Please describe your proposed solution
Cardano seeks to achieve the most decentralized, secure blockchain network as the foundation for an antifragile financial operating system. To achieve this goal, it is essential that the evolution of the Ouroboros protocol be designed and developed alongside a suite of rigorous stress tests and early warning detection systems. Ouroboros Praos established Cardano as a leader in secure decentralized proof-of-stake. The protocol continues to evolve with the anticipated release of Ouroboros Leios which will support Input Endorsers and enhanced optimization of transaction efficiency.
Photrek’s first stress test and detection system will focus on the ability of rogue actors to drain value from the network through manipulation of block production timing. If a multipool or a consortium of SPOs were to share slot leader selection information, that knowledge could be used to strategically withhold block production for the purpose of creating blockchain segments controlled by these Byzantine actors. Figure 1 shows the global block production predictions and experimental measurements. Collusion is difficult to detect because it can “hide” in the inherent uncertainty of Nakamoto proof-of-stake and result in smaller pools failing to mint their block assignments to the broader network. Although this type of event is expected to be a rare event, systematic exploitation could drain a substantial amount of value out of the network overtime. Photrek will quantify the real-world vulnerability, compare that with design analysis, and develop detectors for early warning.
Figure 1: Block Production
The Cardano community needs tools and educational materials to locally monitor the accuracy of the slot leader probabilities. Local alerts are needed when the accuracy of the probabilities drops below historic trends. Furthermore, analytical tools are necessary to identify which cluster of nodes are attempting a Byzantine attack. Furthermore an understanding of how upgrades to the Ouroboros protocol impact these vulnerabilities is needed.
Photrek and its partners will develop a Stress Test and Early Warning System that provides Cardano with
- The ability to simulate vulnerabilities in the timing of block production on testnets;
- A histogram of the probability forecasts of the selected slot leader;
- A visualization comparing the delegation ratio forecasts with the histogram of the actual percentages of slot leader selection;
- Accuracy, Robustness and Decisiveness metrics regarding slot selection;
- Educational resources to enhance knowledge within the SPO community regarding Byzantine attacks and how the SPO Health Monitor can mitigate the threat.
Figure 2 shows an example of a Risk Monitor plot, which will be the first tool developed for the Early Warning Detection System. This example was of an image generator algorithm. The histogram can show the distributional character of the forecasts. The Accuracy shows the central-tendency on a log scale; the Robustness shows the outlier performance; and the Decisiveness shows the ability to forecast the correct class (slot leader).
Cardano protocol developers and SPOs will be able to utilize the Risk Monitor to actively track the accuracy and robustness of the slot leader forecasts. This information can provide an early warning if slot leader selection and thus block production is being manipulated via block withholding or similar network distribution based attacks. The analysis will be completed across the network nodes and for individual nodes. By developing historical records of the Robustness-Accuracy-Decisiveness (R-A-D) metrics, alerts based on significant deviations from historical trends can be created. These alerts can for instance be used to determine whether multipool operators are withholding release of blocks in effort to increase their probability of slot leader selection.
Figure 2: Risk Monitor
<u>How we perceive the problem we are solving:</u>
The problem at hand involves ensuring accountability among SPOs based on the honest behavior outlined by the theoretical predictions in the Ouroboros Praos paper. This involves enabling real-time analysis of each party's data dissemination across the network. The introduction of a numerical risk model would empower developers and SPOs to conduct an empirical analysis of network health, thereby swiftly identifying any Byzantine behavior.
- Another concern that needs addressing is the possibility of multiple pool operators leveraging their size to impact the network negatively. An example of this issue can be found in an article on <https://adapulse.io/multiple-stakepool-operators-are-harming-cardano/>
- Furthermore, an enhancement we are considering is the provision of an "active" local calculation mechanism for block creation probability. This feature would greatly aid in monitoring the performance of the network.
<u>Our reasons for approaching it in the way that we have:</u>
- This approach takes into account real world network behavior instead of theoretical estimates. Nakamoto proof-of-stake allows for a larger validator set (due to being more bandwidth efficient) at the cost of short-term statistical uncertainty. This probabilistic approach coupled with a dynamically available distributed network provides a path for dishonest SPOs to accrue greater than their fair share of rewards.
- Such an effect would be small and would be more easily found by having multiple SPOs analyzing the network concurrently, in real-time.
<u>Who our project will engage:</u>
- The Cardano developer and SPO community will be the primary target for engagement within this project. The team plans to connect with the community through dedicated chat groups on Telegram and Discord, as well as by hosting breakout rooms. The team will introduce the proposed solution and reach out to developers and SPOs who express interest, with the aim of collaborating on testing and validation of new protocol designs and ongoing operation of the network.
<u>How we will demonstrate or prove our impact:</u>
- The impact of this solution will be evidenced through several steps:
- A complete testnet experiment to gauge the impact on probability accuracy. This will provide an understanding of how accurate the solution works in real-world conditions.
- A demonstration of the sensitivity of the model to the accumulation of stake. This will provide insights on the model’s ability to handle changes in stake distribution and their impact on the network.
<u>Explanation of our unique solution, who will benefit, and why this should be important to Cardano: </u>
- Our solution is unique because of its localized approach to pool performance evaluation. Ouroboros is unique among proof of stake blockchains by virtue of its probabilistic nature. Local calculation of the relevant statistics for a pool are vital to understand pool performance.
- Complex model due to a combination of network topology, game theory of information propagation, and leader election process is well suited to a (multi-vectored) risk analysis.
- This comprehensive approach will benefit all participating in the network, particularly SPOs by providing a more robust and transparent understanding of network health and performance. Underscoring why this should be of significant importance to the Cardano community.