Ouroboros Leios, sharding, and layered architecture are three distinct scalability solutions with different implications for L1 fees. In Ouroboros Leios and sharding solutions, fees are used to cover the costs of blockchain operation. However, L2 networks operate somewhat parasitically on L1, with L2 network operators profiting while the parent blockchain may suffer economically. Fees paid on L1 are crucial for the security budget, which will become increasingly important over time. Therefore, scalability must be designed with the long-term economic sustainability of the blockchain in mind. That’s a challenge.
Evolution of Scalability Solutions
The most established scalability solution is the Lightning Network, Bitcoin’s L2. However, it hasn’t gained as much user traction as the numerous L2 networks in the Ethereum ecosystem, such as Arbitrum, Base, and Optimism. There are several dozen L2s within Ethereum’s ecosystem. A channel-based second layer similar to the Lightning Network is being built in the Cardano ecosystem. It’s Hydra.
Initially, the Ethereum team aimed to implement full-fledged sharding but has since shifted towards an L2-centric roadmap. They upgraded the network to introduce BLOBs, special transactions that enable cheaper L2 interactions with L1, thereby reducing L2 fees. However, this upgrade has negatively impacted Ethereum’s monetary policy, increasing the number of ETH in circulation and challenging the narrative of ultra-sound money.
Sharding is an already implemented but not yet time-tested technology. Ouroboros Leios is still in development. These new scalability solutions have a significant advantage over L2s: the fees remain as income for the blockchain itself, rather than going to an external network with its own infrastructure and monetary policy.
While L2s effectively address scalability for blockchains, they can be seen as competitors in terms of the security budget, as third parties primarily benefit from the user fees.
Users incur L1 fees when moving assets between L1 and L2. L2s rely on L1 for ledger data storage, leveraging its security and decentralization. Despite this, L1 fees remain much lower than if all transactions occurred directly on L1.
Economic Impact of L2s on L1
The goal of L2 solutions is to enable users to conduct numerous transactions quickly and affordably. L2 transaction fees typically range from 0.001 to 0.2. These fees are expected to decrease further in the future. For instance, Base by Coinbase processes over 3 million transactions daily at an average fee of $0.01, while Arbitrum handles around 2 million transactions per day with a similar fee. Although fees can be volatile, the daily revenue for these L2s varies from tens to hundreds of thousands of USD. During peak load times, their revenue has reached approximately one million USD per day.
The existence of the L2 ecosystem generates a load on L1 as users move assets between networks, resulting in transaction fees for L1. One might wonder whether the total collected fees would be higher if L1 were congested, thereby increasing L1 fees due to higher demand, or if asset transfers between networks could generate equivalent revenue. However, the goal is not to have a congested L1, making this debate moot.
Additional revenue for L1 comes from transactions initiated by L2s that write data to the blockchain, as these also incur fees. In the case of Ethereum, revenues can be low, as fees for BLOB transactions are negligible. Bitcoin does not have anything similar, so the revenue for on-chain transactions associated with the Lighting Network is higher.
In general, the presence of L2s generates L1 transactions, thus bringing in revenue and improving user experience. A perpetually congested L1 would frustrate users, leading them to migrate elsewhere.
L1 loses fees for transactions that occur on L2s with already transferred assets. For example, if Alice and Bob each transfer assets from L1 to L2 once, and then each makes 10 transactions per day on L2 for the next month, L1 will only receive fees for 2 transactions. In contrast, the L2 network will earn fees for 600 transactions. If the L1 fee is $1 and the L2 fee is $0.01, L1 would earn $2 while L2 would earn $6.
If Alice and Bob move their assets back to L1, L1 earns an additional $2. However, if users remain on L2 for an extended period, such as a year, without needing to make any L1 transactions, L1 could face economic challenges. In this scenario, L1 would primarily function as an onboarding platform for L2, while L2 would be the main beneficiary economically.
It is difficult to predict future developments in L2s, people’s behavior, and their needs to use L1 transactions in a fully developed L2 ecosystem.
Many are already highlighting that L2 solutions act as parasites on L1, depriving it of fees. Examining the data for Ethereum supports this view. Ethereum has become inflationary again, with more ETH being issued than burned. This trend reversed following the Dencun upgrade, which introduced BLOB transactions.
L2 networks have their infrastructure, including nodes or servers. The costs for these resources must be covered, just as users pay to maintain L1. This creates an inevitable competition for fees between L1 and L2 networks. As L2s become more decentralized, secure, and reliable, users will be less inclined to return to the slower and more expensive L1.
At some point, L2s could become significantly more profitable than L1s. Although L2s depend on L1 for their existence, the question arises whether a mechanism will be introduced to transfer a portion of L2 revenue to L1. As L2s become more decentralized and handle a large volume of transactions and smart contracts, their operational costs will increase. Both L1 and L2 full-node operators will seek profit. Additionally, competition among L2s will drive fees down, reducing revenue.
Although this article primarily discusses fees, there is another crucial dimension to consider. If the demand for L1 transactions is high from both L2 networks and L1 users simultaneously, L1 will remain a bottleneck. This means scalability issues will persist. If the network employs a fee market, L1 fees will rise, and a congested L1 will negatively impact the performance and fees of all L2s.
