Empire Podcast Summary: How EigenLayer Unlocks Infrastructure Innovation
EigenLayer's systemic risk? EigenDAO who dat?
Recently, Sreeram, founder of EigenLayer and Calvin, Chief Strategy Officer (CSO) of EigenLayer joined Jason and Santiago in an Empire episode to discuss about EigenLayer. This was a value packed podcast where the founder and CSO explained in detail what EigenLayer is about and gave a sneak peek of what the future of EigenLayer may be. Having written on EigenLayer a while back, there were questions that I had that weren’t addressed previously. Huge thanks to Jason and Santiago for bringing up some of the questions I had, allowing me to gain some clarity on EigenLayer’s roadmap and how the architecture would look like.
Some of the key components discussed in this podcast include the following:
Why was EigenLayer started?
Why did EigenLayer choose to build on Ethereum?
How does EigenLayer work?
EigenLayer’s slashing mechanism and the downsides
What is the use case and potential for EigenLayer?
6 orders of magnitude of improvement in infrastructure through EigenLayer
Risk management methods employed to deal with systemic risk?
Increased cost of attacking
Concept of marketplace for decentralized trust
EigenLayer’s business model
Difference between Cosmos Interchain Security (ICS) & EigenLayer
As usual, portions in italics are my own thoughts. Hope you enjoy this.
The episode started off with Jason giving his take on what he thinks EigenLayer is:
A protocol that allows validators to rehypothecate $ETH to allow for decentralized applications (dApps) to leverage on Ethereum’s security. This is otherwise known as re-staking.
Why was EigenLayer started?
When we look back at the 2011 blockchain space there was Bitcoin, the first application specific blockchain. It was very limited and Bitcoin was built only for one purpose: P2P transactions.
Thus, when people wanted to do other things such as domain name system or build another blockchain, they had to built a whole new Proof-of-Work (PoW) system. And with that, every new innovation needed a new trust network.
Fast forward to 2013/2014, people started to question why there was a need to create a new trust network for every new application. Instead, we could create a general purpose smart contract platform which is essentially Ethereum today. The core idea is that there will be a common trust network which can be made more programmable. This will allow anybody to go and create new applications they want on top of the common platform. This idea started growing massively and as more applications built on the platform, it led to a network effect where applications that complement the existing applications began building on top too. A flywheel effect.
However, there are still limitations in the open innovation landscape permitted by Ethereum.
For example, if someone wants to change the Ethereum’s consensus protocol, it has to be agreed upon a societal consensus that there is a new Ethereum version. There is massive friction in this as it is impossible to have three kinds of consensus protocol running simultaneously. Hence, it is impossibe for one to innovate at any layer of the blockchain stack that is deeper than the smart contract / application layer inside Ethereum’s system. Simply put, Ethereum gives the programmability that allowed for open innovation at the application layer, but not the layer of consensus protocol, scaling, infrastructure etc.
This resulted in a scenario where any new innovation deeper than the application layer had to resort to building their own Layer 1 (L1). Examples of which are Solana, Avalanche, Polkadot and Cosmos.
Sreeram wanted to innovate and had new ideas that he wished to build. However, he was required to build an entirely new L1. This to him seemed insane as creating a decentralized trust network is not a normal job, neither is it the job an innovator/programmer/distributed system builder. This bothered him significantly as the core value proposition of crypto was permissionless innocation, yet there was a high barrier to permissionless innovation at the infrastructure layer. This became a bottleneck and very few experiments could be conducted at the infrastructure layer.
Sreeram pondered over the fact that there were already existing decentralized trust networks such as Bitcoin and Ethereum, and thought about whether there was a way to somehow borrow it and build other infrastructure such as oracles. The concept was as such: Borrow decentralized trust from existing networks and supply it flexibly to anybody who wishes to consume it.
Having felt that the core value proposition of blockchain is decentralized trust, the idea of EigenLayer came to Sreeram, where a marketplace could be built to buy, sell and rent decentralized trust easily.
Why did EigenLayer choose to build on Ethereum?
Santiago posted the following question:
EigenLayer is essentially security as a service. By creating EigenLayer, security is being rehypothecated. What are the security implications and how does one believe that the trust security actually works?
