Sync my mirror: https://mirror.xyz/jojonas1.eth
Blog site: https://jojonas.xyz
——————
This is the first article in the 2022 Blockchain Industry Review series, attempting to discuss some understanding of the Node Layer field. Author's Twitter: @jojonas_xyz Welcome to communicate.
The node layer is an indispensable foundation for blockchain. Although the most encouraged behavior is for users to run their own nodes to make the blockchain network more decentralized and reduce the risk of individual participants attacking the network. However, the technical setup and hardware investment of a network node are not "user-friendly," so hosted node service providers such as Infura and Alchemy, as well as staking service providers such as Lido and Rocket Pool, have emerged.
Node Service Providers
The emergence of Infura and others is very consistent with business logic, that is, achieving economies of scale and cost reduction through specialization and providing convenience to users. However, Infura and others will also bring increasingly severe centralization and censorship resistance issues. Metamask's multiple outages originated from Infura, and Infura was also "unavoidable" when Tornado was sanctioned.
In addition, these nodes are basically hosted on centralized cloud servers such as AWS, which means that the decentralized blockchain you think you have is still largely controlled by others.
source: https://ethernodes.org/networkType/Hosting
Infura actually reflects a typical "Web2 platform" problem, that is, it was originally created to improve efficiency and reduce costs. Once it grows into a giant, whether or not it has subjective intentions, it inevitably carries the "original sin" (the ability to do evil) and becomes a time bomb.
As for the countermeasures, of course, decentralized infrastructure services can be developed, but I think the key point is to solve the problem of "user-built nodes" itself. The reason why users are unwilling to run their own nodes is mainly due to the complicated technical setup and hardware requirements. What if running a node itself is as simple as installing software?
Light clients are a good direction. By using a large number of light nodes and a certain number of censorship-resistant full nodes, the cost of consensus maintenance can be reduced while maintaining the consensus mechanism. However, the behavior of edge nodes is extremely unstable from a behavioral perspective, and the existing incentive system design does not seem to solve this problem.
What about decentralized node service providers like Pocket Network? Pocket itself does not provide node services but creates a bilateral market, one side is the demand similar to Infura, and the other side is various node service providers who provide services and earn income through staking tokens. It seems flawless, but personally, I still don't like this kind of multi-layered design. Adding more layers of protocols actually adds more layers of risks. Pocket's application will encounter a paradox: if the security of the chain is higher than Pocket, using Pocket will increase the risk; if the security of the chain is lower than Pocket, it will lower the credibility of Pocket itself.
Continuing the train of thought from Pocket, going back to the problem mentioned when discussing Infura, another way to think about it is, if we allow centralized hosted node services to continue as the mainstream of the market, is there a way to curb their wrongdoing? Just a random idea, if we use the MPC approach to decentralize the control of hosted assets (possibly certain private keys) by Infura to multiple identities through some technical means. For example, Infura and users each hold a portion, and users hold veto rights, it can effectively avoid wrongdoing.
Staking Service Providers
While node hosting can solve the technical and hardware threshold for setting up nodes, it does not solve the problem of the minimum requirement of 32 ETH per account. 32 ETH is not a very accessible threshold, which means that most users are not eligible to participate in network maintenance (and profit), which contradicts the technical philosophy of blockchain.
Another point is that before the Shanghai upgrade, the staked ETH could not be withdrawn at any time, which was equivalent to depositing a possible deferred time deposit in a bank, and when it was withdrawn, it might be less than half of the original market price. (Just suddenly remembered a stranger in the LUNA wave who staked millions of dollars in Cosmos and couldn't withdraw it, and watched it go to zero... ETH shouldn't be so miserable)
This is where staking service providers come in. As the name suggests, it is a way for ordinary users to participate in blockchain consensus maintenance. Staking service providers receive user deposits, operate their own nodes, or delegate the consolidated assets to node service providers such as F2Pool, Lido, Rocket Pool, etc.
