1、 Price Oracles
The valuation of the price of kUSD will be achieved through a system of price oracles . In this system, each participant is incentivized to determine the correct price of kUSD based on publicly available sources, and to report this price periodically to the kUSD network by sending an oracle stake to a special smart contract. The participants whose prices are within one sigma of the median will receive rewards minted for this purpose. This approach incentivizes participants to discover and report a trustworthy and accurate consensus value of the market price of kUSD.
As each block is generated, a small amount of the newly minted amount, called the oracle deduction , is placed into a specialized smart contract, called the oracle fund , and these accumulated fees are paid out to oracles as rewards when they report prices. The oracle deduction is determined by the following formula:
The function oracleDeductionFraction(b) is not defined here but will be return values between 0 and 0.01 in such a way that, in normal circumstances, the oracle fund will have a balance sufficient to pay out anticipated rewards to oracles.
We use the term minted reward to describe the the portion of the minted amount that remains to be given to proposers after the oracle deduction is deducted. The amount of the minted reward is simply:
Normally, the reward to oracles is set equal to 4% of the total minted reward over the last 900 blocks (approximately 15 minutes):
where B(m,b) is the subset of the blocks greater than b-900 and less than or equal to b where miner m was the proposer.
However, no reward is paid to oracles unless the proposer is qualified as an oracle during block b. To be qualified as an oracle, a proposer must maintain a minimum stake of mUSD (currently 6 million) and not be banned by a majority of other oracles. This reward is further limited by the actual balance in the oracle fund:
Here is the complete formula for the reward, which we term the oracle reward :
2、 Mining Reward
In Konsensus, the crucial work will be done by proposers. The proposer’s role requires significantly more resource and energy expenditure than the validators’. The proposer must interrogate the blockchain to determine the amount to be minted (see Stability Mechanism 1: Minting Algorithm , above), assemble the block and ensure it is distributed properly — all in a short timeframe — which is non-trivial. Indeed, the processing of blocks would halt without a proposer, but, with a proposer, can continue even if any validators (or even all of them) leave the network.
The total reward earned by the proposer for his or her work is called the mining reward, and is equal to the sum of three components: 1) the minted reward, 2) any applicable oracle reward, and 3) the compute reward:
where m is a miner and b is a block number. The following diagram shows these three components, and the source of their respective values:
3、 Proposer Eligibility, Proposal Stake
Konsensus aims to punish dishonest or free-loading miners by making them ineligible to propose blocks (and thus earn rewards) for a game-theoretically sufficient number of blocks. For example, a would-be proposer must have participated in the prior 10,000 blocks as a validator in order to be eligible to propose the current block.
In order to receive mining rewards, proposers will be required to stake a minimum number of mUSD. For performance reasons, the roster of actively mining nodes has a maximum size (currently 500 nodes). Once the roster is full, a prospective mining node will need to offer to stake that is higher than the lowest stake of all nodes in the roster. A proposed stake that is greater than this lowest stake will be accepted and will bump the node with the lowest stake off the the roster. Furthermore, a proposed stake must be less than the absolute minimum stake amount (currently 30,000 mUSD).
A special-purpose smart contract will be set to track mining activity and stakes. Using this contract, the mUSD owner will specify three things: the staked amount of mUSD, the node that is authorized to mine, and the address that will receive the kUSD mining rewards. Note that the reward receipt address is independent of the node performing the associated mining.
The amount staked is labeled a proposal stake . Normally, the proposal stake is reclaimed after the proposal is accepted or rejected, but if the proposal is determined to be dishonest, the stake is forfeited.
4、 Transaction Speed
In Konsensus, there is no concept of post-creation confirmation. Blocks are irreversible once created. Compare this to Ethereum, where 6 or more 15-second block cycles may need to pass before a block is considered non-reversible.
We are actively investigating optimization of Konsensus, with the goal of processing 7,000+ transactions per second with a typical per-transaction processing time of one second. This performance would compete favorably with real-world commercial payment systems such as Visa.
5、 Reduced Energy Usage
Because the Konsensus is designed to avoid the need for the expensive search for hash solutions, there is no incentive for miners to use powerful, energy-wasting mining hardware to outperform their fellow miners. This approach provides security by incentivizing the distribution of blockchain validation-by-consensus across many independent parties (because only active miners receive validation rewards) but should use very little electricity compared to typical cryptocurrency consensus mechanisms. For example, one analyst estimates that there are between 5,000 and 100,000 Bitcoin miners who collectively consume approximately 774 megawatts of electricity [7]. By comparison, we estimate that 100,000 miners could run kUSD client 7 software on low-power machines and consume less than 5 megawatts of electricity.
6、 Continuous Agent-Based Modeling
We have created a sophisticated agent-based software model to test the simultaneous use of all three mechanisms that form the Kowala Protocol. We have run a significant number of simulations against various permutations of the model, including:
○ variations in constants used by the Minting Algorithm
○ variations in constants used to define the Stability Fee
○ variations in starting conditions
○ market demand fluctuations, including mass panics
○ rapidly increasing and decreasing numbers of participants
○ excessive optimism and pessimism of arbitrageurs and prospectors
The algorithms and constants described in this white paper reflect the results of our testing, and the Kowala Protocol incorporates only the behaviors which consistently and reproducibly yielded the best results in our models.
Future agent-based modeling will be based on actual historical market data. Kowala will regularly create large-scale simulations of established marketplace behavior and conduct predictive research into further refinements to the Kowala Protocol. These enhanced simulations will feature agents whose behavior is derived from genetic algorithms informed by actual, historic market decisions. Kowala will also model potential attacks by malicious actors in order to preempt them and to uncover other unforeseen vulnerabilities.
7、 Current Research Area: PID Controllers
Our approach of continuously incorporating market feedback into our coin-supply adjustments is broadly reminiscent of the concept of a PID-controller, a sophisticated engineering model of goal-seeking action coupled with continuous feedback from a sensor. We are currently investigating the possibility of improving our mechanisms by expressing them as PID-controllers.
8、 Current Research Area: Artificial Intelligence
We are currently investigating the incorporation of Artificial Intelligence into the Kowala Protocol in two areas. First, we are investigating the use of deep learning to improve the responsiveness of the minting and burning rates of the stability mechanisms. Second, we are investigating the use of adversarial AIs that would make highly sophisticated attacks against the network and exploits of the stability mechanisms, and also simulate complex economic challenges. The results of these experiments will inform future development, especially in terms of security and overall resilience in the face of adversity.
Conclusion
This white paper has identified the problem of volatility in cryptocurrency and proposed the Kowala Protocol as a robust solution to this problem. Although we have established through extensive modelling that the protocol worked in many anticipated scenarios, more work is needed to demonstrate with higher certainty that the specific mechanisms described here will work in a real-world market. We invite others to contribute to improving the Kowala Protocol by visiting https://kowala.tech and to participate in the development of the kUSD client software located at https://github.com/kowala-tech/kcoin .
If you have an exchange about blockchain learning, you can contact me in the following ways:
Add me to WeChat, note the note: Blockchain learning
Micro signal: elninowang
github: https://github.com/omnigeeker
telegram:
twitter:
medium: https://medium.com/@elninowang_cn
steemit: https://steemit.com/@elninowang
zhihu: https://www.zhihu.com/people/elninowang/activities
jianshu: https://www.jianshu.com/u/103321b574ba