We have already seen this scenario with Bitcoin. When Bitcoin became congested due to Ordinals and fees soared to tens of dollars, users were reluctant to open and close Lightning Network (LN) channels. On-chain fees often exceeded the amounts users wanted to transfer via LN channels, making LN usage economically unviable at that time. If Bitcoin fees were permanently high, using LN for many use cases would not make sense without further innovations. For instance, transferring $50 to LN would be impractical if the on-chain fee were typically $30.
When L1 fees rise, it’s inevitable that L2 fees will also increase. Businesses developing L2 solutions need to account for this and adjust their fees accordingly. This dependency on L1 fees is far from ideal.
Another issue arises when multiple L2s are introduced, leading to fragmentation. Users may be compelled to operate across various networks, spreading their capital. Different users might favor different L2s, making interactions less secure and potentially more costly. Additionally, bridging assets between these networks could incur L1 fees.
Ideally, there would be a unified network or ecosystem where revenue is fairly distributed. Sharding and Ouroboros Leios’ solutions seem less problematic in this regard, as fees are collected by a single protocol that manages the reward mechanism. Even from the point of view of scalability, it also appears to be a more suitable solution, as the blockchain does not remain a bottleneck in the ecosystem. However, this does not mean that newer solutions will be free of bottlenecks.
L1 Scalability as a Path to Economic Sustainability
Sharding and Ouroboros Leios can be described as L1 scalability solutions. They have similar characteristics.
L1 scalability offers the benefit of keeping fees within a single network, as it can handle all transactions. This eliminates the need to share fees with another network. All active full nodes contribute to scalability and are rewarded directly by the protocol.
From the user’s perspective, this is simpler because they consistently use the same wallet and addresses, avoiding capital fragmentation. Since all users are on the same network, direct interaction is always possible.
The downside of L1 scalability is that transactions are processed by a distributed network, which is inherently more expensive compared to the often centralized L2s. L2s can achieve fast settlement times and very high throughput, which can be challenging for L1 scalability.
A more centralized network will always be more efficient than a decentralized one.
Users new to blockchain solutions will likely seek the lowest possible fees, balancing between cost, security, reliability, and speed, even if it means sacrificing decentralization. This poses a challenge for L1.
As mentioned earlier, L2 fees will always depend on L1 fees. To keep L2 fees low, L1 fees must also remain low. Therefore, it’s crucial for L1 to scale effectively, even though most transactions will occur on L2s.
L1 can generate the same revenue with low TPS and high fees or high TPS and low fees.
For example, if the average fee is $1 and the network handles 7 TPS, it can earn about $25,000 per hour. Let’s ignore the fee market. If a network handles 700 TPS with a fee of $0.001, it will earn the same amount per hour.
However, in terms of scalability and user comfort, only the second option is viable.
Users will typically select a business solution based on cost. In the context of blockchain, this means they prioritize scalability while also considering decentralization and security.
L1 scalability can secure the long-term economic sustainability of blockchain. Users are generally more inclined to pay lower fees for similar services. If blockchain networks consistently charge fees in the tens of dollars, users may eventually abandon them.
The advantage of L1 scaling is that it offers a more decentralized network compared to L2s. The true value of blockchain lies in its decentralization. The discussion around fees will likely focus on the trade-off between fee levels and the degree of decentralization. From my perspective, moving assets from decentralized blockchains to centralized, often custodial services, doesn’t make sense. Centralization tends to lead to abuse of power. Many L2s still don’t fully adhere to decentralization principles and yet continue to exploit the blockchain industry.
Ouroboros Leios and sharding offer more decentralized solutions compared to L2s. Revenue from fees serves as a crucial part of the security budget. Ensuring that users remain on the blockchain and are willing to pay for decentralization is essential for long-term economic sustainability, and achieving this requires effective scalability.
Bitcoin requires special attention due to the high operational costs of PoW. Maintaining security will be challenging if fees don’t compensate for the decreasing block reward, which halves every four years. Currently, there isn’t a high demand for Bitcoin transactions. This might be because when demand spikes, fees become prohibitively high, deterring users. Consequently, demand drops and average fees fall below $1. As of now, the fee is $0.3. For one block, a miner receives an average of 800 USD from fees, while the block reward is roughly 200,000 USD.
Conclusion
Predicting the future of scalability solutions and the level of decentralization users will demand is challenging. Estimating when and what economic sustainability issues L1 might face is even harder. Efficiency will be crucial for success. A blockchain that provides the best decentralization and security at the lowest cost will have a competitive edge. However, it’s not guaranteed that users will prioritize decentralization.
Ethereum users, for instance, are comfortable using L2s, with the top 5 L2s processing around 200 TPS collectively. The Lightning Network hasn’t gained significant traction yet, with Bitcoin handling about twice as many transactions (4.4 TPS) as LN (2.5 TPS for non-private transactions). Most LN users opt for custodial solutions.
I’m a strong advocate for L1 scalability. However, even L1 scalability will need to rely on L2s in the short term if there’s a significant increase in adoption. Even if most cryptocurrency users end up primarily using L2s, we must consider the importance of onboarding through L1. L1 needs to be both affordable and efficient. Improved L1 scalability is crucial since all L2s rely on it.