To which Sreeram answered:
There are two essential ingredients that the EigenLayer team found was very useful for creating flexible programming semantics on top of a common decentralized trust network: (1) Proof-of-Stake (PoS) system (2) Full programmability. These are the two factors that Bitcoin does not possess.
PoS System
This system possesses the unique ability to not only provide positive incentives for stakers to opt tin, but also negative incentives when there is misbehavior. This allows for the construction of a completely decentralized pseudonymous world with in built tight enforcement (where it is made sure that validators and nodes are behaving correctly).
The PoS system is able to build negative incentives (e.g. slashing) into the protocol because of its additional degree of freedom (DOF), which comes about because the stake itself is a digital representation of an asset. When a node misbehaves and acts in a manner that is malicious, the protocol will be able to burn up the corresponding stake.
On the other hand, the design space is very constrained for a PoW system. There is only positive incentives in a PoW system. It is not possible to have negative incentives in a PoW system. For example, if a node misbehaviors, it is impossible for the protocol to burn the mining rig.
Full Programmability
The slashing that PoS displays is that of programmable slashing, where stakers are opting in to not just one type of misbehavior, but the defined misbehaviors across various protocols.
How does EigenLayer work?
It should be noted that EigenLayer is not a new L1 or a network on its own. It is a series of smart contracts on Ethereum that allows stakers to express their intent in participating in additional network.
There are two sides to this decentralized marketplace:
Stakers: Express interest to opt in due to favourable economics
Protocol: Build on top of EigenLayer. Stakers will download and run the node. The protocol has to make sure that stakers are behaving well and thus create slashing conditions on top of EigenLayer. Once the staker misbehaves, EigenLayer will slash the stake.
EigenLayer’s slashing mechanism and the downsides
Santiago asked the following question:
If stakers get slashed at the smart contract layer for the new protocol, does this introduce latency into the validation of the L1, which in this case is Ethereum.
Sreeram’s answer:
To answer the above question, we first have to look into the two distinct types of restaking on EigenLayer.
Liquid Restaking
Liquid staking derivatives (LSDs) such as staked $ETH can be locked into EigenLayer’s smart contracts, which can be committed to other services. When stakers do liquid restaking, they are giving explicit permission for EigenLayer contracts to slash the $ETH if they misbehave according to a certain set of slashing conditions.
In this case, there is nothing to do with the Ethereum core staking protocol. The LSD platform gives a layer of abstraction.
A questions that I have in mind for this would be: Doesn’t the $ETH ultimately lie within the core Ethereum protocol? Because the LSD platform stakes to the core Ethereum protocol. Hence, once the slashing has been conduced on the LSD platform, the underlying $ETH will be slashed too.
Native Restaking
This is a much more powerful way of restaking.
Native restakers refer to solo stakers or home stakers that wish to provide decentralization without having to trust any operator set such as Lido or Coinbase. Native stakers are able to specify the withdrawal credentials, meaning they can choose who has the power to withdraw the staked $ETH.
Normally, native stakers will set the withdrawal power to their own address. In this case, EigenLayer will ask stakers to set the withdrawal power to EigenLayer’s smart contract, which will give EigenLayer the power to withdraw the stake on user’s behalf to their address.
In the case of malicious actions, the EigenLayer smart contracts can trigger the withdrawal from Ethereum. This will allow Ethereum to know that the staker is no longer honest and will not be allowed to continue staking $ETH.
What is the use case and potential for EigenLayer?
Below is the point of view from Calvin on ‘why EigenLayer’:
For developers
The crux issue here that we face is that any innovation beneath the application layer will require the building of your own validator set.
When a new network is started and tokens are launched, a story has to be told about how this initially worthless asset is actually worth something. This is because the token can be staked in the network and serve as economic security for the activity that is happening on top of the network. Essentially, the economic security for a network has to be bootstrapped from zero.
However, the core competency that developers and builders bring about are developing infrastructure on top of the blockchain, and not building a validator set. Especially so when building a validator set is a very time and effort intensive process which includes:
Talking to validators
Launching of tokens
Attracting market to stake for the network to even work
Leveraging on Ethereum’s Highly Secure Trust Network
Ethereum by far, has the most decentralized validator set. Protocols building on EigenLayer can tap on this highly decentralized set quickly, renting a huge amount of staked security from Ethereum.
The vision that we can have is: Ethereum validator set will serve as a platform. And Ethereum would just be one application that resides on this platform, alongside many other applications built on it.
