Editor's note: This article is from medium.com, author: Jacob Arluck, translation: Leo, published with permission.
Preface: Is Tezos' LPOS the same as EOS's dPOS? what is the difference? The block producers of EOS are fixed super nodes, with high performance as the priority. On the other hand, the block producers of Tezos are dynamic, and the threshold for block producers is lower, with security as the priority. In addition, Tezos block producers not only have block production rights, but also have voting rights for on-chain governance. Will this lead to the centralization of interests? Different consensus mechanisms and governance models are in practice, and what kind of results will there be? Welcome to leave a message to discuss.
Tezos allows token holders to transfer or delegate validation rights to other token holders without transferring ownership. This confuses many observers, ignoring Tezos' consensus mechanism and confusing it with EOS or Lisk's dPOS model.
In dPOS, in order to achieve network consensus, it is necessary to elect a fixed set of block producers, the so-called delegated representatives. In the Tezos network, delegates are only optional. The consensus mechanism of Tezos is more like "LPOS (liquid POS)". The goal of LPOS is to maintain a dynamic validator group to facilitate token holder coordination and accountability management.
secondary title
Block Production in Tezos
To obtain the right to create blocks, Bitcoin miners need to solve mathematical problems, which is the mechanism of PoW. Tezos allocates block production rights based on equity, a mechanism called PoS. Each block is generated by a random stake holder and verified (or notarized, approved) by other 32 random stake holders.
Like Bitcoin, Tezos also rewards producers who participate in the consensus by issuing additional tokens and transaction fees. To incentivize honest behavior during block production, Tezos requires producers to stake tokens for a few weeks. If a block producer attempts to double-produce or double-sign a block, the staked security deposit is forfeited.
Token holders may not be interested in producing blocks themselves, so they can entrust others to produce blocks, and at the same time, the ownership of the token has not been transferred.
As said before, "for me, this is the producer's life": since not every token holder is interested in becoming a producer, tokens can be delegated to others. The delegated delegate does not own or control the token. In particular, delegated delegates cannot spend these tokens. However, if the stake represented by these tokens is randomly selected for block production, then the rights belong to the delegated representatives.
secondary title
dPOS
Because Tezos allows for delegated delegation, its consensus mechanism is often described as dPOS. This confuses many people and confuses it with EOS, BitShares, ARK, LISK, etc.
However, in the standard model of dPOS, the number of block producers or witness representatives is fixed. For example, EOS has a fixed number of 21 block producers, but it is optional in Tezos. The use of dPOS consensus is usually a matter of scalability, and dPOS adopters generally believe that scalability is the main obstacle for dapps to enter the mainstream. In the words of the EOS white paper:
“The EOS.IO software introduces a new blockchain architecture designed to enable vertical and horizontal scaling of decentralized applications. … The ultimate technology is a blockchain architecture that can scale to hundreds of Tens of thousands of transactions, remove user fees, and deploy decentralized applications quickly and easily.”
The dPOS adopted by EOS also adds BFT, and Dan Larimer claims that the EOS network can reach 99.999% final state in less than a second. In other words, dPOS may finally let us abandon the era where a group of cryptokitties can cause congestion to the world's computers. (Translator: meaning that you don’t have to worry about Ethereum’s congestion anymore)
In order to keep a group of block producers honest and accountable, the dPOS protocol continuously votes for a group of representative nodes to publish and verify blocks. In EOS, token holders elect 21 block producers and 100 backup block producers, and the results are calculated every two minutes. In Lisk, token holders continuously elect 101 delegates. The voting rights in the dPOS protocol are usually proportional to the rights and interests of token holders.
Larimer believes that compared with Ethereum and Bitcoin, dPOS has a greater degree of decentralization because mining activities have already tended to be centralized. Tezos allocates block production rights according to the proportion of stake. On the other hand, dPOS distributes block production rights evenly among a group of active block producers. In addition to the risks of Sybil attacks and cartels, the 21 most-voted validators technically have as many votes as the most popular block producers.
With a fixed set of validators, dPOS also requires low inflation to incentivize block production. Of its 5% annual inflation rate, only 1% is allocated to block producers, with the remaining 4% allocated to an on-chain treasury that rewards contributions to the network.
secondary title
Towards LPOS
Unlike dPOS, in Tezos delegation of authority is not an important aspect of PoS. The project prioritizes decentralization, coordination, and security over scalability that will soon be available, and carefully tuned throughput (currently 40tps).
In fact, Tezos’ style of PoS is more akin to Liquid Proof of Stake than dPoS. In order not to be confused, and to better explain Tezos' PoS, the consensus mechanism of Tezos is called LPOS (Liquid Proof-of-Stake).
A few weeks after the testnet launch, Tezos had 411 block producers. However, as the graph below shows, some churn in the delegate market can be seen, as delegates compete based on fees, payment frequency, reputation, and other metrics.
In Tezos, delegate misbehavior (e.g., not paying rewards) can be quickly discovered and the community condemn delegates for charging high fees. Token holders can change delegates effortlessly, so there is a deterrent effect on delegates, incentivizing them to coordinate.
The above is the distribution of block producers on the Tezos testnet from the first to the fifteenth cycle. Since its release, there has been a clear trend towards decentralization of block production, and with the adoption of better infrastructure, it will move further toward decentralization.
secondary title
The Impact of On-Chain Governance
All of these are important for on-chain governance. Because in Tezos, consensus participants are also decision makers for protocol upgrades. The token holder grants the representative the block production right and also grants the voting right.
On this issue, many well-known bloggers have warned that on-chain governance is destined to be captured by central interests. Even, this already has real proof in Lisk. However, observers often extrapolate the experience of the dPOS protocol directly to Tezos, and dPOS has fixed nodes. Few observers realize that Tezos is based on different models with different priorities.
Of course, this is not to say that in Tezos, powerful voting coalitions will not appear, or that some large entities, such as exchanges, will not become important stakeholders. However, the most important thing is that both token holders and users can coordinate the rearrangement of power and make the central interests accountable. (Translator: Here, Tezos tries to paint a picture of self-correction and evolution, of course, all this will take time and practice to prove)
To achieve this goal, LPOS should ensure that the threshold for block producers is low enough to maintain a dynamic validator group. But, like information security, accountable governance is a challenge, one that can never be fully resolved.
Alternatively, block production and voting rights could be completely separated, reducing the link between consensus and on-chain governance. This requires changing the strong default, where users may delegate block production rights to one party and voting rights to other parties. Other potential balances aside. However, this is an interesting area of research.
