Rollups, especially the optimistic rollup, are being considered by the community as the most promising short to medium term Ethereum scaling solution at the moment. The reason is that fundamentally, rollup solutions use data on-chain techniques instead of data off-chain techniques used by other scaling solutions such as Plasma. This choice removes the need for semi-trusted operators and enables generalized scalable side chains.
Currently, there are a number of teams developing rollup solutions either optimistic rollup or zk rollup, and more teams are coming. Unfortunately existing rollup solutions don’t have a viable business model YET to fund the standalone operation of this Dapp infrastructure for the long-term. This will become the critical obstacle for mainstream Dapps to consider and eventually adopt one of the solutions. At the end of the day, no product or service will take the risk to adopt an infrastructure that will go away tomorrow.
On the bright side, the permissionless and merged consensus characteristics of rollup technologies make it easy for Dapp developers to incorporate rollup technologies into their products or services and solve the scaling issue by combining the operations of both Dapps and the rollup infrastructure. It is not an ideal approach since it will add additional unnecessary technical complexities into Dapps. But Dapps have incentives to absorb the cost of operating the rollup infrastructure.
90% Dapps won’t need to worry about scaling challenge
As we all know, scaling is a hard challenge due to the Blockchain Trilemma, e.g. satisfying all three properties: security, decentralization, and scalability, is very hard. Does rollup solve the trilemma and if it does, what is the tradeoff? It kind of does and it has to scarifice “on-chain settlement time”, an implicit assumption in the Blockchain Trilemma.
If the new property: on-chain settlement time, is added into the ring, the Blockchain Trilemma will transform into the Blockchain Tetralemma as shown in Figure-1. In Tetralemma, a system can have three out of the four properties. Rollup solutions incur delays in on-chain settlement time and these delays come from waiting for transactions to aggregate, challenge period in optimistic rollup, proof generation time in zk rollup, withdrawal latency, etc.
Figure-1: Trilemma vs TetralemmaObviously, only products or services that do not require fast on-chain settlement time are better suited to build on top of rollup infrastructures. Luckily, 90% (author’s own opinion) of Dapps don’t require fast on-chain settlement time! For example, in Commerce 3.0, there will be no user experience difference if a commerce transaction is confirmed and settled at the end of the day instead of immediately. And even if a delayed on-chain settlement creates a problem, merchants can always withhold the shipment of the products or take other actions to mediate the problem generated at the underlying infrastructure layer.
The remaining 10% of Dapps such as decentralized exchanges (DEX), real-time cryptocurrency price feed oracles, that require fast on-chain settlement time will not benefit much from rollup solutions. However, if 90% of Dapps can be offloaded to rollup side chains, the Ethereum scaling situation will be in a much much better position!
Rollup solutions don’t have a viable business model for the long-term yet
Indeed, from the technical solution point of view, rollup is becoming the most promising short to medium term Ethereum scaling solution at the moment. But, from the potential adoption point of view, existing rollup solutions lack a viable business model to fund the continued operation of this Dapp scaling infrastructure for the long-term. This is a huge risk for Dapp developers since nobody would like to build products or services on top of an infrastructure that will disappear tomorrow!
A separate token is not needed and is actually detrimental to the rollup process. Hence, different from blockchain miners, operators of the rollup infrastructure cannot monetize from token rewards. It is no surprise that the majority of existing rollup projects plan to monetize from transaction fees by acting as operators of their rollup solutions, with some assuming they will be the primary block producers.
Relying on transaction fees is not a viable business model. Academic research (even though on Bitcoin with fixed block size) shows that total transaction fees depend on congestion of the blockchain and user willingness to pay higher fees to reduce transaction processing delay. As shown in Figure-2, most of the time, the Bitcoin system is not congested and the Bitcoin blocks have capacity to include additional transactions. When the system is not congested, the transaction fees per block are low and for users who are willing to bear delays, they don’t even need to pay.
Figure-2: Congestion and Transaction Fees (Source: Jacob Leshno, Economics Professor at University of Chicago)By design, rollup solutions reduce congestion and at the same time introduce delay. Hence in a rollup system, the willingness for users to pay for transactions is low and total transaction fees in the system should head towards 0. The same academic research also shows that users will still pay a low competitive transaction fee even if the Bitcoin system becomes monopolistic. Applying the same logic, in a permissionless rollup system, even acting as the primary block producer won’t force users to pay higher transaction fees.