I think there are two main reasons for the existence of staking service providers. First, like node service providers, users sacrifice a certain level of security for convenience. Second, it lowers the participation threshold, such as the 32 ETH staking requirement for Ethereum. The problems that staking service providers face are basically the same as node service providers, and the risks are even higher in the delegation model. (From this perspective, users' demand for convenience and low thresholds is actually higher than security, even if they don't admit it in words; perhaps because their perception of risk is not as direct as the former)
Both Lido and Rocket Pool choose the delegation model. Lido DAO has the right to choose node service providers, while Rocket Pool adopts a permissionless approach, where node service providers provide 16 ETH and the remaining 16 ETH is provided by user assets, and they need to stake an additional 1.6 ETH worth of RPL tokens.
These are two distinct paths. The former tries to use community governance to deal with the wrongdoing of node service providers in advance, while the latter adheres to the permissionless concept and uses the slash approach. Personally, I prefer the latter because I have little confidence in inefficient DAO governance. How does the community identify the behavioral tendencies of node service providers? I'm not sure about the details, but the first thing that comes to mind is rent-seeking. On the other hand, slash has been adopted from ETH to Optimism, and I believe there is a more complete mechanism.
And Rocket Pool's mechanism made me think of a possibility: a "market of autonomous node service providers." Suppose there are three node service providers, A, B, and C, each of which stakes a deposit and 16 ETH. Users are actually unsure how to choose among A, B, and C. But suppose the staking yield is 5%, and the protocol charges a unified fee of 0.5%. Originally, users would get 4.5%. Now, if ABC can choose their own fee strategy, things get interesting (imagine community group buying):
Service Provider A says: From September 15th to November, the yield is 4.7%, we only need 2 more ETH to start, hurry up!
User: I have 1 ETH here, it will be available in November, count me in!
...
DVT Decentralized Validator Technology
As mentioned earlier, a major problem with staking service providers is that it is difficult to ensure that they will not do evil when delegated to node service providers. Solutions like Rocket Pool have introduced a slash mechanism, but it is ultimately a band-aid solution and not a fundamental solution. Therefore, the DVT approach emerged, which fragments the private keys of the validating nodes (I actually looked at the technical description of DVT, and it seems that there really is MPC as I mentioned before).
The big name in the DVT field is ssv.network. SSV is the predecessor of DVT. After understanding what DVT does, you will find that some things I mentioned earlier have changed: Lido can completely solve the biggest problem it faces by cooperating with SSV, which will be the biggest challenge to Rocket Pool's rise.
There are even more ambitious projects, such as Diva, which directly combines LSD and DVT to create a permissionless staking service. Users directly stake ETH and hold their own private keys, and Diva also provides a machine gun pool. Of course, this kind of all-in approach can either be very exciting or very disastrous. Regardless of the outcome, it still makes Lido and similar projects feel a sense of crisis.
LSD
LSD, which stands for Liquid Staking Derivatives, actually refers to the staking service providers mentioned earlier, which gained some attention recently.
The general logic of the market is as follows: compared to other chains, the staking rate of ETH has always been low (due to various reasons, such as before the Shanghai upgrade, staked ETH could not be withdrawn immediately; ETH itself is an asset with excellent liquidity, and its value does not need to be realized through staking; non-direct stakers need to bear additional protocol risks).
After the Shanghai upgrade, ETH can be flexibly unstaked, and even if it only reaches the average level of other chains, based on the current staking value of over 27 billion USD, LSD will become a trillion-dollar race.
source: https://www.stakingrewards.com/
After entering the LSD field, Rocket Pool is no longer the only competitor of Lido, but the competition will expand to the entire DeFi space. The reason is simple: users do not necessarily need to stake their ETH to maintain network security, and they may not even have this awareness. The key factor that will determine the outcome of the LSD war is the secure and stable yield. Whoever has the higher yield will have the greatest advantage.
The previous wave of DeFi stacking and Lego stacking must have left a deep impression, right?
