Background:

FIP-0070: Allow SPs to move partitions between deadlines is currently in scope for inclusion in the v21 Watermelon upgrade proposed to land on mainnet Nov 29.

However, a bug has been identified that may cause inconsistencies in the expiration epoch handling for recovered sectors. This bug raises concerns about sector lifetimes and power crediting, potentially leading to network vulnerabilities and fairness issues.

This matrix ONLY deals with options available in terms of whether or not to include FIP0070 as part of nv21 Watermelon upgrade. It does not delve into the options available in terms of fixing this bug.

Decision Matrix

Decision Option Governance risk/consideration Resource Requirements (Governance/FF) Technical Resource Requirements Timeline Impact Alignment with Strategic Goals Risks Recommendation
Option 1: Keep FIP0070 in-scope for nv21 This will maintain the existing scope of nv21 as already widely communicated with community members. Low - Nothing changes in terms of governance. FF Governance will provide other support as needed to communicate possible delays. High- this will require more time commitment to bug fixing and, full testing cycle. Fixing this will most likely mean further delays to landing nv21 and high likelihood of moving the upgrade to early 2024 after holidays are over. Partial - Risks outweigh benefits Likely hard to estimate the scope of fixing the bug. This uncertainty means a lack of clarity as to when nv21 might land on mainnet. In addition, to meet the timeline, it might require additional patching which is not recommended. Not recommended
Option 2: Descope FIP-0070 from nv21 Upon agreement by implementation teams for this FIP to be descoped, this FIP will remain as “Accepted” and no further action needs to be taken as such in terms of its status. High - Community members will have to be kept informed through various comms means. FF Governance will coordinate with comms team at FF to coordinate adequate messaging. All public documentation will have to be updated. Simple - avoids immediate technical debt. Will allow more time for implementation teams to investigate and prepare FIP for inclusion in future upgrades. No delay to nv21. Strong - Focuses on stability and predictability Hard to predict community sentiments regarding this option possibly from SPs Recommended
Option 3: Delay nv21 beyond Q4 2023 to include FIP-0070 Comprehensive solution to bug, and maintain scope. No governance implication as such High - Detailed & comprehensive comms and stakeholder liaison to effectively communicate this delay. Coordinate with implementation teams to understand and agree on timing Medium - depends on bug complexity nv21 will likely not be feasible in 2023. Upgrade occurring early Q1 2024 will further push out already pre-scheduled upgrade and timelines communicated to implementation teams https://github.com/filecoin-project/core-devs/discussions/155. Weak - Delays can delay network growth which would have been introduced through other FIPs and improvements. This will possibly not be perceived well by community stakeholders and could signal overall network instability. This option will delay other improvements and critical features that should have landed in pre-scheduled upgrades. There is also the risk of scope creep and the likelihood of considering adding more FIPs to the upgrade thereby stretching limited resources. Consider with caution
Option 4: