FIL-RetroPGF-1: Nominations for Protocol Research & Development

[kkarrancsu]

Protocol Research & Development

This is the thread to nominate projects for the Protocol Research & Development category of FIL-RetroPGF-1. As a reminder, this category is intended to capture work to improve the Filecoin protocol, to include both theoretical contributions as well as implementations and prototypes are in scope.

Nomination Template

When nominating a project, please answer the following questions in a comment to this discussion:

1. The project name
2. A description of the positive impact of this project for this specific category over the past six months (Oct 2023 - Mar 2024)
3. A link to the project’s GitHub or Twitter
4. Contact info for the project or project lead

Nomination Considerations

Things to consider when nominating a project:

• Use your best effort to describe how this project has delivered impact to the Filecoin ecosystem in the previous six months (Oct 2023 - Mar 2024). Use the template provided in each category’s thread when nominating a project.
• Make sure to nominate projects in the relevant category. If you feel that a project spans multiple categories, please choose the one category that you feel it best aligns with. Do not nominate the same project across multiple categories!
• Note that the organizers of FIL-RetroPGF-1 may choose to reassign a project’s category from its nominated one, if they feel it better aligns elsewhere.
• Before nominating a project please check if this project has been nominated already, so we can avoid duplication.
  RetroPGF is an experiment that Filecoin plans to iterate on regularly. Community feedback is key to helping the process improve with each round. We welcome your notes, reflections, and feedback at any step. Please use this discussion to submit feedback!

Project Nominations will close by Mar 31, 2024.

See here for general guidelines on project nomination.

Note: These threads will be moderated to only include project nominations. Off-topic comments that show support or hype for a project or meta-commentary on the program itself will be removed to keep the thread clean. If you want to show your excitement for a nomination, use the reaction feature!

[jnthnvctr]

The project name:
Banyan

A description of the positive impact of this project for this specific category over the past six months (Oct 2023 - Mar 2024)

• Banyan is one of the early teams seriously pushing enterprise adoption in the Filecoin ecosystem. Notable in their approach has been the open sourcing of their own research to the community: which has provided critical early feedback to core devs on gaps in the protocol, helped shape prioritization (to enable Filecoin to become a more critical component of an enterprise offering), and reduced the barrier of entry for new start ups looking to build enterprise offerings in the space). Example: https://www.youtube.com/watch?v=uU96bIyruwo
• Banyan has also been an active contributor to the creation of specs and standards to help enable composability between collab-apetitive teams: /~https://github.com/banyancomputer/on-chain-compliance-attestations

A link to the project’s GitHub or Twitter
/~https://github.com/banyancomputer

Contact info for the project or project lead
c@banyan.computer

[Joss-Hua]

The project name:
FIP-0021, FIP-0029

A description of the positive impact of this project for this specific category over the past six months (Oct 2023 - Mar 2024)

• As is well known, "Secter quality" is the base of the Filecoin network consensus, and accurate values are a hallmark of the core contribution of SP in the network;
• To provide more flexibilities for Filecoin finance market, without this implementation, there would be no current lending market(DEFI);

A link to the project’s GitHub or Twitter
/~https://github.com/filecoin-project/FIPs?tab=readme-ov-file#fips

• /~https://github.com/filecoin-project/FIPs/blob/master/FIPS/fip-0021.md
• /~https://github.com/filecoin-project/FIPs/blob/master/FIPS/fip-0029.md

Contact info for the project or project lead
venus@ipfsforce.com

[marco-storswift]

Project:
StorSwift

Description of impact in last 6 months:

1 StorSwift provides SP with an MPC-based wallet based on lotus repo, which improves the convenience of network users using the wallet, lowers the usage threshold, and improves the security of the wallet.
2 The wallet API based on lotus provides the offline sign function to ensure that the user's private key is not stored in the production environment and ensures the security of the user's large assets.
3 The team optimized the gas overhead of buildin Actor, enriched and improved the availability of the network.
A link to the Github / Twitter:
mpc ：/~https://github.com/storswiftlabs/imfil branchonline-v3-mpc
offline-sign:/~https://github.com/storswift/lotus-offline-sign
network optimized:
filecoin-project/lotus#8696
filecoin-project/lotus#8854
filecoin-project/lotus#8545
filecoin-project/go-state-types#125
filecoin-project/builtin-actors#722
filecoin-project/lotus#11063
filecoin-project/lotus#10928

Contact:
mfkrgxa@gmail.com marco@storswift.com

[jsoares]

The project name

EC Finality Calculator

A description of the positive impact of this project for this specific category over the past six months (Oct 2023 - Mar 2024)

Filecoin's Expected Consensus comes with probabilistic finality and a 900-epoch soft finality threshold. Up until the release of the EC Finality Calculator in Jan 2024, there was no rigorous way to determine the reversal probability of a given tipset/chain. Therefore, network participants like exchanges and L2 operators had no other option but to enforce a worst-case wait time of 100-900 epochs (i.e. hours) before clearing transactions or releasing funds.

