a16z: Dencun, The Biggest Upgrade to Ethereum Since The Merge

a16z: Dencun, The Biggest Upgrade to Ethereum Since The Merge

a16z: Dencun, The Biggest Upgrade to Ethereum Since The Merge

Last week, Ethereum underwent its largest upgrade since The Merge. “Dencun” is a combination of “Deneb” and “Cancun,” following the tradition of naming upgrades after stars and cities. “Dencun” bundles together nine proposed network changes.

One of the most anticipated changes among these proposals, known as Ethereum Improvement Proposals (EIPs), is EIP-4844. It is considered an important milestone on the scalability roadmap and is also referred to as “protodanksharding,” inspired by developer Diederik Loerakker, also known as protoolambda, and Dankrad Feist.

So, why is EIP-4844 important? Firstly, it introduces the concept of “blobs” – a place to store additional temporary data on Ethereum blocks. In simple terms, a blob is a new location for storing rollup data added to the network. Rollup is a Layer 2 (L2) service that reduces network load by processing transactions off-chain and then bringing them back onto the chain. Since rollup only temporarily needs this data, the blob data (mostly) will be forgotten by the blockchain.

Moreover, because blobs are temporary – like Instagram Stories (although designed to expire after 18 days in this case) – they reduce Ethereum’s reliance on permanent data storage. This is also a step towards storing more blob data by implementing data availability sampling.

Here’s a helpful analogy from a16z crypto engineer Noah Citron to summarize why all this is important: Imagine Ethereum as a highway. Mainnet transactions are people driving alone. Rollup is like a bus that gathers people together and helps alleviate traffic congestion. EIP-4844 essentially adds a “bus-only lane” to Ethereum, making the network more efficient. The Dencun upgrade also paves the way for adding more “bus lanes” in the future.

The benefits and results are significant. Imagine if rollup data never expired. This would add approximately 83.7 GB of data to the blockchain per month (about 31 days) and 985.5 GB per year. And the number would only continue to increase because, remember, the blockchain permanently stores information. By regularly expiring, blobs limit the demand for excessive data storage, especially when other data can be stored off-chain through rollup. (To get a more concrete understanding of blob data size: Each Ethereum block aims for 3 blobs, with a maximum of 6 blobs per block. Each blob is ~128 KB of data, a vector containing 4096 elements, with each element being approximately 32 bytes.)

EIP-4844 has significantly reduced costs. For example, a transaction on rollup provider Optimism now costs less than 0.1 cents (source: l2fees.info) – about 1000 times cheaper than before the upgrade. It’s important to note that these immediate cost savings are unlikely to be sustained as more people put more transactions into rollup, which could increase fees due to induced demand. (If you’re interested in tracking the cost market for blobs, check out the Dune dashboard created by Citron: the dashboard shows the current base cost of blobs and the current target base cost percentage in use.)

Some believe that the Dencun upgrade could reduce costs by 10-1000 times (this is purely an estimate). However, a future upgrade called PeerDAS or “full danksharding” aims to make rollup even more efficient, increasing transaction throughput by 32 times. The key innovation is to add more shards to improve efficiency without incurring too much additional cost. Therefore, full sharding technology will allow for adding many more “bus lanes” at the price of one, potentially leading to significant growth in throughput in the future.

Lower transaction costs are important for everyone, as cheaper transactions unlock new application categories that wouldn’t make sense at higher fees.

With Dencun, the concept of transient storage (EIP-1153) is also added to the Ethereum Virtual Machine (EVM). Smart contracts can now store data only during transactions, rather than permanent storage or only during specific contract invocations. This means developers can do cooler things at much lower costs because they now have a “mid-term” memory for smart contracts. Think about the impact of different types of volatile memory on semiconductor innovation…

The Dencun upgrade also brings other benefits for developers, including more liquidity staking protocol tools to understand what’s happening on the beacon chain (from the EVM), which helps decentralize these protocols. Another one is the mcopy opcode, which, together with Dencun, makes some memory-involved smart contracts more energy-efficient.

In summary, while the long-awaited “Merge” is one of the biggest technical feats so far – transitioning Ethereum from energy-consuming proof-of-work to proof-of-stake – we are now entering the “Surge,” where ongoing updates can further scale Ethereum. Like all other updates, this one has been in the works for a long time (Ethereum held a trusted setup ceremony for it).

But most importantly, all these upgrades are the result of countless developers coordinating and contributing through open-source efforts.