Our clever Frax is here.
As one of the few survivors in the stablecoin field (Frax is somewhat stable), Frax created higher yields by stacking frxETH and sfrxETH, although the operation is more complicated, it may be a gameplay that DeFi veterans prefer.
source: https://www.convexfinance.com/stake
In addition to Frax, lending protocols, DEXs, and others are also potential competitors. Let's think of ETH as a big pie, and cutting off a large part of it for pure liquidity without staking, Lido, Frax, MakerDAO, etc., will share the other half. Although the Shanghai upgrade is good for Lido, this dividend may not be exclusively enjoyed by Lido, but it is more likely that DeFi veterans will leverage their stacking and Lego stacking nature on top of ETH's deflation and instant unstaking to come up with more innovative ideas that Lido couldn't even think of. After all, everyone wants to lie down and make money, and this thing cannot be without ambition...
Looking at it from another perspective, LSD, which can be unstaked at any time, is actually competing with traditional DeFi (DEXs, lending, etc.) for liquidity. So the basic protocols of traditional DeFi should think about how to position LSD (if they continue to use, for example, stETH as the underlying, their own pirated ETH may become less and less relevant because, in the end, the liquidity incentive provided by LSD can be considered real yield) and how to seize the opportunity on the new underlying. I think there will be a day when the swords are drawn.
Node Reuse (Re-staking)
(Node reuse is a term I made up, generally referred to as re-staking)
I've talked a lot, but the project that impressed me the most this year is EigenLayer.
I mentioned it on my previous tweet:
Compared to some "Lego stacking" and "Parasitic" innovations, I think the innovation of EigenLayer is more advanced. These innovations are impressive, but in the end, the former often has beneficiaries who suffer losses, while the latter can truly achieve a win-win situation. In economic terms, it achieves Pareto improvement.
Let's discuss what it can do: The ETH used for node staking, in addition to maintaining the ETH network, can also serve as the node layer for other protocols/networks to maintain their security.
Let's analyze:
1⃣️ For ETH, any protocol that uses EigenLayer is actually entering the ETH ecosystem. The growth of these protocols will correspondingly bring growth to ETH. And this will give ETH an unprecedented level of influence (unprecedented because it is already high).
2⃣️ For stakers. In addition to earning ETH from staking, they can also receive incentives from other protocols at the same time. Why not?
3⃣️ For major protocols. The previously troublesome consensus security issue is pleasantly resolved. And they can be tied to ETH dad. And developers can finally focus on solving real problems (without cursing).
4⃣️ For staking service providers like Lido. First, they probably don't need to fight against DeFi veterans anymore; second, LSD has become the consensus hub driving various networks and protocols, what more benefits can they ask for?
5⃣️ For EigenLayer itself. EigenLayer is still essentially a platform, with stakers on one side and major protocols on the other. Similar platforms are actually difficult to fully leverage the network effect of both sides because one side often holds the initiative. Just like Pinduoduo using three tactics to forcefully enter the e-commerce market, this is because users are free to come and go, and new platforms are cheaper, so they can leave whenever they want...
What about EigenLayer? Stakers ➡️ Network security ➡️ Attracting major protocols ➡️ Attracting stakers. New competitors will not be able to offer comparable yields (can't be like Pinduoduo), nor can they provide satisfactory security for major protocols (can't be like JD.com). Of course, here I assume that EigenLayer is exclusive, meaning that stakers who use its service cannot use a second re-staking protocol, otherwise, the moat is not that high because rational stakers will definitely use all available options, and EigenLayer would be like making a wedding dress for others.
Node reuse will be a major trend, and its fundamental logic is the same as DeFi Lego stacking, which is the same as what we like to call "composability." This is one of the few ways to improve efficiency in our redundant decentralized world.
References
https://web3edge.io/fundamentals/master-web3-from-node-to-network/
https://newsletter.banklesshq.com/p/ethereum-censored-flashbots-centralization-lido
