are methods in place to check whether the sequencer is committing fraud (Optimistic Rollups) or b) The sequencer also submits a proof validating its transaction data as authentic (ZK Rollups).

Breaking down the two types of rollups and their features will require a long write-up of its own (which is coming soon), but for our purpose, here are the basic differences between the two.

The sequencer has to be a powerful computer to handle the kind of volumes a rollup is expected to deal with. When this happens, the efficacy of the rollup is limited by the amount of data it receives and by the amount of data it can send to Ethereum. And that brings us to sharding.

Sharding

The Ethereum mainchain currently serves as an execution layer, a data layer and a consensus layer. The consensus part will be handled by the proof-of-stake Beacon chain when Ethereum finally merges with it. And the execution part will be handled/is being handled by L2 solutions like rollups and networks like Polygon (Ethereum devs are still discussing implementing sharding on execution). Which leaves us with the data.

To increase data availability, Ethereum plans to implement data sharding. The data layer will be split into a bunch of “shard chains” (64 as of now), which will massively increase throughput and prevent network congestions (which leads to high fees). Vitalik Buterin estimates that while rollups with compression can hit 1500–4000 TPS, rollups with sharding will be able to reach 100,000 TPS.

The extra data from sharding means that nodes on Ethereum will have higher resource requirements to deal with all of it. But what if they didn’t have to check all the data to verify its veracity? That’s where Data Availability Sampling (DAS) comes in.

Put simply, DAS is a technology that allows nodes to check whether all the rollups data has been published by checking random pieces from it.

There’s yet more to this sharding business, but it gives me a headache so I’m going to speedrun it.

1. Danksharding: A new and improved form of sharding designed by Ethereum researcher Dankrad Feist (Guinness nominee for most memorable first name). It does away with a fixed number of shard chains and needs a feature called Proposer-Builder Separation (PBS) to work.
2. Proto-Danksharding: A proposal to implement most of the logic and “scaffolding” (eg. transaction formats, verification rules) that make up a full Danksharding spec, without actually implementing any sharding. Comes from the mind of Diederik Loerakker aka protolambda.
3. Proposer-Builder Separation: Consensus nodes (miners/validators) today serve two roles — building the actual block and proposing it to other consensus nodes who validate it. PBS splits these roles and creates specialized block builders who assemble blocks, then bid for proposers (validators) to select them. Validators can efficiently verify blocks thanks to DAS. PBS also requires the implementation of Censorship Resistant Lists (crLists) to prevent builders from screwing around and censoring transactions.

Now we have a basic understanding of rollups and sharding, but how do they fit into the Ethereum roadmap? Well, rollups are already live, and sharding is the main goal of one of “The Surge” stage. But now that the fundamentals are sorted out, it’s time to go through the roadmap and see how it all fits together.

1. THE MERGE

In a few weeks or months, the Ethereum mainchain will merge with the consensus layer Beacon chain and finally transition from Proof of Work to Proof of Stake. When this happens, users will not have to upgrade ETH (there will be no ETH2 tokens) or their dApps, but they will have to update MetaMask and their nodes if they’re running any.

Around 6 months after the merge, there will be a hard-fork which will enable withdrawals of staked ETH. These withdrawals will mostly be queued/limited because a sudden drop of staked ETH poses significant security threats to the network.

Also, if you’re expecting the merge to lower transaction fees on a PoS Ethereum, I have some bad news for you: they will remain high and will likely go even higher. What will lower fees are L2 solutions, which brings us to…

2. THE SURGE

We’ve already covered sharding and all the other stuff that goes with it, so I’ll keep this short.

There are 2 measures of a blockchain’s speed: Latency (the time taken to confirm a transaction) and Bandwidth (number of transactions per second). While data sharding will certainly boost Ethereum’s bandwidth, Vitalik has said that it will probably also cause the latency to increase as well (block creation time might go from 12 secs to 16 secs).

Ultimately, transaction fees get cheaper because if the chain can handle more transactions, then there’s more space and users aren’t competing and bidding for the same slots. High bandwidth = low cost.

3. THE VERGE

The Ethereum state is currently stored using something called Merkle Trees. A Merkle Tree is a data structure in which data is hashed together in pairs until a single root is obtained. By reversing the hashing process from this root, nodes can access and verify the data stored in the tree. Proofs obtained from Merkle Trees are used to check the authenticity of data in the case of Optimistic Rollups.

