Harmony CTO, Rongjian Lan recently concluded an AMA session with a leading staking service provider, Wetez, on Mar 27. Wetez is a professional staking-as-service provider with 10,000 delegators and $50M worth of assets staked in more than 20 projects including Cosmos and Tezos.
Below is the transcript of the event, edited lightly for clarity:
Q: What problem is Harmony trying to solve? What is your vision for Harmony?
[Rongjian]: When Harmony started in 2018, the most prominent problem in blockchain was Ethereum’s performance problem. Many other projects have also been trying to solve this problem. However, these proposed solutions cannot improve performance significantly without sacrificing other critical aspects, such as security and decentralization. Highly centralized systems like EOS only have a small number of supernodes on the network. Whereas Harmony uses sharding and PoS as a solution thus preserves decentralization and security.
There are 1000 nodes on our mainnet, processing transactions on four shards concurrently. Harmony not only shards the network communication and transaction validation but also shards the blockchain state. This makes us a fully scalable blockchain since the total transaction throughput increases linearly as the number of shards grows. We want to achieve high-performance, low-cost but maintain Ethereum’s level of security and decentralization.
In terms of vision, we want to provide the world with a scalable and secure public blockchain system that can support the emerging decentralized economy and enable application scenarios previously not feasible. We want Harmony to become a globalized infrastructure for decentralization. At the current stage, daily applications have a throughput level of thousand TPS, which is only possible through sharding.
Q: Compared to other sharded blockchains, e.g. QuarkChain, Near, Elrond, what is Harmony’s advantage?
[Rongjian]: Each project features different sharding approaches.
For example, QuarkChain originally did sharding on POW consensus and then shifted to State Partitioning.
For NEAR protocol, there is no shard chain in the network, but only a main chain where data is sharded and processed separately.
Elrond launched on Binance after Harmony and we share many similarities! However, their consensus algorithm is based on Algorand, and their PoS Staking mechanism has not yet been finalized.
Harmony’s major advantages are:
- BLS multi-sig, 8-second-block propagation, instant finality;
- Secured sharding, VRF + VDF eliminates last-revealer attack;
- EPoS mechanism avoids stake centralization;
- Mainnet with 1000 nodes and integrates with major exchanges;
- 8-second transaction confirmation on MathWallet, TrustWallet, Ledger, Safepal.
Our key innovations are in state sharding (secure staking and resharding with decentralized randomness). A shard only processes, validates its transactions, and stores its state instead of the whole blockchain, processing only transactions that are relevant to itself. Other sharding approaches can increase throughput, but without state sharding, shards are not independent thus transactions must be broadcasted to the whole network. Whereas, state sharding doesn’t require cross-shard communications apart from cross-shard transactions, thus enabling scalability.
Our network contains a beacon chain and multiple shard chains. The beacon chain has more functionality, serving as the randomness beacon and identity register, and more importantly, processes all staking transactions. The shard chains store separate blockchain states and process transactions concurrently. Harmony’s sharding structure is more balanced with no bottlenecks.
In terms of security, Harmony proposes an efficient algorithm for randomness generation by combining Verifiable Random Function (VRF) and Verifiable Delay Function (VDF), which ensures the sharding process is unpredictable and unbiased thus preventing 1% attack. 1% attack, i.e. each shard only has 1/n of calculating power and voting power of the whole network, thus an attacker can corrupt a whole shard to double spend with only 1% of the total calculating power if there are 100 shards in total. To guarantee the security of a single shard, the amount of voting shares by malicious validators needs to be kept below ⅓ of all the voting shares in that shard. This is guaranteed algorithmically by using Adaptive-Thresholded PoS and DGR to separate the voting power of big stakers into different shards.
In terms of decentralization, Harmony is the first sharded POS project. With over 400 nodes on each shard, our network can support thousands of nodes. Compared to other POS chains with only 21–100 validators, we are more decentralized and open to non-professional validators.
Lastly, we keep optimizing our protocol. It is known that network capacity is one of the major bottlenecks for blockchain systems. We have adopted Kademlia-based routing as the routing mechanism for cross-shard messages, enabling O(logN) for cross-shard communication and significantly reducing the overall network load in a sharded blockchain.
