In blockchain, there exists an impossible Triangle, which is that security, decentralization, and scalability cannot be achieved simultaneously. Bitcoin and Ethereum chose the first two but are lacking in the latter. A large volume of transactions in a short period can lead to network congestion and high transaction fees.

The concept of scalability was first proposed in the Bitcoin ecosystem, with the hope of building a virtual second layer on top of Bitcoin for handling transactions, while the main chain would be used for settlement. Ethereum has tried to achieve scalability through various means, such as State Channels, Sidechains, and Plasma, but none have been entirely successful. On September 5, 2018, Barry Hat proposed the concept of Rollup on Github. Eventually, the Rollup technology gained community approval, and the Ethereum Foundation called it the only Layer 2 technology.

Five years have flown by in the blink of an eye, and Ethereum’s Rollup, which has the highest market attention, has recently been frequently reporting new developments in RaaS (Rollup as a Service). Will we soon see a Rollup summer soon? This article analyzes the overview, ecosystem, and future development of RaaS, hoping to get a glimpse of the bigger picture.

RaaS Overview

In terms of technical analysis about Rollup, its implementation is relatively complex, requiring high levels of professional skills and development capabilities. The high barrier to entry for launching a chain is contrary to the concept of blockchain’s permissionless.

Rollup-as-a-Service (RaaS) packages Rollup as a service, providing a more user-friendly and convenient experience for deploying Rollup for enterprises, organizations, and individuals, similar to Cosmos SDK and Polkadot Substrate.

Like what has been done for Layer 1 chains, RaaS provides a general SDK for Rollup, allowing autonomous Rollup development and deployment through simple configurations. The customizable feature maintains the project’s sovereignty. Some RaaS projects even offer a no-code one-click chain launch function, enabling users without programming skills to deploy their own Rollup.

Rollup is highly modular. Sequencers and Provers can be iteratively upgraded separately. In RaaS, some projects specialize in the design and development of Sequencers and Provers, offering services to all Rollups.

RaaS brings about the following changes:

1.Cheaper, more efficient, and equally secure application chains: Rollup moves expensive computational processes off-chain, making transactions cheaper and more efficient. By using the underlying public chain as the Data Availability (DA) layer and smart contracts for proof verification, it can achieve the same level of security as the underlying public chain.

2.A testing ground for innovative ideas: Rollup uses the same virtual machine environment as the underlying public chain but at a much lower cost. It can serve as a battle-test environment for the underlying chain, allowing community proposals to be thoroughly tested on Rollup before being migrated to the underlying public chain.

3.Higher interoperability: Rollups using the same set of RaaS services, due to the identical technological architecture, can easily define a set of message rules. They can transmit messages between each other without needing to bridge for cross-chain operations, achieving high interoperability between them.

RaaS Ecosystem

Broadly speaking, all projects contributing to the issuance of Rollup belong to the RaaS ecosystem. This article, based on the principle of modularity and from the bottom up, categorizes the RaaS ecosystem into four levels: DA (Data Availability), SDK (Software Development Kit), Sequencer, and No-Code (deployment).

Some projects offer services across multiple levels and will be discussed comprehensively when they first appear; they will not be repeated in the subsequent text.

2.1 DA（Data Availablity）

Theoretically, any public chain can serve as the DA layer to store Rollup’s transaction data. However, without a stable and correctly functioning DA layer, Rollup will not be able to verify the reliability of state transitions.

For Rollup, there are two choices. One is Smart Contract Rollup, which is the mode chosen by most Rollups today. It actually relies on the underlying public chain for settlement and data availability. The other is Sovereignty Rollup, which separates data availability from settlement and only relies on the underlying public chain for data availability while handling settlement itself.

The former typically chooses public chains with complete functions, such as EVM-compatible chains, Cosmos-compatible chains, or Solana. The latter’s needs have spawned projects that specialize in data availability, including Celestia, EigenLayer, and Avail.

Celestia

Celestia is a PoS chain built using the Cosmos SDK, using a modified Tendermint consensus algorithm and RS codes for block data encoding. By utilizing data availability sampling techniques, Celestia further reduces the verification cost for light nodes, which only need to download part of the block data to verify data availability.

Moreover, for checking whether a block has been correctly encoded, Celestia uses the Optimism mechanism. It optimistically assumes correct encoding initially. If no fraud proofs are received after a certain period, it confirms that the block has been correctly encoded. While this Optimism mechanism improves runtime efficiency, it does introduce some latency.

Avail

Avail is a project supported by Polygon Labs, using the BABE+GRANDPA consensus algorithm. It also uses data availability sampling technology. Unlike Celestia, Avail uses validity proofs to verify that blocks have been correctly encoded, using the more efficient KZG proofs instead of Merkle Proofs.

EigenLayer

EigenLayer is essentially a solution for heavy staking, aiming to use Ethereum’s staking liquidity to provide economic security for projects. With EigenLayer, new protocols do not need to build their own distributed verification network; they just need to leverage the security of Ethereum’s heavy staking through EigenLayer. EigenLayer performs excellently in lightweight, permissionless, decentralized scenarios, making it an ideal use case within the narrative of Ethereum scaling.

Because the DA layer doesn’t perform transaction computations but only encodes and commits to transaction data, the requirements for nodes are lower. Since it uses the PoS algorithm, staking liquidity directly reflects the blockchain’s security and availability, which is precisely the opportunity for EigenLayer to shine.

EigenLayer exists as a smart contract on Ethereum, using KZG validity proofs to check the correct encoding of blocks. However, EigenLayer has not yet adopted data availability sampling technology, which may be related to Ethereum’s next phase of upgrade plans.

2.2 Sequencer

The job of the sequencer is to sort the user transactions received, and all subsequent execution and block production will proceed in this order. In the architecture of Ethereum, because sorting and execution are handled by the same entity, it leads to excessive power for the validators. This results in phenomena like MEV (Miner Extractable Value) and censorship, which greatly impact user experience.

Separating sorting from execution is a reflection of the PBS (Propose Builder Separation) proposal’s concept of separation. However, current Rollup architectures still widely rely on centralized sequencers to determine transaction order, posing single-point-of-failure and censorship risks, and so a decentralized solution continues to be sought.

Astria

Astria offers a shared sequencer solution. User transactions from different Rollups are collected into the Astria sequencer. For Rollup nodes, data can be directly fetched from Astria to get soft confirmations with lower latency; or they can wait until Astria commits the data to the DA layer to obtain the strongest finality.

Since the data submitted by Astria contains transactions from multiple Rollups, each Rollup needs to filter out invalid transactions (including those from other Rollups) based on its consensus mechanism before processing them. Astria only provides the data, leaving the choice of consensus to Rollup nodes, thus preserving the sovereignty of the Rollup.

OP Stack

OP Stack’s default configuration uses a single dedicated sequencer for transaction sorting. A simple modification is to use a permissioned set of sequencers, which could mitigate the likelihood of malicious activity through a PoS mechanism.

After introducing the concept of Superchain in OP Stack, a shared sequencer becomes an inevitable choice, bringing atomic cross-chain functionality and improving interoperability between Superchains.

Espresso

Espresso aims to utilize the liquidity of Ethereum stakers and gain shared security through heavy staking. Espresso integrates the sequencer and DA, providing sorting results for Rollups through a REST API, abstracting away the DA details. The security of consensus is verified by smart contracts located on L1, providing stronger reliability.

Saga

Saga initially served a role similar to Cosmos Hub, providing shared security for application chains on Saga that use the Cosmos SDK.

With the surge in popularity of Rollups, Saga collaborates with Celestia, utilizing Celestia as the DA while converting its validators into sequencers. It exchanges information with upper-layer Rollups through Optimistic Rollup IBC, offering shared security.

SUAVE

Unlike other sequencers, SUAVE is specifically targeting the MEV market. Flashbots are the absolute leaders in the MEV space, and SUAVE is their product aimed at capturing cross-chain MEV, claiming ‘The Future of MEV is SUAVE.’ With the shared sequencer provided by SUAVE, atomic cross-chain transactions become possible, contributing to the efficiency of different chain capital markets.

EigenLayer

We’ve previously discussed EigenLayer’s use case on the DA level. Decentralizing sequencers is also EigenLayer’s forte.

Since sequencers are only responsible for sorting and not for execution, the node requirements are very low. The key to decentralization lies in lowering the potential for malicious activity through penalty mechanisms. EigenLayer can provide deep staking pools, using Ethereum’s decentralization to cultivate the decentralization of Rollup sequencers.

2.3 SDK（Software Development Kit）

Similar to Cosmos SDK, the SDK provided by RaaS allows developers to reuse a large number of software modules and customize the Rollup they need at the lowest cost, reducing the difficulty of development.

Rollkit（Optimism）

Rollkit was initially incubated by the Celestia community and has now become an independent project. Rollkit uses Celestia as the DA layer, providing an ABCI-compatible client interface, serving all ABCI-compatible Rollups (Cosmos chains).

At this stage, Rollkit uses a single centralized sequencer and supports the integration of Cosmos SDK, Ethermint, and CosmWasm. Users can choose the execution environment they want. In the future, Rollkit will continue to develop to support more configuration services.

Dymension（Optimisim）

Dymension divides services into front-end and back-end. The front-end supports custom RollApps, supported by Dymension RDK (modified Cosmos SDK). The back-end Dymension Hub coordinates the entire system, handling DA and sequencing.

Dymension uses the Optimism mechanism. The Dymension Hub optimistically accepts status updates from the sequencer. If a valid fraud proof is received, it will rollback the state change. RollApps can achieve an average latency of 0.2 seconds and a maximum of 20,000 TPS.

Dymension adopts a flexible block generation scheme, when there are no transactions in the block, it will stop generating blocks, which significantly reduces the operating costs of the sequencer.

At this stage, the Dymension product is still under development, and has already supported the execution layer of EVM, and there is no clear choice for the DA layer.

Sovereign（ZK）

The Sovereign SDK offers zk-Rollup as a Service, providing general modules for building blockchains and a zkVM that hides the underlying details of zero-knowledge proofs, allowing developers to program in Rust. The SDK can compile it into an efficient zk-friendly form.

As the project name indicates, Sovereign SDK emphasizes sovereignty. Rollup determines the legality of state transitions through custom consensus rules, without requiring verification from the DA layer.

Currently, Sovereign SDK has already adapted to the DA layer of Celestia and Avail, supports Risc0’s zkVM, and can achieve the deployment and demonstration of Rollup.

Stackr（Unknown）

Stackr proposes a more radical innovation, wanting to transfer the microservice architecture in the traditional internet to the blockchain, introducing the concept of micro-rollup.

The relationship between regular rollup and micro-rollup is like that between virtual machines and containers. Using the Stackr SDK, developers only need to define the required data structure and state transition functions; the rest is handled by Stackr.

Stackr supports multiple execution environments, such as EVM, Solana VM, FuelVM, etc. Users can choose the environment they wish to use.

AltLayer（Optimism）

AltLayer, as a decentralized and flexible RaaS, offers an SDK aimed at developers and a No-Code Dashboard that doesn’t require coding experience, capable of one-click chain launches.

AltLayer provides a unique flexible Rollup, called Flash Layer. When application demand surges, a Rollup chain can be quickly deployed. When demand returns to normal, settlement can be done on L1, abandoning that Rollup, realizing the common horizontal scalability in the internet system.

AltLayer aims to support multi-chains and multiple execution environments and has already realized support for EVM and WASM.

OP Stack（Optimism）

OP Stack is built to support the Optimism Superchain, a proposed network sharing security, communication layer, and joint development stack from the L2 network. After the Bedrock upgrade, Rollups created using OP Stack will be natively compatible with Superchain. Of course, components of OP Stack can be modified to acquire customized features; base and opBNB are both Rollups developed based on OP Stack.

The security and availability of OP Stack have been fully tested by the OP Mainnet, Base, and other chains. However, there are still current issues like the lack of fraud proofs, centralization in the sequencer, etc. OP Stack is exploring new directions, such as adopting a cheaper DA layer, using ZK Proof, sharing sequencers, etc.