Ethereum wants to eventually replace these Merkle Trees with a similar data structure called a Verkle Tree, which uses vector commitments instead of hashing and give only the pieces of state that are accessed in a particular block along with a proof that they are correct. As an idiot, I am in no position to explain the exact working of this new tree, but here’s the bottom line: Verkle Trees take up less space.

Proofs obtained from Verkle Trees are around 200kb while those obtained from Merkle Trees are over 40mb. For reference, the Ethereum state database is over 40gb.

The Verge will also see the introduction of stateless clients, which is a concept that enables nodes to validate a block without storing the entire Ethereum state.

4. THE PURGE

There are two types of legacy data stored on Ethereum: state (which is like a country’s record of all its citizens) and history (which is a country’s record of its birth and death records). The Purge will see the removal of old data in order to enable the running of even lighter nodes.

With History Expiry, clients will no longer store history over a year old. Since rollups and certain dApps do require this older data, it will be store by centralized (Etherscan, etc) and decentralized (The Portal Network, etc) protocols, from where it can be retrieved.

Ethereum also plans to eliminate ‘technical debt’, which essentially involves simplifying the protocol and removing complexities from the code.

5. THE SPLURGE

This phase will involve implementing more features like:

• Account abstraction: Will put smart contract wallets on par with regular user accounts.
• Improvements to PBS
• ZK-SNARK the EVM and everything else that can be ZK-SNARKed.

A Utopian Vision of Ethereum

If everything goes according to plan, Ethereum would have solved the blockchain trilemma in the following manner:

But what would this Ethereum of the future look like? It is already the largest blockchain ecosystem today, boasting hundreds of DeFi apps and NFT collections, and even several Metaverse platforms despite the network’s poor throughput and prohibitively high transaction fees.

While the update will (hopefully) take care of these 2 issues, Vitalik has also stated that in the future, Ethereum plans for all its transactions to take place on Layer 2s, with the mainnet providing only security and consensus. This means that smart contracts will be executed not on the EVM (Ethereum Virtual Machine), but on the VMs of Polygon, Optimism, zkSync, etc.

This would remove a wide range of current restrictions and vastly broaden the scope of innovation. Developers will be able to choose the protocol that best serves their purpose and build better products faster, leading to more value creation which will further boost the blockchain economy.

To draw parallels with the past, consider the growth of web2. For the longest time, internet users were hobbled by the painfully sluggish speeds granted by dial-up modems. In 1996, the fastest speed available to internet users was 56,000 bits per second, which is over a thousand times slower than the speeds a lot of us are accustomed to today! Despite this, the world still saw the birth of utilitarian companies like Amazon, Yahoo!, eBay and AoL. And as these companies drove up consumer demand for a faster internet, the world rushed to build better infrastructure to provide it, which eventually led to the proliferation of high-speed broadband internet in the early 2000s. This changed the game.

No longer did users have to wait minutes to just load a website with a simple graphic. Now they could stream videos on YouTube, scroll endlessly through social media feeds and do a lot of other cool shit that simply wasn’t possible before. The provision of a better internet infrastructure allowed people to build innovative products more easily and more quickly. And look where it led to: some of the biggest companies today are internet giants like Google, Amazon and Meta, which have made their founders billionaires several times over.

And I feel that something similar is going to happen with the Ethereum upgrade. But unlike the internet boom, which heavily favoured American companies because that’s where the tech and infrastructure was evolving, the boom of blockchain tech is going to be a global event, for both users and developers/builders/creators. So instead of having a handful of billionaires who monopolise the space, we might see several millionaires, or hopefully, several billionaires, who will create value across different parts of the ecosystem. We’re on the brink of the next great technology era, and I for one can’t wait to see all the stuff that’s going to be built on Ethereum over the next decade.

Additional resources:

1. For a technical but comprehensive low-level breakdown of the roadmap, read this excellent write-up by Jon Charbonneau: The Hitchhiker’s Guide to Ethereum or watch his Bankless video: Guide to the Ethereum Roadmap
2. Learn more about rollups: A non-technical introduction to rollups and An incomplete guide to rollups
3. To get a thorough understanding of the different stages of the Ethereum upgrade, watch this Bankless video: Endgame with Vitalik Buterin