Jason then clarified regarding the technology:
In that case, applications that are built on top of EigenLayer are not really the front end live applications that users interact with. Rather, it goes one or two layers lower the application layer: infrastructure, middleware, bottomware. Examples of which are rollup sequencers, bridges, oracles and RPC nodes.
Sreeram followed up with an answer on EigenLayer’s use case, primarily focused on the concept of open innovation.
Ethereum brought about an environment where pseudonymous founders, people without prior reputation or public identity, can create dApps that are worth millions and billions, creating the open innovation phenomenon. This is possible because users do not have to know who are the founders, they just have to look at the dApp and its smart contract to understand what is going on. With this, there is no need to trust the founder, only the Ethereum network needs to be trusted. These founders are essentially borrowing Ethereum’s trust, which is the underlying source of a pseudonymous economy.
If the above philosophy of anybody can create a dApp is taken, and unlocked at the infrastructure level, this would mean that anybody can come and upgrade the infrastructure. This is a powerful force that can transform the current crypto landscape.
The above vision can also be supported by how the Ethereum roadmap has developed. The core idea was sharding, where the validator set is split into small groups and each group will execute a sub set of the program. In this original roadmap, only one team will be working on a single idea that has to be planned and deployed over many years. This warranted the need to be conservative and the set of ideas that could be explored was very minimal, leading to a bottleneck in innovation. With the migration to a rollup-centric roadmap, anybody can come and create the execution layers, with all of them competing in a free market. This unlocked a massive tsunami of innovation where the competition leads to the best innovation in the execution layer, cost and other aspects.
EigenLayer takes this concept of the rollup-centric roadmap and turbo charges it to one layer deeper. One key differentiation that EigenLayer has from rollups is that it allows for protocols to change the consensus mechanism they wish to employ.
6 orders of magnitude of improvement in infrastructure through EigenLayer
The following section depicts the level of impact Sreeram predicts EigenLayer having.
One way to view the blockchain or Ethereum’s performance is: It is a place where people are writing data, and all these data is being published into a common ledger. Execution is off loaded to Layer 2 (L2), and the subsequent executed transactions are published publicly on Ethereum, together with proof of what will happen when the transactions are executed.
However, all of the above is still ultimately dependent on the data bandwidth of Ethereum (rate at which data can be written onto Ethereum), where the final data is published.
As of this current state, should we assume Ethereum is not used for other functions (computation etc.), and is purely used to write the data for the executed transactions, the throughput is 83 kb/s.
When we look at bandwidth requirements of Ethereum validator nodes, it is around 2 - 3 mb/s. The theoretical bandwidth available on Ethereum is huge while the practical usage is very low. EigenLayer is challenging the paradigm to find out how permissionless competitive innovation can unleash this.
The current landscape only allows Ethereum Foundation to upgrade the data bandwidth. They are constrained in the sense they have to take a conservative approach and cannot induce any abrupt change given the amount of economic activity taking place on top of them.
By overlaying an intermediary layer - EigenLayer, people can essentially take any feature of Ethereum and challenge to make it better.
Risk management methods employed to deal with systemic risk?
Santiago posted the question:
If a lot of validators interact with EigenLayer, does that make Ethereum riskier?
If there is a smart contract bug that occurs within some of the dApps interacting with EigenLayer, does it affect Ethereum and will this break Ethereum at some point?
Sreeram’s answer:
There are two risk management methods that are employed by EigenLayer
EigenDAO to manage smart contract bugs on EigenLayer / protocols built on EigenLayer
This revolves around the concern that bad actors are not slashed or honest actors being slashed. In this scenario, $ETH stakers will lose their $ETH. This does not come across as a risk to just the protocols building on top of EigenLayer but to the ecosystem as a whole.
Theoretically, however, this systemic risk should not exist given that stakers who opt in understand what they are opting into. However, the current sophistication level of smart contract audits and formal verification is not yet at the stage where there will not be any discrepancies and mistakes. As such, the effects of $ETH slashed will still flow through the system.
To manage the above risk mentioned, EigenDAO will be setup. EigenDAO adds a layer of human subjectivity by appointing reputed Ethereum members to form a slashing committee to man this slashing veto. This is the process:
Step 1: Slashing is triggered. At this stage, the stake is not slashed immediately as there resides a possibility that the slashing is illegitimate.
