A deep dive into the world of Layer-2s

In 50 years, we might all be telling our grandchildren about the good old days when we could interact on Layer 1 for less than $100. It’ll be as credible to them as if you were to say that Tchaikovsky was inspired to compose his ballet after watching Barbie Swan Lake. (Obviously, he did not). 

Why? Already at the present stage, we’ve seen Ethereum fees spike whenever a new hyped mint occurs or a memecoin launches that everyone needs to ape into for a slight chance at generational wealth. A few days ahead of the Dencun upgrade, fees would range from $150 to $200 for a swap of assets worth $1k. 

While increased demand and activity is good overall, it’s not feasible to scale this to a global computer - as Ethereum’s ambition has always been. Fortunately, all is not lost in the scaling race, as L2s have matured to become multi-million dollar layers home to vibrant communities. 

What are L2s to begin with? 

L2s are one answer to solving the blockchain trilemma. The trilemma states that you can only achieve two of the three: decentralization, security, and scalability. Currently, Ethereum is relatively decentralized and secure but lacking in scalability. It cannot support millions of users transacting simultaneously except if we add L2s to the equation. Layer-2s provide a cheaper and faster way to execute ethereum transactions.

The most common form of L2s these days are rollups, which bundle transactions and send them as a single piece of data to the mainchain, consequently reducing the computing power required. Rollups are like having an assistant who sorts and prepares all your paperwork to the degree where you (the L1) just have to sign the final version - instead of doing all the work of filling them in yourself. 

Among rollups, we have optimistic and zk rollups. 

• Optimistic rollups: assume that transactions are valid by default. They have a challenge period during which transactions can be proven “dishonest”.  
• Zk rollups: use zk-proofs to confirm to Layer-1 that all the computation on the rollup has been in accordance with consensus rules. 

As with any trending narrative in crypto, we’ve witnessed countless new Layer-2s launch in the past few months, leading to the question.

Why is everyone launching L2s? 

During our last Twitter Space, we asked our speakers working at different L2s precisely that question. They identified a few reasons for that, including: 

• The ease of doing so has drastically increased thanks to a rise in modularity. This is further enhanced by the availability of chain-building kits such as OP Stack or Polygon CDK. 
• No one layer can solve all the problems. This means that similar to how different regions of the world rely on different payment providers, L2s might start localizing and optimizing geographically. 
• In addition to localization, our speakers foresee that L2s will focus on different areas such as privacy, DeFi, compliance, and gaming solving for use cases with Layers instead of being general-purpose. 
• There is a market share to be captured. It’s a free market, and more L2s mean more options for users, which isn’t a bad thing. 
• Running Layer-2s opens up an attractive design space as you decouple execution from consensus. Take Eclipse, a Layer-2 that runs Solana’s Virtual Machine on top of Ethereum. 

And, of course, for some time, it was a relatively easy thing to raise funds for and get degens to ape in, especially if you were backed by Paradigm. 

Needless to say, just having points isn’t a sufficient differentiator. Here is how the Layer-2s in our space thought about their unique selling point. 

Examples of what different L2s specialize in

• Manta Network started with a focus on privacy and then pivoted to focusing on DeFi, especially LSDs and LSTs. Overall, their efforts aim to bring new users to the ecosystem and reward the users already there. 
• Kroma shared that they are the first OP rollup with validity proofs. They also have a vast open-source zk library and are further differentiating themselves by a strong local presence in the Asian market and expertise in gaming. 
• Etherlink: is the first enshrined L2 on Tezos, meaning it has decentralized sequencing from the start and resistance to censorship. 
• NEON: highlights that what really sets them apart is leveraging Solana’s parallelization - removing the usual bottleneck of EVM: sequential execution. 

There are many more L2s, and you can learn more about them in detail on L2Beat. 

One thing that stands out when you do that is that, for many, in case of sequencer and proposer failure, it becomes hard for users to withdraw funds, especially for those not familiar with how to force a transaction via an L1. 

“So much TVL is in rollups with centralized sequencers when we are building a censorship-resistant future of finance. Realistically, is any user going to use an escape-hatch or self-propose?”  - Sasha, Etherlink

How can that be in a supposedly decentralized, permissionless ecosystem? 

The answer is simple: centralized sequencers. 

On decentralizing sequencers 

Sequencers play an important role for L2s as the entity that handles, receives, organizes, and executes transactions before submitting them to be included in blocks. However, centralized sequencers are optional. In theory, L2s can use the underlying L1 for sequencing. Examples of that are NEON EVM, which relies on Solana for execution, and Etherlink, where Tezos mainnet takes on the role of ordering and executing transactions. 

“Neon has no sequencer, so we don’t have that problem. The currently centralized component for us is a proxy operator that executes Ethereum transactions on Solana. We’ll work on making that permissionless eventually”  - Simeon from NEON

Doing so on Ethereum would reduce the cost savings and hamper user experience, which is why most L2 teams have chosen centralized sequencers, introducing a single point of failure. Still, teams are acutely aware of the paradox that is having a decentralized base layer and then on top tech that does not offer the same guarantees. 

Decentralizing sequencers is eventually on everyone’s roadmap, and some projects, like Espresso or Astria, are working on solutions such as shared sequencing, which allows rollups to share a sequencing network. Benefits of decentralized sequencers include: 

• Immutability 
• Censorship resistance 
• Resilience 
• Reduction of intermediaries. 

Or put otherwise, simply what Web3 was meant to provide. It might be some time until we see any of the decentralized sequencing networks go live in production. Fortunately, there are many other exciting things happening in the L2 space. 

What does the future hold? 

With the Dencun upgrade, Ethereum has made it cheaper for rollups to store their data. This should decrease gas fees on L2s, improving the user experience. And still, just a few days later, gas fees on Base hovered around $10, driven by an increase in interest. 

Clearly, blobs aren’t the solution yet, which explains why networks like Manta chose to use Celestia for DA. Future upgrades should further improve the cost of using Ethereum for DA, but it’s a long way off. 

In the meantime, there are other trends to look at. 

“There is quite a lot of buzz around the BRC ecosystem and hyper around Bitcoin L2s. I’d also keep an eye on Merlin.” - Shubam, Manta Network 

Sasha from Etherlink thinks that Enshrinement might be the next big thing. It’s a “good buzzword because no one really knows what it means.” 

In regards to the tech and decentralizing, Steve from Kroma points out that currently, most ZK rollups rely on a centralized prover. He foresees the creation of prover networks that’ll be shared across zk rollups. 

Lastly, if you are trying to run an L2, here are some recommendations from people who are doing it: 

• Focus on the problem you are solving. Don’t build just because it’s a trendy thing. 
• Ethereum DA is expensive  - might not make sense if you have very little traffic
• Remember decentralization
• Embrace the journey. 

For more insights, you can listen to a recording of our recent Twitter space here.