NOTE: THIS PLANNING DOC IS DEPRECATED AND IS NOT ACCURATE AT THIS TIME

The following table is a summary of identified projects with the highest impact in the shortest amounts of time

Project Impact Timeframe Resources Parallel effort? Priority (TBD)
[Q1] Grindability of WindowPoSt challenges in different partitions (https://pl-strflt.notion.site/Q1-Grindability-of-WindowPoSt-challenges-in-different-partitions-b0e951e643d64f31bb170e51f17747bf) There are open security implications, but this work was historically not prioritized. CNL urges that it could already be exploited today and should be patched. Requires mandatory network upgrade. 2 weeks to code ready; lotus timing and integration TBD nemo (with team code review when ready) Work can be done in parallel P1(?)
;[Q1] Synthetic PoRep (a particular form of NI-PoRep) (https://pl-strflt.notion.site/Q1-Synthetic-PoRep-a-particular-form-of-NI-PoRep-e99bb5a6f5b2466fa812ba0fd8333112) Allows miners to free up disk space sooner after PC1, with minimal impact to current pipelines. This would be an optional feature accepted by the protocol, therefore an optional network upgrade. 2-4 weeks to code ready; lotus timing and integration TBD nemo and Jake (with team code review when ready) Work can be done in parallel P1(?)
Evaluating SupraNational Improvements SN has made exceptional performance claims related to data onboarding that have yet to be verified by FilDev. We would like to gain access to necessary code and configurations to verify this and push to release this to the community asap. This does not require a mandatory network upgrade. Highly depends on SN cooperation and assisting with Lotus integration; TBD, but we are optimistically estimating 3-8 weeks. Lotus team, FilCrypto, SupraNational Work can be done in parallel P0
De-risking Testudo If SN improvements can be integrated, it’s not clear what the impact here is. Otherwise, Testudo can improve Proofs performance significantly over time and eliminates the need for future trusted setups after the initial one (which has a positive security implication). De-risking does not require a mandatory network upgrade, but paves the path for one in the future. CNL estimates that this collaboration will take 10-20% dedicated time from some FilCrypto resources alongside their team allocations. FilCrypto estimates 20-25% dedicated time. Volker and Jake at~ 20-25% time; CNL resources (nikkolasg / TBD) Work can be done in parallel P1(?)
Inactive CC-Sectors CNL has proposed that this ability can be mfairassive, particularly with regard to the sales of sectors from SaaS providers. It is likely this will be a mandatory upgrade for sector acceptance. FilCrypto does not have full details, but we are told this could be ship-ready on the 2-3 week timeframe; lotus timing and integration TBD Nicola, nemo and Jake (with team code review when ready) Work can be done in parallel P0
[Q1] Defaulting Proofs GPU support to CUDA instead of OpenCL (https://pl-strflt.notion.site/Q1-Defaulting-Proofs-GPU-support-to-CUDA-instead-of-OpenCL-1fbaf744fa79489bb16052960932dd08) Most SPs are already running CUDA as it’s more performant than the default (OpenCL). This would simplify SP deployments as it would be enabled by default. ~1-2 weeks or less to ensure it’s plumbed through the stack (to FFI / Lotus) and tested; lotus timing and integration TBD Volker and nemo (with team code review when ready) Work can be done in parallel P1(?)

Previously identified items [Older, see above for latest]

The following are tasks that we believe can be completed in either the Q1 or Q2 timeframe (i.e. code ready and tested internally; ready to coordinate with Lotus on testnet/shipping windows). We are in the process of gaining feedback in order to prioritize them.

The following are known tasks, with a concrete timeline being unknown — however, we believe that neither task will extend beyond being ‘ship ready’ at the end of Q2; again with all due respect in finding an appropriate ship window with Lotus team.