Layer 3s, in the context of exponential adoption, are seen as a life-saving solution to soaring fees, even on level 2 blockchains. Unlike their underlying counterparts, layer 3s are mostly for specific use and therefore derive their scalability from application fractalization. Let’s find out how they work in detail and examples of these ecosystems.
What is Layer 3?
Victim of their adoption, the scalability solutions of the Ethereum blockchain, whose goal is to decongest the latter, must face the very problem that they must solve. To compensate for this surge in costs, the developers decided to look into creating level 3 networks—the latter benefit from the security of the parent blockchain by recursion.
Unlike layers 1 and 2, the 3rd layer is called “specific use”. In other words, each layer 3 blockchain is built with a use in mind, for which it will be optimized. This separation of uses allows layer 3 to be very efficient in addition to benefiting from fractal scalability.
At the time of writing these lines, the development of layer 3 solutions is still at an infant stage. For the moment, only Avalanche and StarkWare are seriously considering this option in order to improve the experience of their users.
Building a Layered Smart Contract Blockchain
The Blockchain Trilemma
When designing a blockchain, developers have to deal with the famous blockchain trilemma. This can be represented in the form of a triangle in which each vertex corresponds to a crucial parameter of a blockchain: security, scalability, decentralization.
Figure 1: The blockchain trilemma
This trilemma is notorious because it has never been solved perfectly. It is very common to have a blockchain that ticks two of the boxes while bringing together the 3 characteristics that remain a mirage for the moment. For example, Ethereum (ETH) ticks the boxes of security and decentralization while Ripple (XRP) ticks those of security and scalability, completely neglecting decentralization.
Figure 2: The different layers of blockchain technology
A possible solution to this trilemma is the establishment of layers on top of a parent blockchain. This way, the parent blockchain would only have to deal with security and decentralization while the upper layers could focus on scalability.
The Layers 1
The idea of a layered blockchain is not new. Indeed, the first decentralized networks like Bitcoin or Ethereum already used this architecture when they were created. Blockchains, as we understand them today, are actually layer 1s. They are the first rung of the pyramid of layers and do not depend on any other blockchain (layer 0s being sets of components like miners and not blockchains).
In the vast majority of cases, layer 1s, and in particular those whose goal is to host decentralized applications, choose security and decentralization. The reason for this choice is that it is very complicated, if not impossible, to improve these two parameters at the upper layers.
There are still counterexamples, such as Solana (SOL) or the BNB Chain (BNB). These solutions were both built with scalability at the heart of the thinking. As a result, both decentralization and security have been neglected.
The layers 2
Layer 2 is, in a way, overlays based on layer 1. These take the form of full-fledged blockchains and intervene in order to decongest the mother blockchain. Their range of action being limited by the dependence on the security of the lower layers, the vast majority of layer 2 focuses on scalability. Solutions like Optimism, Arbitrum or Polygon are prime examples of layer 2 seeking to relieve their parent blockchain, Ethereum, through a faster and cheaper blockchain.
Although focused on scalability, some protocols choose to focus on anonymity and confidentiality. This is the case with solutions like Aztec or Nightfall on Ethereum.
This two-layered architecture has allowed some blockchains like Ethereum to come closer and closer to solving the blockchain trilemma. Although integration is not frictionless, layer 1 takes care of security and decentralization while layer 2 improves scalability.
The Two-Layer Architecture Problem
The vast majority of layer 2s are general-purpose blockchains. In other words, they can accommodate all types of decentralized applications. On Polygon, for example, we can see decentralized exchanges like Uniswap cohabiting with non-fungible token platforms (NFTs) like OpenSea.
This versatility, while convenient for users, comes with its own set of problems, the main one being congestion. Indeed, the low transaction costs and the speed of the solution have the main consequence of the influx, not only of new users but also of regulars of layer 1 seeking a faster and less expensive blockchain. This limit is called the vampirization of the parent blockchain.
Beyond the abundance of new users, the versatile nature of layer 2 is attracting new protocols from all walks of life who want to deploy without having to pay exorbitant fees. Thus, a person simply wanting to exchange one cryptocurrency for another may be exposed against his will to the explosion of fees due to NFT protocols. With these three phenomena added together, layer 2 can see itself failing in its primary goal: to reduce costs and offer a decongested experience.
Figure 3: Polygon blockchain fee history
This layer 2 congestion occurs in particular during fashion effects. Polygon is one of the first solutions to have paid the price during the rise of “Play-to-Earn” applications, these video games in which players are paid to play.
