Why blockchain governance matters
Governance and decentralization of blockchain protocols are as important as much spoken scalability and transactions per second. Almost every brand new public blockchain protocol claims to process 1000+ transactions in the testnet and be able to scale to 10 000+ TPS however very few put emphasis on governance and decentralization which are as important as scalability. You can set up a small cluster of servers worth 20 000 USD and process more transactions per second than any blockchain could process at the moment. What really differentiates public blockchain protocol is that it brings in decentralization of power and with the use of cryptography establishes trust between unknown to each other participants without a centralized entity solely responsible for the whole network. Decentralization means that your transaction or data cannot be censored or aborted, that you don’t have to trust centralized institution to process your transaction or store your data, that there is no single point of failure, that barriers to entry (or to vote) are low so there is almost perfect competition between providers and the lowest possible price for resources. But when a network meant to be decentralized become centralized it turns to be even worse than a centralized entity, because in such case you must trust several providers (for example, 21 in EOS) to process diligently your transactions and host your data. Instead of perfect competition you will get cartels, price manipulation and exploits of price inefficiencies.
Both PoW and PoS networks tend to become more and more centralized as requirements to full nodes increase and less and less network members want to participate in transactions processing themselves. However centralization of tokens distribution in dPoS networks is even worth, for example, EOS has 21 block producers being elected by the EOS tokens holders but 90% of all EOS tokens are owned by 1.6% addresses, top 10 EOS wallets hold 50% of all tokens so it means people or organizations behind those addresses (or EOS whales for short) can decide on who will be block producers which is a great threshold for cartels (e.g. block producers compete for whales votes by proposing higher payouts in exchange for votes which has nothing to do with community interests and drives prices for computation resources higher and higher). RAM prices in EOS network are heavily criticized and as the mainnet was launched only in June its too early to draw conclusions however whatever change is proposed without the EOS whales nothing will be implemented. Ethereum has more decentralized governance model which is nevertheless also heavily criticized on being inefficient as for any protocol change to be implemented you have to perform a hard fork and be sure that key network participants from core developers to miners support the change what makes the process of implementing any change extremely long. Moreover, Ethereum uses “the code is law” model which means that if a hacker abuses a bug in the code or hacks centralized exchange and stoles tokens worth hundreds of millions dollars from thousands of users the only way to get money back is to hard fork Ethereum network which is very hard to do and leads to possible split of the network participants which negatively affects network’s value. This inefficiency also leads to the increasing number of well-resourced hackers who are striving hard to find and exploit bugs in the code 24/7 which is again bad for a network.
Tezos — blockchain protocol focused on providing the best governance model for a highly decentralized network of nodes that still runs through law suits and tons of FUD due to the series of mainnet delays and insoluble disputes between the founders nevertheless valued at 726M USD (as at 14.08.2018). Tezos calls its consensus mechanism liquid proof of stake rather than delegated proof of stake as you can start “baking” (mining) new blocks with as little as 10 000 XTZ (equivalent to 12 500 USD) and the number of “bakers” (nodes) is not limited to a particular amount (e.g. 21 in EOS) and can scale to tens of thousands bakers. However, Tezos might be prone to staking pools as the more XTZ tokens you have (or the more XTZ tokens were delegated to you by other XTZ holders that don’t want to directly participate in baking and governance) the higher your chances to “bake” next block. If there is a couple of bakers with a huge stake your chances to bake as an individual node will be getting smaller and smaller. Today there are two nodes controlling almost 20% of roll outs (every 10 000 tokens owned or delegated gives 1 roll out) not counting the Foundation bakers. Existing bakers already offer lucrative conditions for delegating your tokens to them (xtez offers 5.25% return annually and no need for paying deposit fee). Delegating doesn’t mean you send your tokens from the wallet, you just delegate a right to bake and vote on behalf of your stake in the network, you can take back this right at any time while your tokens never leave your wallet. The whole delegation process feels convenient, it will be interesting to see whether Tezos will succeed in attracting lots of bakers and there will be fierce competition to become a delegate and very little voting power assigned to each baker or it will eventually become EOS 2.0 with 20 bakers controlling all roll outs and votes which we think is undesirable by both Tezos community and Foundation.