Those are our technology design, we have made substantial progress since mainnet launch in June 2019. Since then, we have implemented state sharding, cross-shard transactions, view change and resharding against 1% attack. Sharding is also one of the main goals of Ethereum 2.0. I am proud that Harmony is one of the first projects that has launched a mainnet with staking and sharding. We are going to launch open staking on our mainnet very soon, which also means our community will truly participate in our governance. Thus, we have an advantage in both design and execution.
Q: I appreciate the article “setting foundations for the DeFi economy” by Nikos. What does sharding mean for DeFi?
[Rongjian]: DeFi has been hot since last year. I think this area has the best use case for blockchain: the current financial system is a trust-based model that can be supported by blockchain. DeFi is still experimental at the current stage: for example, people can withdraw eth from MakerDAO at zero cost when it plummets. Also, other issues happened on public chains, e.g. orders are not cleared in time, data is not updated from oracle.
Harmony’s goal is to build a high-performance, low-fee public chain as a frictionless platform for DeFi applications. Apart from better infrastructure, I think it also needs a better ecosystem. 1. Oracle should be a part of this ecosystem — we have integrated with Chainlink, the leading oracle service provider. 2. It should also include Stablecoin, which is the “cornerstone” of any financial applications on crypto. Carbon is helping issue an FDIC-insured stablecoin native on Harmony. 3. Another important participant in DeFi is DEX. We have cooperated with DEX including ViteLabs and Hydro Dex to give users a better experience on Harmony.
We are also focused on a very promising market: Cross-fi, i.e. cross-border transaction. Two obvious pain points exist currently: i) high transaction fee; ii) long clearing time. We are working with many licensed partners in Southeast Asia and India to solve this problem for cross-border labors.
Q: Wetez has participated in Harmony’s open staking testnet. What challenges did you guys have in the testnet launch?
[Rongjian]: In Introducing Harmony’s Effective Proof-of-Stake (EPoS), I gave an overview of our staking mechanism. We have innovated in many ways to prevent stake centralization, which also complicates the network logic.
A potential challenge is the high processing speed. High throughput not only comes from hundreds or even thousands of validators but also millions of delegations. We have solved this problem by optimization in our algorithm.
The second challenge comes from the slashing mechanism to prevent non-conforming behaviors. Validators will be kicked out of the committee because of unavailability (downtime). We have one-on-communication with offline validators and provide 24/7 support.
Finally, because EPoS is a relatively un-intuitive mechanism, it takes some time to truly understand and capture the best strategy to maximize returns. Our product designers have spent a lot of time on the staking dashboard to guarantee good user experience.
Q: Please give a short overview of Harmony’s economic model!
[Rongjian]: Most staking models are designed with an optimal staking ratio, i.e. staked tokens/circulating tokens. A good design can navigate usage of tokens, balancing the tradeoff between staking and other usages, e.g. transaction fee or DeFi collaterals. Most chains have a specific target number on this, e.g. 60% for Cosmos and 50% for Polkadot. We haven’t set an optimal number but our expectation is over 60%.
Harmony features a constant annual issuance of 441M ONE. An advantage of this simple model is that it becomes easy for validators to project their future rewards. Yield changes invertedly to staking ratio thus encourages validators to step in and capture return when staking ratio is low and vice versa. Meanwhile, although annual issuance is constant, issuance rate (issuance/total amount) gradually approaches zero after many years., which maintains the value of ONE token. The 441M ONE translates to a 3.5% initial issuance rate. We project the first-year yield of a validator between 10% and 50%, based on the staking ratio.
Q: How is ONE token performing on the secondary market? What is the long term value for investors?
[Rongjian]: Our issuing price at the Binance launchpad in June 2019 was $ 0.003175. It then went up to 0.03$ at its peak, making us one of the best performing projects at that time. However, the whole crypto market was influenced by the economic slowdown in the second half of 2019. Our rank in the crypto industry is relatively stable, around 150 on CoinMarketCap.