One of the solutions to this price spike due to the versatility and low cost of these layer 2s is the development of a new level 3 layer.
Layer 3: Scalable Blockchains for Specific Use
Fractal Hyper Scalability
To put it very simply, layer 3 are to layer 2 what they are to blockchains like Ethereum or Bitcoin. The development of layer 3 comes with the aim of solving the problem due to the versatility of level 2 blockchains. Like the latter, the third layer is based on all the preceding levels.
The main advantage of this architecture is the fractal scalability it generates. Indeed, if each overlay makes it possible to multiply the processing capacity by 1,000, we can quickly arrive at a solution capable of processing more than one million transactions per second.
This fractal hyper scalability notably allows the deployment of high-speed applications such as video games, payment protocols or social networks. Until now, centralized versions of these sectors have dwarfed their decentralized counterparts with their near-free nature.
Level 3 blockchains derive their security from layer 2 on which they are based, like the latter with their parent blockchains. Thus, the security of layer 3 is similar to a pyramid in which everything rests on the lowest level by recursion.
Figure 4: Architecture of a layer 3
Solutions for Specific Use
Beyond this mathematical hyper scalability due to the stacking of layers, layer 3 also benefit from its specialization. These are so-called “application specific” blockchains. In other words, and unlike the majority of Layer 1 and 2, a Tier 3 solution is created with a specific use in mind.
Bitcoin’s Lightning Network is a perfect example of such special-purpose blockchains. This only integrates peer-to-peer transactions, which allows it to process nearly a million transactions per second. Sorare, the famous French application mixing fantasy football and blockchain, also announced in July 2019 that it would deploy on a personalized version of StarkNet, a layer 2 of Ethereum.
The main advantage of separating into several blockchains for specific uses is to separate the uses and, therefore, provide horizontal scalability. Each protocol or sector will not only be able to benefit from a custom-developed blockchain, but its adoption will not come at the expense of the overall user experience.
As briefly mentioned before, these blockchains will benefit from a suite of custom settings and tools to best meet its requirements. A GameFi protocol may, for example, favour a fast network, while a network dedicated to decentralized exchanges will have security and speed at the center of concerns. Experimental applications will also be able to be deployed on canary networks, which are faster and very inexpensive, without hampering more mature protocols (as Kusama does with Polkadot).
An example of layer 3: StarkNet
As of this writing, the development of Tier 3 solutions is still in its infancy. Indeed, although large blockchains like Avalanche have already raised the idea of fractal scalability through the deployment of rollups on subnets, it is the StarkWare team that is the only one to have really invested in detailed research on layer 3.
StarkWare has been developing scalability solutions for Ethereum since 2018. Among their current products, we find StarkNet, a layer 2 of ZK Rollup, and a development engine called StarkEx, allowing the deployment of scalable decentralized exchanges. The latter is notably used by giants like dYdX or Sorare. However, at the time of this writing, StarkEx and StarkNet are both layer 2 and are therefore entirely separate from each other.
StarkWare’s idea is to create a unique level 2 network, a new StarkNet, based on Ethereum directly. It would thus be the only one to directly benefit from the security of its mother blockchain. Many layer 3 would then be built on this unique layer (including the current StarkEx), thus creating an ecosystem of specialized scalable blockchains.
Figure 5: StarkNet Network Ecosystem
An example of such an ecosystem is shown in the diagram above. Several potential applications can be distinguished:
- StarkEx using rollup or validium technology for decentralized exchanges like dYdX or games like Sorare;
- StarkNet forks focused on different aspects such as security, anonymity or scalability ;
- Classic general-purpose StarkNet networks to simply benefit from fractal scalability;
- StarkNet is specific to certain decentralized applications. For example, we can think of a layer 3 dedicated entirely to Uniswap not only to decongest the mother layer but also to improve its performance and user experience.
Layer 3, an intermediate step before infinite scalability
Layer 3 promises hyper scalability coupled with better control of the technology stack and fragmentation of the user base while maintaining the security guarantees provided by Ethereum or another parent blockchain. This is intended to decongest layer 2 blockchains, which are supposed to solve this same cost problem.
Beyond layer 3, companies like StarkWare also propose continuing this momentum of stacking layers by creating layer 4. These would then be based on level 3 layers and would allow quadratic scalability beyond the expectations and needs of current users.
This stacking of scalability layers for specific purposes could allow current blockchains to reach the speed levels of centralized giants like Visa or Mastercard.