Project: Name of Project
Rocketpool Mechanistic Modeling Research: Universal Variable Commission
What is the work being proposed?
We have a great number of suggestions for how to improve the Rocketpool protocol, but an immature way of analyzing the effect that these suggestions would have on the network. Arguments on LEB4 or RPIP30 tend to focus on the qualitative rather than the quantitative, and what quantitative analyses exist are few and far between as they require considerable work from a member of the community. This leads to uninformed choices by the community in which emotion is a large component. The community needs hard numbers to make informed choices.
Mechanistic modeling allows us to gain additional quantitative insights into expected results from a proposed change. It could find a loophole that we missed or it could give us additional confidence that the expected behavior is as predicted. Other use cases include testing the viability of attack strategies, optimizing settings, or comparing multiple implementations.
The proposed topic of this round of mechanistic modeling is for Universal Variable Commission. The deliverables would include:
- The release and further development of an open-source mechanistic model (written in python) that can be used to gain insights into rocketpool dynamics
- The usage of this tool to analyze up to four implementations of universal variable commission. The comparison will include both how well it is able to balance pETH and NOs as well as how much it would cost in gas. Implementions include:
- UVC by commission changes voted for by pDAO
- A PID implementation designed by sckuzzle
- Up to two more ideas solicited from and put forward by the community
- A report that summarizes the problem UVC aims to solve and what would be expected from various solutions
- If desired, a call that explains the findings and answers any questions
The result of these will lead to a community that has more comprehensive information, making more informed choices and better guiding the direction of Rocketpool.
Alternate Topics
I view UVC as one of the more pressing needs of Rocketpool at present, which is why it is the proposed topic. If the GMC is interested in research on other topics, some other options include:
- MEV Theft
- RPL Tokenomics Changes (like RPL-as-rent / direct capture)
- Any of the other ideas submitted in the options draft
Is there any related work this builds off of?
Not for UVC (other than the RP sim I’ve already built).
Will the results of this project be entirely open source (MIT, GPL, Apache, CC BY license or similar)? If not, which parts will not be, why, and under what license will they be published?
Yes. The simulation framework, analysis files, and the report will all be under GPL.
Benefits - enter N/A where appropriate
This work will help ensure a timely and efficient implementation of Universal Variable Commission. The following are the benefits of UVC:
| Group | Benefits |
|---|---|
| Potential rETH holders | An efficient implementation of UVC will help to ensure that there is room in the deposit pool for new rETH to be minted. |
| rETH holders | When there are an excess of node operators, rETH APR will increase over time. |
| Potential NOs | UVC will encourage both that the deposit pool is not empty (new minipools can be spun up) and also that the commission for nodes is as high as it can be. |
| NOs | When there is excess demand for rETH, the commission for NOs will increase. |
| Community | The primary purpose of this work is to inform and guide the community on potential changes to the protocol. Information on what the optimal strategy is and making sure incentives are aligned helps ensure an efficient protocol. This includes the selection of both what changes to put into an RPIP as well as voting on RPIPs. Both the analysis of UVC implementations and simulations showing what will likely occur should help the community and dev team to more rapidly enact beneficial changes to the protocol. |
| RPL holders | As the rocketpool protocol grows, RPL will become more valuable. |
What other non-RPL protocols, DAOs, projects, or individuals, would stand to benefit from this grant?
None will directly benefit, but the algorithms and analysis may help to inform other protocols how to implement their features.
Team
Who is doing the work?
Primarily sckuzzle.
Community members (such as Valdorff) have been helpful in analyzing results and suggesting scenarios to look at. I am sure that community members will be instrumental in giving technical review / comments, and have suggested a monetary reward to be made available to those reviewing the report. Collaborators and requests for analysis are welcome. I most frequently talk about the status in the Modeling Thread which is open to anyone in the rocketpool discord.
What is the background of the person(s) doing the work? What experience do they have with such projects in the past?
I build simulations and models in biotech. While some domain knowledge (cell metabolism, reaction kinetics, etc.) is not useful here, how to build a simulation, what assumptions to make, and how to perform analyses is very transferable. Other knowledge such as control theory and algorithm design are useful when designing and modeling universal variable commission.
I have been involved with rocketpool for slightly over half a year. I am a node operator and have primarily been joining discussions in #research and how rocketpool should function.
What is the breakdown of the proposed work, in terms of milestones and/or deadlines?
Each scenario could be roughly divided into:
- Initial Modeling (2-3 weeks)
- Report Creation (1 week) – Rough draft report released to community members
- Technical Review from community (2 weeks)
- Edits / additions in response to community (1-3 weeks) – Final draft released on forum, discord
For a total of 6-9 weeks for UVC.
Time taken for other topics will vary heavily depending on the work to be done.
How is the work being tested? Is testing included in the schedule?
The simulation framework has already been created and was tested by comparing RPIP30 reward curves as well as topup strategies. It will be expanded to include proposed universal variable commission implementation and tested to match expected behavior.
How will the work be maintained after delivery?
Since the work is for an analysis, maintenance is not a priority. The simulation frame will be updated whenever there is additional modeling work funded. I’m also usually happy to answer questions and discuss implications the report doesn’t cover.
Backward compatibility in the framework (such as when RP mechanics get updated) is not included both to save cost and because it us unlikely to ever be used.
Payment and Verification
What is the acceptance criteria?
For the model:
- Model accurately predicts known outcomes
- Sckuzzle’s PID implementation demonstrates stability in all scenarios while maintaining responsiveness
- Model is able to track gas-consuming operations
For the work:
- A comparison of multiple UVC implementations with relevant metrics such as gas usage and responsiveness (how much time the DP spends full, max queue size for minipools, etc.)
- If found, a description of potential vulnerabilities to attackers or network conditions that would harm RP
- A final report that addresses reviewers feedback and requests for additional analysis (if in scope)
What is the proposed payment schedule for the grant? How much USD $ and over what period of time is the applicant requesting?
| Item | Timeline | Amount |
|---|---|---|
| Use and release of existing model in open-source | Immediately on grant acceptance | $8k |
| Release of rough draft | ~1 month after work starts | $14k |
| Release of final report | ~2 months after work starts | $16k |
I also think funding 3x $200 incentives for reviewers would ensure high-quality feedback and critical review.
How will the GMC verify that the work delivered matches the proposed cadence?
There will be periodic modeling results in the modeling thread along with discussion. The rough draft will be made available to community members (which the GMC may include themselves in) and the final report will be released on discord and posted to the forum.
What alternatives or options have been considered in order to save costs for the proposed project?
The report, what detail to include, and the target audience would affect the cost. My view is that the report should include additional background to be applicable to more of the community.
I also considered the effort that would be required to make the simulation framework ready for release as a package. The documentation, unit testing, and integration testing required for what would likely not see significant use was discarded to lower cost.
Other components of the research could be excluded, such as alternate implementations, attack testing,
Conflict of Interest
Does the person or persons proposing the grant have any conflicts of interest to disclose? (Please disclose here if you are a member of the GMC or if any member of the GMC would benefit directly financially from the grant).
None
Will the recipient of the grant, or any protocol or project in which the recipient has a vested interest (other than Rocket Pool), benefit financially if the grant is successful?```
No