Step 2: There will be a time lapse in which the slashing committee will determine whether this slashing is indeed accurate. Once it has been confirmed that it is indeed slashing bad actors, the slashing committee will activate the veto and the slashing will be carried out.
This creates a two layer system: (1) Smart contract to trigger slashing (2) Veto committee to approve of the slashing. This buffer system will protect against slashing cascading risk.
Santiago questioned whether the DAO might be corrupted. To which Sreeram answered:
There will be this fundamental one layer of trust in which stakers and services have to trust before they opt in into EigenLayer. When the EigenDAO has a change in committee members, there will be a one month delay before the actuating of the committee. During this time gap, should stakers and services not trust the new committee, they can choose to opt out. To this, I personally am skeptical and think that it brings about a lot of new issues. Example of which is where will these services go? Should they already have users, how do they suddenly find a new chain to migrate to within a month to achieve the same level of security?
Nonetheless, the roadmap looks towards transitioning away from this veto committee eventually.
The other question brought up by Santiago was:
We are talking about specialised smart contracts for oracles, storage, sequencers etc. These sectors require a a high level of technical sophistication from the veto committee. How will it be possible to make sure the committee members chosen have the technical expertise to make a proper assessment as to whether to approve the slashing. This seems like a huge bottleneck by placing the trust on a small committee that is required to have the technical expertise across a vast number of use cases. I also question the scalability of this.
Sreeram’s answer: The committee will consist of individuals such as Ethereum core developers, L2 builders, MEV solutions builders. The committee will essentially be made up of deep distributed system builders. This is not a subjective judgement that will have human bias factored in. It is a very clear and obvious decision as to whether there is a mistake that has been made by the node.
Ethereum’s economic security layer is constantly increasing
First order of reasoning: When new applications are built on Ethereum, it increases the benefits of attacking it. This would mean that it is an unstable system and fundamentally cannot have applications built on it. The profit of attacking increases whereas the cost remains the same.
Second order of reasoning: Ethereum was worth 100 million just four years back. Today it is a 300 billion ecosystem and the $ETH token has grew significantly too. This is because when a new dApp is built, it does not immediately increase the amount of security, but it increases the amount of fees paid back to stakers. This value accrual to the $ETH token increases the staking returns. This compound effect will increase security in the long term.
Third order of reasoning: By expanding the dApp programmability and bring able to specify what each node in the Ethereum ecosystem does, it massively increases the total value being written on Ethereum, with some of this value flowing into the $ETH token, and once again increasing the economic security layer.
Increase cost of attacking
Sreeram commented on how EigenLayer brings about a higher attacking cost.
dApps today depend on Ethereum’s consensus but also relies on other services such as Chainlink, bridges, data availability layers etc. The cost of attacking these dApps is the minimum of the security of such auxiliary services. Yet, with such low cost of attacking, there are massive failure points created by attacking them. For example, by attacking a cross chain bridge, the ecosystem using it will fall and funds will be lost.
With EigenLayer, all dApps are built on Ethereum’s security. To attack any one service, a malicious attacker will have to attack the entire Ethereum economic stake.
Concept of marketplace for decentralized trust
Jason commented on the decentralized trust marketplace concept.
Given that marketplaces are essentially about demand and supply, and there are different stakeholders that EigenLayer deals with: Validators, Stakers, Protocols, which is the most critical party that has to be attracted to adopt EigenLayer.
Calvin’s answer:
There is a flywheel effect to this. When protocols building on EigenLayer can get substantial amount of staking, this will attract more developers to build on top of EigenLayer. This generate fees which will go back to stakers, and this economic incentive will attract more stakers.
Hence, it should start with attracting stakers and validators.
EigenLayer’s business model
This is still in development but it will potentially be a marketplace between developer and stakers. Stakers will stake their assets into EigenLayer for developers to use. EigenLayer will then keep some of the fees in the treasury to serve as insurance.
Difference between Cosmos Interchain Security (ICS) & EigenLayer
This question was brought up but didn’t managed to be answered due to the lack of time. But this is a very interesting point that I would be interested in knowing more about.
All in all, this podcast was filled with a lot of value and gave us more insights and concrete ideas on what EigenLayer could possibly be like.
If you liked this, do follow us at @ProblyNth on Twitter and subscribe to our substack for more of such content.