Upgrade to the future; a much-needed industry measure; an oasis for DAOs

Mar 18, 2024

Feature: Understanding the biggest upgrade to Ethereum since The Merge

with Noah Citron and Valeria Nikolaenko

Ethereum just underwent its biggest upgrade since The Merge last week. Called “Dencun” — a portmanteau of “Deneb” and “Cancun”, in keeping with a tradition of naming upgrades after stars and cities — Dencun bundled together nine proposed changes to the network.

Of these proposed changes aka Ethereum improvement proposals (EIPs), the most-anticipated was EIP-4844 — considered a major milestone on the road to scalability. EIP-4844 is also known as “protodanksharding” (as inspired by the names of developers Diederik Loerakker aka protolambda and Dankrad Feist).

Why it matters

So why does EIP-4844 matter? First, it introduced the concept of “blobs” — a place to store additional, temporary data on Ethereum blocks. Simply put, blobs are a new place to store rollup data that’s being added to the network. Rollups are layer two (L2) services that reduce the network’s load by processing transactions off-chain and then bringing them back on-chain. Since rollups only need this data temporarily, the blob-data is (mostly) forgotten by the blockchain afterwards.

And since blobs are ephemeral — like Instagram Stories (though in this case designed to expire after 18 days) — they lessen Ethereum’s dependence on permanent data storage. It's also a step towards enabling Ethereum to store even more data blobs via data availability sampling.

Here’s a useful analogy from a16z crypto engineer Noah Citron to help sum up why all this matters:

Think of Ethereum as a highway.
Mainnet transactions are the people riding alone in cars.
Rollups are the buses that group people, helping alleviate traffic.
EIP-4844 basically adds a “dedicated bus lane” to Ethereum, making the network more efficient.
The Dencun upgrade also paves the way for additional “bus lanes” to be added in future.

Benefits and results

Imagine if rollup data didn’t expire. That would add roughly 83.7 gigabytes of data to the blockchain every month (~31 days), and 985.5 gigabytes per year. This figure would only keep increasing, because remember: blockchains store info permanently.

By regularly expiring, blobs limit the need for that exorbitant data storage — especially as the rest could get stored off-chain via rollups. [To get a more concrete sense of data size for blobs: There’s a target of 3 blobs per Ethereum block, at max 6 blobs per block. Each blob is ~128 KB of data (a vector of 4096 elements of ~32 bytes each).]

EIP-4844 has already drastically reduced costs. A transaction on rollup provider Optimism, for instance, costs less than one tenth of a cent right now [source: l2fees.info] — about 1000x cheaper than transactions cost before the upgrade.

Note, these immediate cost savings are not likely to last. As more people stuff more transactions into rollups, fees will likely increase due to induced demand. [If you’re interested in tracking the blob fee market, check out this Dune analytics dashboard created by Citron: The dashboard shows both the current blob base fee as well as the percentage of the target base fee currently being used.]

Some suggest that the Dencun upgrade could settle anywhere between a 10x-1000x reduction in costs (note: this is purely an estimate). However, a future upgrade called PeerDAS or “full danksharding” is designed to make rollups much more efficient, increasing transaction throughput by another 32x. The key innovation there is adding more shards, and therefore more efficiency — yet without much additional cost.

Full danksharding would therefore allow many bus lanes to be added for the price of just one bus lane — leading to potentially huge throughput increases in the future.

Implications and applications

Lower transaction costs matters for everyone, because cheaper transactions unlock entirely new categories of applications that didn’t make sense with higher fees.

Because Dencun also adds the concept of transient storage (EIP-1153) to the Ethereum virtual machine (EVM), smart contracts can now store bits of data only for the duration of a transaction — rather than permanently, or only during the execution of a particular call into a contract. This means developers can do a lot more cool stuff than before, and for a lot cheaper, because they now have a sort of "medium-term" memory for smart contracts. By analogy, think of what different types of volatile memory did for semiconductor innovation…

Other benefits of the Dencun upgrade for developers include more tools for liquid staking protocols to understand what’s happening on the beaconchain (from the EVM), which helps decentralize those protocols. Still another is the mcopy opcode, which, with Dencun, now makes some smart contracts involving memory much more gas efficient.