In terms of long term value, I would love to mention a few things:
First, a large number of ONE token will be staked after we launch open staking which is strong support on the demand side.
Meanwhile, we will integrate with many other applications on an ecosystem level. We had an $8 million joint acquisition of Quidd with Animoca — Quidd is a marketplace for collectibles. It has 6.8 million users and 2 billion licensed items. Quidd has already achieved $10 million user purchases and averaged 6 transactions/sec. Users will spend Harmony tokens on Quidd.
Our official partner Lympo incentivizes a healthy lifestyle and corporate fitness with reward tokens. They are fully integrated with Samsung Wallet; they have hundreds of thousands of users and millions of exercise challenges completed. There are 140 million profiles for mobile fitness alone. Lympo’s reward tokens will be exclusively built on Harmony and our data-sharing marketplace.
Finally, this year we will also implement decentralized governance. Token holders can participate in global governance with proposals.
Q: Will there be a large number of tokens controlled by Binance? Will Binance become a staking service provider?
[Rongjian]: Binance was holding about 3.8 billion ONE tokens before the token swap. Two months ago, they integrated with the Harmony mainnet and 90% of the tokens have been converted into native ONEs. In other words, Binance no longer holds a large number of Harmony tokens but may still have less than 1 billion tokens.
Binance will be one of our validators and provide corresponding staking services for Binance users.
Q: What are Harmony’s countermeasures to prevent stake centralization? How can decentralized nodes like Wetez contribute to the Harmony’s network?
[Rongjian]: “The rich get richer” problem prevails in all blockchains. Even POW chains like bitcoin or Ethereum with fully decentralized mechanisms cannot avoid voting power centralization.
This problem is more obvious in POS chains. For example, the top 10 validators in Cosmos hold more than 50% of the total staked tokens. This “rank-stake in atoms” chart shows how much atom is staked by each validator. It’s a very skewed distribution.
In most PoS blockchains, validators earn block rewards pro-rata to their stake. Big validators get high block rewards without “extra work”. We introduce Effective Proof-of-Stake, an efficient staking mechanism that avoids stake centralization by forcing big validators to earn higher block rewards through “extra work”. In EPoS, validators are rewarded in proportion to their effective stake, which is capped/floored in a range around the median stake.
Big validators are economically incentivized to decentralize their nodes, i.e. spin up new validator machines to separate it into multiple small nodes. Small validators can get the same return by running a single node.
Higher-ranked validators(in the yellow area) are economically punished to stake too much in a single validator and the lower-ranked validators(in the green area) are enjoying the extra reward. The effective stake is acting as an equalizer that pushes for a more evenly distributed stake among validators, thus avoiding stake centralization.
Q: U.S. stock trading experienced a series of temporary halts in mid-March as reactions to the coronavirus pandemic. What is your opinion on the global market turmoil?
[Rongjian]: The stock rout, spurred by tumbling oil prices and growing fear of the economic impact of the coronavirus, has put the vitality of the longest-ever bull market in jeopardy. The Fed has dropped its benchmark interest rate to zero and launched a new round of quantitative easing program with $700 billion worth of asset purchases of Treasuries and mortgage-backed securities. This only reveals how systematic risks accumulated in the current system leads to cyclical crises inevitably.
Credit risks and mismatching are two major problems in our current system: subprime mortgage and inaccurate credit ratings triggered the 2008 financial crisis; this time hedge funds failed to hedge risk due to forced liquidation under Volcker Rule. Because financial institutions (humans) are playing trusted third parties, the system is prone to collapse due to inside jobs.
Blockchain is the answer. It enables code-based trust to replace human-based trust. Any predefined rules can be executed by a smart contract in a transparent and non-reversible way, which removes the middleman. Bitcoin is designed to be a payment system that removes the central bank to avoid over-issuance. Many DeFi applications are also removing the middleman in the centralized finance world, e.g. lending and market-making.
Opportunities are entailed amid the turmoil. In terms of creating trust, humans will be replaced by algorithms gradually, especially in Finance. Bitcoin might receive recognition on a larger scale through this crisis. As for our team, we have remained very conservative since funding, thus we are still in a good position.