|
| 1 | +--- |
| 2 | +title: Weekly Summary – June 3, 2025 |
| 3 | +authors: |
| 4 | +- will |
| 5 | +tags: [progress, update, weekly, ci, formal-methods, pseudo-mainnet, topology, rust-simulation, conflicts, incentives, mainnet-scale] |
| 6 | +--- |
| 7 | + |
| 8 | +This week, the Leios team focused on infrastructure improvements, formal methods advancement, and large-scale network simulation. The team successfully resolved outstanding CI issues, enhanced the formal specification with Full-Short Leios support, and began simulating a realistic 10,000-node pseudo-mainnet topology. |
| 9 | + |
| 10 | +## Infrastructure improvements |
| 11 | + |
| 12 | +- Fixed outstanding CI bugs [#368](https://github.com/input-output-hk/ouroboros-leios/issues/368) and [#379](https://github.com/input-output-hk/ouroboros-leios/issues/379), enabling all CI checks to pass. |
| 13 | + |
| 14 | +## Formal methods advancement |
| 15 | + |
| 16 | +- Added Full-Short Leios as a special case of Short Leios to the [formal specification](https://github.com/input-output-hk/ouroboros-leios-formal-spec/tree/yveshauser/full-short-leios) |
| 17 | +- Implemented trace verification capabilities for Full-Short Leios. |
| 18 | + |
| 19 | +## Pseudo-mainnet topology simulation |
| 20 | + |
| 21 | +- Designed and initiated comprehensive simulations for 1 through 300 TPS using the new [pseudo-mainnet topology](https://github.com/input-output-hk/ouroboros-leios/blob/main/data/simulation/pseudo-mainnet/) |
| 22 | +- Created a realistic 10,000-node network with: |
| 23 | + - 2,657 block producers and 7,343 relay nodes |
| 24 | + - Realistic stake distribution and geographic distribution |
| 25 | + - Two relays per block producer with realistic latencies |
| 26 | + - 298,756 total connections with 6-hop network diameter |
| 27 | +- Observed significant performance challenges with large-scale simulation: |
| 28 | + - Rust simulation: 6 minutes of network time in 10 hours at 1 TPS |
| 29 | + - Performance degradation at higher TPS rates (1 minute network time in 10 hours at 300 TPS) |
| 30 | + - Haskell simulation requires optimization for practical large-network analysis. |
| 31 | + |
| 32 | +## Rust simulation enhancements |
| 33 | + |
| 34 | +- Implemented random sampling of transactions from the Leios memory pool to ensure different IBs contain different transactions when possible |
| 35 | +- Added simulation support for Leios variant where IBs contain transaction references rather than full transaction bodies |
| 36 | +- Enhanced transaction handling for high-traffic scenarios. |
| 37 | + |
| 38 | +## Analysis of conflicts and incentives |
| 39 | + |
| 40 | +- Completed comprehensive analysis of transaction conflicts, ledger design, and fee incentives |
| 41 | +- Key findings on conflict management: |
| 42 | + - Honest duplicates and conflicts are unavoidable with local sortition |
| 43 | + - Memory pool rules can minimize conflicts through prompt transaction removal |
| 44 | + - Collateral requirements for failed transactions conflict with Cardano's guarantees |
| 45 | +- Identified block producer compensation strategies for handling conflicting transactions |
| 46 | +- Proposed EB-level optimization through bitmap-based transaction validation to reduce persistent storage of duplicates and conflicts. |
| 47 | + |
0 commit comments