There is bunch of new projects experimenting with novel governance and token economics models that could (theoretically) provide sufficient decentralization and security levels for the required throughput. Dfinity is one of those projects. Dfinity proposed Blockchain Nervous System (BNS for short) that is governed by the dfinities holders and is able to implement any protocol change without hard forking as soon as the BNS agrees on a proposal. For instance, when the price for dfinities rockets BNS may reduce the amount of dfinities needed to be deposited therefore increasing supply side of difinities, BNS can change almost everything in Dfinity Network, it acts as a sole governance body driven by the voting power of neurons (nodes). It works the following way: “neurons” (nodes) make a proposal by depositing set amount of difinities which are returned after the proposal is adopted and pay a fee which is held by BNS as a reward for those who review the proposal. After that all neurons vote to adopt, reject or pass on a proposal. To earn difinities or participate in Dfinitity governance you must deposit dfinities in neuron. The more dfinities deposited the higher relative voting power of a neuron, to ensure long-term thinking you won’t be able to withdraw deposited dfinities from the neuron for 3 months after you call a function to dissolve your neuron. Neurons will have two options when participating in Dfinity governance: to vote by itself or to follow decision of another neuron that is owned by core team developer or researcher or somebody else you trust to decide on a particular subject (all proposals will be divided into several groups: Economics, Policy, Protocol, Client upgrades etc.). This looks similar to Tezos governance model however Dfinity decouples earning dfinities by staking and mining from voting on proposals and participating in Dfinity network governance. In Tezos you cannot delegate your voting power without delegating your right to bake (mine) new blocks what leads to concentration of both baking and voting power in hands of nodes with the most roll outs. Dfinity governance model is designed to encourage participants to follow the most knowledgeable and trusted neurons rather than those with the biggest deposit. Even though there is lots of negative sentiment on Dfinity fundraising we still think it is a very promising project from the both technical and business (adoption) perspectives. Moreover, every system using PoS consensus (and its variations) is theoretically prone to hostile takeovers in case an adversary stake enough tokens in order to process wrong transactions or adopt bad decisions that destroy the whole network (for example, to freeze every single neuron) therefore the higher the valuation of PoS network, the more distributed it is and the more long-term money it attracted from strategic investors (rather than “strategic flippers”) the higher chances to prevent an attack from an adversary.
Governance and decentralization are as important as scalability for public blockchain protocols which seem to focus heavily on the second issue. Without a proper governance model and sufficient level of decentralization no blockchain protocol will be able to succeed and leverage all the disadvantages of centralized entities. We will be putting more attention to decentralization and governance when assessing the idea of a project as we believe those are the key elements for the long-term success of any public blockchain protocol and should be designed and set in stone before TGE as it would extremely hard to change in decentralized network (or meant to be decentralized) in the future. If there is efficient governance model in place it will be possible to implement layer 1 or 2 scaling solutions and increase throughput but if the governance model is broken and inefficient its almost impossible to change as hard fork would most probably split the network into two even sides.
We decreased Chromapolis long-term score from S to A and Aergo long-term score from A+ to A. Both networks will launch pretty much centralized. Chromapolis team will choose by themselves at least 12 providers at a time of a mainnet launch and then will let existing providers to decide on new nodes. Chromapolis is still on a very early stage of development but its unclear how with such governance model will incentivize providers to keep hosting prices competitive and how the changes to the protocol will be implemented. Aergo plans to have over 20 block producers and approximately 50 candidates for block producers. Enterprises utilizing private chains may choose any number of block producers and candidates. First 2 million blocks will be process by only two block producers — Aergo and Blocko (parent company). Aergo seems to be very centralized in terms of governance, it will be interesting to see whether tokens distribution will lead to the EOS problem (90% of tokens owned by 1,6% addresses) or not. However in general we think dPoS is more appropriate for blockchain protocols target at enterprises as the end users.
Spreadsheet — https://goo.gl/Bk2J5i
TG announcement — https://t.me/token_metrics
Twitter — https://twitter.com/Token_Metrics