It is possible that rollup infrastructure operators can monetize from providing users liquidity for faster withdrawals. This type of monetization does not seem to be enough either. First, very few users will need fast withdrawals since dapps that are better suited to use rollup solutions don’t require fast on-chain settlement time. Second, since anyone can provide liquidity, competition among these liquidity providers will drive down the fees for faster withdrawals.
Rollup projects as well as the community need to and will continue searching for viable business models that can provide long-term support for operating standalone rollup infrastructures. The good news is that even without viable business models, rollup technology can still be used by developers to solve the scaling challenges in their Dapps.
Dapps have incentives to absorb the cost of operating rollup infrastructure
To most Dapps, compared to the financial gain by using rollup technology to solve the scaling challenge, the cost of running rollup infrastructure is small and can be absorbed by Dapps as part of the total cost of running the Dapps. On the technical side, rollup technologies, by design, rely on Ethereum mainnet for merged consensus and security. This makes it easy for developers to incorporate rollup technologies into their products or services and combine the operations of both Dapps and the rollup side-chain.
The best way to understand how Dapps can incorporate rollup technology and combine the operations is through an example. Let’s use the decentralized commerce (Commerce 3.0) application as an example because commerce transactions don’t require fast on-chain settlement time, a great fit for rollup solutions.
In the Commerce 3.0 network, merchants, especially big merchants which sell many products, have incentives to absorb the cost and operate the optimistic rollup infrastructure to ensure their transactions are executed since otherwise, they won’t sell their products and make money. The costs and/or incentives of operating a optimistic rollup infrastructure include:
- the incentives for block producers to aggregate transactions, produce a block for the rollup side-chain, and publish required data on Ethereum mainnet.
- the incentive for validators to validate the blocks produced by block producers and generate fraud proofs if needed.
- the cost of using Ethereum mainnet that provides merged consensus and security for the rollup chain.
Accordingly, one possible solution to cover the cost and operation of optimistic rollup infrastructure is:
- Block producers: Naturally big merchants can act as block producers. Each big merchant will focus on producing blocks for its own transactions and hence don’t need incentives. For small merchants that won’t run their own block producer nodes, they can just pay big merchants to include their transactions (assuming no filtering due to competitions).
- Validators: As block producers, big merchants have incentives to act as validators as well since they want to ensure the blocks they produce will be added to the correct chain. Multiple competing merchants (block producers) can watch each other and hence maintain security.
- Ethereum gas fees: Big merchants will need to absorb the Ethereum gas fees incurred when block producers publish related data on Ethereum mainnet.
From economics perspective, big merchants may even save money by acting as block producers to process their own transactions. First, they don’t need to pay fees for standalone operators of rollup infrastructure to process their transactions. Second, the Ethereum gas fees will be the same since standalone operators will have to transfer gas fees to merchants eventually. Third, big merchants can charge fees to process transactions from small merchants. Of course, the main incentive is to have a reliable scaling infrastructure for the long-term to ensure the smooth continued operation of the Commerce 3.0 network.
Apparently, this approach won’t be ideal since it adds technical complexity into Dapps. In addition, each Dapp may choose to operate its own rollup sidechain instead of joining other Dapps’ rollup chains. The potential impact of fragmented rollup chains needs further study.
Please note that in this article, the developer incentives are not addressed even though the rollup technologies need to be developed and maintained. Top teams are funded by venture capital, and some such as Optimism, Matter Labs plan to monetize via running the rollup infrastructure while some such as Arbitrum, StarkWare plan to monetize via enterprise plans to provide technology consulting services.
Conclusion
Rollup is becoming the most promising short to medium term Ethereum scaling solution. Unfortunately there is no viable business model yet because transaction fees are likely to be insufficient to support operating standalone rollup infrastructure for the long-term. But Dapp developers should not hesitate in moving forward with their products or service due to scaling challenges. Because, they have the option to combine the operations of both Dapps and rollup infrastructure.
Acknowledgement: Thanks JosephC, John Adler, Haseeb Qureshi, Yi Sun, Daniel Lubarov for their reviews and discussions.