Arbitrum Orbit（Optimism）

On June 22, Offchain Labs released tools for issuing the Arbitrum Orbit Chain. Orbit Chain is a Layer3 built on top of Arbitrum Layer2, and can choose to settle on one of the three Layer2s, namely Arbitrum One, Arbitrum Nova, Arbitrum Goerli. Users can choose to use Rollup or Anytrust technology; the difference is that Anytrust uses DAC without submitting transaction data to the chain, at a lower cost but with weaker security. Orbit Chain has the advantages of a simple chain launch process, interoperability with the Arbitrum ecosystem, immediate updates from Nitro, and EVM+ compatibility provided by Stylus (supports Rust, C, C++, running on the WASM virtual machine). Users can customize and issue any Orbit Chain, but must settle on Arbitrum Layer2; otherwise, they need to contact Offchain Labs or Arbitrum DAO for authorization.

ZK Stack（ZK）

On June 26, zkSync published an article, announcing that in the coming weeks they will modify the existing open-source code and launch ZK Stack, allowing users to build their own customized ZK super chains. Unlike Arbitrum’s Orbit Chain, ZK Stack emphasizes sovereignty and interoperability; users can completely customize according to their needs. Chains built using ZK Stack can achieve bridgeless interoperability. ZK Stack can be used to build both Layer2 and Layer3; the officials have not restricted it and have not required it to settle on zkSync. From this perspective, the sovereignty provided by ZK Stack seems stronger.

Starknet Stack/Madara（ZK）

Madara’s original positioning was as a sorter on Starknet, leveraging technological accumulation, successfully developed Starknet Stack based on the original product, helping build application chains Rollup on Starknet. Using Ethereum as the DA layer, and using Starknet’s shared prover for settlement on Starknet. From the usability point of view, Madara has already helped teams issue application chain Rollups within 24 hours at the PragmaOracle hackathon, and also provided a video demonstration; compared to zkSync’s ZK Stack, the completion level is higher.

2.4 No-Code（No-Code Deployment）

Codeless deployment is a lower-threshold solution that offers a one-click chain launch option for non-developers, promising to further increase adoption.

Caldera（Optimism）

Caldera Chain is a fully customizable one-click chain launch solution. On the execution layer, it supports OP Stack and Arbitrum Orbit, and the settlement layer can choose between Polygon, BSC, Evmos, and other EVM-compatible chains. The DA layer is supported by EigenLayer and Celestia.

In addition to the Rollup chain itself, Caldera also provides a range of supporting infrastructure, such as blockchain explorers, testnet faucets, oracles, and Hyperlane-supported bridges, further reducing the cost of launching a chain.

Eclipse（Optimism+ZK）

Eclipse has a high level of customization, supporting EVM and SolanaVM on the execution layer, and integrating Celestia, Avail, and EigenLayer on the DA layer. The settlement layer provides Optimistic settlement and is developing ZK settlement with RISC0 zkVM.

Users can also choose the chain’s access (permissioned/unpermissioned), Gas tokens, whether to charge a Gas fee, whether to allow MEV, specific Opcodes, block size, etc., giving it a high degree of flexibility.

Opside（ZK）

The most noticeable feature of Opside is that it has built a decentralized ZKP market. Originally, I wanted to describe the Prover separately as a layer, but gave up due to the small number of projects. Zero-Knowledge Proofs (ZKP) have high computational requirements. As the market share of zkRollup gradually increases, the decentralization of ZKP is the major direction for future development.

Opside adopts a permissionless PoW consensus mechanism to attract miners to generate ZKPs, maintaining the security and availability of zkRollup, without the need for the chain initiators to consider the problem of proof generation. At the verifier level, it adopts a PoS mechanism to lower the entry barrier and promote the centralization of verifiers.

Opside offers customized services; users can choose among zkSync, Starknet, Polygon zkEVM, and other zkEVMs, and can also modify the economic model and adjust gas fees.

The Future Outlook of RaaS

More ZK

Compared to Optimistic Rollup, zkRollup upgrades the security from economic to cryptographic, making it more secure; eliminates the need for a long waiting period for challengers, thus reducing confirmation delay; and compresses data more efficiently, making DA cheaper.

