-
Notifications
You must be signed in to change notification settings - Fork 1
Open
Description
Goal
Write a paper about WT wallet management with A1CP method of fee sponsorship. Present the model (on paper and through code) of WT and its environment. Simulate several case studies.
- 4 person stk-man hedge fund
- 5 stks 2 mans finance team of a BTC trust company
- ...?
Deliverables
- Design, formalisation, and analysis of WT wallet management algorithms
- Quantification of operational costs for provision of security by a WT
- Quantification of WT wallet risks
- Quantification of stakeholder activity
Tasks (incomplete)
High level
- Finish first pass of formal model for WT wallet & environment (Jacob by 23/08)
- Finish review and feedback of WTSM and WTSim (Antoine by 23/08)
- Finish Algorithm Design section (Jacob by 28/08)
- Settle on results to generate (by 28/08)
- Complete first draft (by 02/09)
- Edit and final touches (by 08/09)
- Submission & Code release (by 09/09)
Data collection
- Historic fees per block #15
- Update estimateSmartFee historic data (low priority)
Code
- Use estimateSmartFee for Cancel tx & CF Tx
- Use integers for all values with units in satoshis
- Add random time interval between balance low & re-fill trigger (to simulate slow stakeholder), to investigate time-at-risk.
- Handle
V_m = 0
for early blocks - Replace feerate data src with own node info (indexed by block number)
- Remove possibility for inconsistencies in progress of blocks with WTSim
- Compute & plot the amount wasted by overpaying fees due to inadequate availability of coins
- Track all operational information for initial deposits & actions from WT
- Optimise run time (to speed up generating & interpreting results)
- Spends/cancels that use multiple vaults
- Algorithm review
- General line-by-line review
- Consider ordering of outputs in fee generator for CF Tx
- Reconsider all failure modes (then compare with enumeration in paper) and recoverable errors
- Create paper-ready template for plotted results (fig size, text & tick sizing, sim params)
- Tests module?
- Prepare code for release (PEP8 formatting, documentation, etc.)
Paper
- Introduction
- What is Revault?
- Summary of Revault WT
- Related Work #20
- Objectives
- Paper organisation
- Methodology & Scope
- Scope: Specify analysis for 1 WT in operational phase
- Methodology: rationale for choosing LTS modelling language
- WT Wallet model
- Why A1CP? Why prepare coin pool? Why not CPFP? #16
- How / when to feebump Feebump algorithm #3
- Spend/cancel transitions that use multiple vaults
- WT environment model
- Formalise operation (event timing, deployment specific parameters)
- Fix notation (if there is some "canonical" LTS notation)
- Transition sequences with batched spend attempts
- Algorithm Design
- Justify design choices
- Determine trade-offs
- Determine constraints
- Results
- What to use as case studies? (n_stk, n_man, time-lock, expected number of vaults & spends, etc.) #17
- Comparisson of algorithm versions (which algorithms?)
- Interpret results
- Format paper for FC'22 Submission
Questions to study
- Extending results to system of watchtowers (either off-hand, or formally)
- Is it necessary for WT to output risk status info to stk (to make better decisions about re-filling)?
- Is it necessary to simulate feebump requirement? If there is enough in the reserve, WT will pay with amount at least as large as conservative estimateSmartFee value with target 1 block. Enumerate why this could fail.
- Compute the amount wasted by overpaying fees due to inadequate availability of coins
- How to notice and exploit the low-fee periods?
Metadata
Metadata
Assignees
Labels
No labels