To summarize: While the much-awaited “Merge” was one of the biggest technical feats yet — moving Ethereum from the more energy-intensive Proof of Work to Proof of Stake — we are now into the “Surge”, where ongoing updates can further scale Ethereum. This update, like all others, was in the making for a long time (Ethereum ran a trusted setup ceremony for it). But most importantly, all of these upgrades are the result of countless developers around the world coordinating and contributing via open source.

related resources:

▶️📹Distributed data storage, data availability, and danksharding by Valeria Nikolaenko — explaining approaches to distributed data-storage; different approaches for achieving consensus around data-availability; and how data availability sampling (DAS) including danksharding works

📄📝Data availability sampling and danksharding: An overview and a proposal for improvements by Valeria Nikolaenko and Dan Boneh — surveying how data availability in danksharding works; plus proposing some modifications to the underlying technique to improve data recovery

📄📝Cryptoeconomic security for data availability committees by Ertem Nusret Tas (guest talk in a16z crypto fall research seminars 2022) — on data availability committees (DAC) and their shortcomings; plus proposing a more optimal and secure DAC protocol

📄📝Atomic and fair data exchange via blockchain (2024) by Ertem Nusret Tas, István András Seres, Yinuo Zhang, Márk Melczer, Mahimna Kelkar, Joseph Bonneau, and Valeria Nikolaenko — introducing a blockchain Fair Data Exchange (FDE) protocol, enabling a storage server to transfer a data file to a client atomically; includes application of the protocol (with open source implementation) to the protodanksharding and danksharding data availability scheme on Ethereum

🎧🎙️Of data availability & danksharding with Valeria Nikolaenko, Dan Boneh, and Robert Hackett — on the bigger picture for data availability sampling; why it matters, more

Resource: Understanding ‘LVR’ – a tool for measuring the costs of market-making in DeFi

with Tim Roughgarden

A couple years ago, LVR (“lever”) introduced a much-needed framework to help DeFi liquidity providers to think about whether and when to provide liquidity to an AMM (automated market maker). And while several papers, talks, and more have since been published on LVR, there still hasn’t been an easy way to explain just what it is and why it matters.

Now, LVR co-inventor and Head of Research at a16z crypto Tim Roughgarden offers a new video explaining LVR — in just 3 minutes:
check out the video

Why does this work matter, in the big picture?

In traditional finance, participants rely on order books, brokers, centralized exchanges, and other market-making intermediaries to help them trade and hold their assets — including granting those intermediaries access to their personal information to do so.

But in decentralized finance (DeFi), math and code enable liquidity providers to pool assets (for a share of transaction fees and/or tokens), through automated market makers (AMMs) — allowing anyone to transact directly on decentralized exchanges.

However, in such decentralized finance systems, what’s the cost of providing liquidity to an AMM? The original paper introducing LVR (by Jason Milionis, Ciamac Moallemi, Tim Roughgarden, and Anthony Lee Zhang) in 2022 answered this question with a new running-cost metric they proposed, called loss-versus-rebalancing (LVR). Unlike the traditional, but widely criticized, metric of “impermanent loss”, LVR isolates the adverse selection/ information costs of liquidity provision — independent of any market risk component. The authors also suggested how LVR can inform liquidity provision decisions and the design of AMM protocols:
see the paper

▶️▶️ see also:

“LVR: Loss versus rebalancing” with Tim Roughgarden, presented at Stanford’s Science of Blockchains Conference, SBC 2022

“Automated market making and arbitrage profits in the presence of fees” by Ciamac Moallemi, presented at a16z crypto summer research series 2023

“Arbitrage extractable value” by Anthony Lee Zhang, presented at a16z crypto summer research series 2022

“Liquidity provision and automated market making” by Ciamac Moallemi, presented at a16z crypto summer research series 2022

In the news: A breakthrough for DAOs as legal entities

with Miles Jennings and David Kerr

This past week, Wyoming passed a new law to recognize DAOs (decentralized autonomous organizations) as legal entities — enabling such blockchain networks to operate within the bounds of applicable laws, and without compromising their decentralization. This is a major breakthrough as it will provide DAOs with much-needed protections and empower them to keep blockchain networks open.

Wyoming has a long history of supporting innovative legal entity structures: The state was the first to adopt the limited liability company (LLC); the first to adopt the unincorporated nonprofit association (UNA); and the first to introduce a subset of its LLC statute to be utilized by DAOs. The new Wyoming law incorporates many of the provisions we proposed in model legislation.

Since this new entity structure is likely to become the industry standard for blockchain networks created in the United States, here’s everything you need to know about Wyoming’s Decentralized Unincorporated Nonprofit Association (DUNA). It answers questions such as why DAOs need it; non-profit vs. for-profit; securities and tax law implications; and other FAQs.
read the post
share this thread on Twitter
check out this AMA on Farcaster

web3 letter from a16zcrypto