Although the Optimism solution has a high market share due to its technological maturity and early advantage in product release, ZK as a revolutionary technology will play an increasingly important role in the future. In his Montenegro speech, Vitalik placed ZK technology on an equally important footing with blockchain technology, which also indirectly reflects the importance of ZK.

As the technology continues to improve, more zk-Rollup as a Service projects will come into public view, providing users with more choices.

More Non-Ethereum

To this day, the Ethereum ecosystem still occupies an absolutely dominant position in the entire blockchain industry. Although other communities are constantly iterating and innovating, they still can’t shake Ethereum’s throne.

In RaaS, things seem to have changed a bit. Due to Ethereum’s low data storage capacity and high prices, people can choose cheaper DA layers like Celestia, Avail, or Polygon; Ethereum is non-modular and difficult to modify, people can opt for the highly modular Cosmos SDK; the EVM is inefficient, people can choose more efficient Solana VM, Move VM, CairoVM.

‘A single flower does not make spring, a hundred flowers in full bloom make spring everywhere.’ Various non-Ethereum ecosystem solutions will blossom with new vitality in RaaS.

More Modularization

The role of modularization can be divided into two points: one is that each module can iterate quickly, improving development efficiency; the second is that modularization can greatly reduce the complexity of customization.

In the current market environment, it’s almost impossible to independently develop a one-stop solution; the overall innovation speed can never catch up with the rapid iteration of small modules. Extreme customization requirements will lead to further refinement of module division; if one doesn’t modularize themselves, they will eventually be broken down by other projects, like OP Stack and Arbitrum Orbit being separated into execution layers by Caldera.

More Customization

As scalability technology matures, transaction fees are getting lower, and infrastructure is improving, people realized, ‘Fool, the key is in the application.’ Every application has its specific operating rules and models; a single solution cannot adapt to a complex application ecosystem, so we need more customization.

From block size to data structure, from transaction fees to transaction latency, from admission mechanisms to security assumptions, from contract engines to token empowerment, the degree of customization for future Rollup will gradually upgrade, providing more flexible solutions for applications.

More Interoperability

As mentioned earlier, the Ethereum ecosystem’s dominant position in the blockchain ecosystem is largely related to its locked huge liquidity. In the crypto market, because each chain exists independently, liquidity cannot exist simultaneously in two chains; the increase in the number of Rollups will lead to further fragmentation of liquidity, which becomes a serious problem.

Stronger interoperability can reduce cross-chain friction and allow liquidity to flow smoothly between different chains, even sharing liquidity. Consistent with the ideas promoted by Cosmos, OP Stack, Arbitrun Orbit, ZK Stack, Starknet Stack, etc., are all trying to build a large application chain ecosystem. Rollups built using the same technology stack, due to their same technical architecture, can obtain native interoperability without needing to build cross-chain bridges.

More Overcollateralization

Many services in RaaS currently use the PoS model, using economic penalties to increase the cost of malicious activities to improve security. But economic security requires deep collateral assets as a guarantee, leading to low capital utilization and increasing the startup cost for service providers.

Overcollateralization is a good solution, using Ethereum’s large staking pool to provide shared security for other services while increasing revenue for stakers and improving capital utilization. EigenLayer and Espresso are currently doing related work, and it can be expected that more services will be guaranteed by overcollateralization for economic security in the future.

In summary, The real beneficiary of RaaS development is the application chain. Can the application chain concept proposed by Cosmos and Polkadot in the early years sprout, and explode in the RaaS ecosystem? We shall also wait and see.

Perhaps, only innovation at the application layer can drive the explosive growth of the RaaS ecosystem. After all, even the best road is not considered a good infrastructure if there are no cars running on it.

Disclaimer：

1. This article is reprinted from [Marsbit]. All copyrights belong to the original author [CynicLeo]. If there are objections to this reprint, please contact the Gate Learn team, and they will handle it promptly.
2. Liability Disclaimer: The views and opinions expressed in this article are solely those of the author and do not constitute any investment advice.
3. Translations of the article into other languages are done by the Gate Learn team. Unless mentioned, copying, distributing, or plagiarizing the translated articles is prohibited.