FRC-0089 introduces a new, entirely client-side method that allows participants to compute finality thresholds in view of the actual chain history. Under normal operating conditions, it brings down the confirmation from hours to ~15 minutes while maintaining the same level of security, greatly improving the UX for applications and for the overall network. The FRC specificies the calculator and is accompanied by a prototype stand-alone implementation that can be used today by any interested party.

A link to the project’s GitHub or Twitter
/~https://github.com/filecoin-project/FIPs/blob/master/FRCs/frc-0089.md
/~https://github.com/consensus-shipyard/ec-finality-calculator

Contact info for the project or project lead
jorge@protocol.ai (@jsoares), guy.goren@protocol.ai (@guy.goren)

[jsoares]

Some notes:

• This could fit under tooling too, but this seemed like the right place; please reassign if needed
• I'm self-nominating as one of the project members
• The project members are employed by PL but this has been a sort of "20% project" for us, outside of our core duties
• While the calculator is ready and usable today, PL isn't funding an out-of-the-box implementation in Filecoin clients, which would greatly increase usability, reach, and impact; we hope to use any rewards for the past work to fund a community bounty for said implementation

[ansermino]

The project name

Forest

A description of the positive impact of this project for this specific category over the past six months (Oct 2023 - Mar 2024)

• Highly performant Filecoin client running on Mainnet and testnets.
• The only non-Go Filecoin node in the ecosystem, written in Rust 🦀, directly interfacing with the FVM without FFI bindings.
• Supported all network upgrades in the past six months, with performant state migrations that helped bootstrap non-Forest nodes after the upgrade.
• A significant reduction in average resource consumption versus other implementations reduces the node operators' cost. Additionally, operations such as snapshot generation and full archival data exporting are to be done on minimal hardware
  • Able to produce snapshots with 4 core CPU, 10GiB memory; twice as fast as Lotus
  • 50% reduction in the size of archival data exports, with a slower rate of growth over time
• Contributions to network upgrades, including catching critical bugs and bringing sanity to the release process
• Direct cooperation and collaboration with Lotus and Venus maintainers.
• Enabled bootnode capabilities to bring more decentralization to the network

A link to the project’s GitHub or Twitter

github.com/ChainSafe/forest

Contact info for the project or project lead

forest@chainsafe.io

[jennijuju]

It was really great to see all the progress Forest has made in the past year, including starting to be one of the clients backing up the Fileocin calibration testnet, bootstrap nodes, and snapshot services! And working with forest devs on upgrade tests is super fun!

[patrickwoodhead]

The project name

Spark Protocol

A description of the positive impact of this project for this specific category over the past six months (Oct 2023 - Mar 2024)

Spark samples file retrievals from Filecoin Storage Providers to improve the health and performance of the Filecoin Retrieval Market.

Between October 2023 and March 2024, the number of (Station) nodes running Spark requests went from essentially 0 to ranging between 10k and 20k daily. Spark is now making around 20m retrieval requests a day to Filecoin Storage Providers and it is requesting roughly half a million different file CIDs each week.

In March, the Spark protocol added the ability to calculate retrieval success rate by Storage Provider, by linking up keys in IPNI and in on-chain records. This means that the Spark Protocol can now understand the retrieval performance of each Storage Provider at a file retrieval level.

N.B. Spark is making file CID retrievals not piece CID retrievals.

A link to the project’s GitHub or Twitter

• filspark.com
• /~https://github.com/filecoin-station/spark

Contact info for the project or project lead

spark@meridian.space

[momack2]

@patrickwoodhead is there a dashboard or leaderboard for SP success rate now?

[patrickwoodhead]

There is a public dashboard but it has not yet been migrated into the Space Meridian domain yet. We will share as soon as it is ready. I will DM you the current link.

[nicola]

Project name: Future PoRep Research

A description of the positive impact of this project for this specific category over the past six months (Oct 2023 - Mar 2024)

Description

• The New PoRep Research WG has been researching a new PoRep and to ensure that it is secure
• In 2023/24 we rebooted research effort for designing a new proof of replication that could substitute SDR with the primary goal of achieving fast retrieval and secondary goal of achieving simpler proof system with lower hardware requirements.
• Outcomes from this work:
  • First, we de-risked the SDR to be secure in the cost model (after the Supranational improvement, the security margin was thought to be 2x - which would make SDR with current parameters only secure in the cost model for a few more years). We showed that there are tighter bounds which lead to a more concrete cost calculation, showing that the current security margin is instead 168x (this is a 84x improvement!). This means not only that SDR is secure for many years, but also that there is a large margin of improvement for new PoReps.
    • Better cost analysis via adding costs for storage https://www.notion.so/pl-strflt/Better-Cost-Analysis-via-Adding-Costs-of-Storage-and-Computation-532e034e9389412099368ac0da22bf8c
  • Second, we wrote the first summarized paper about SDR and designed SPR, a tighter construction with better bounds, which could be a stepping stone for future PoRep
    • Proofs of Space with Maximal Hardness https://eprint.iacr.org/2023/1530
  • Third, we produced several design documents for potential constructions
    • Why Better Predecessor Robustness or Depth Robustness are crucial: https://www.notion.so/pl-strflt/Why-Better-Predecessor-Robustness-or-Depth-Robustness-are-Crucial-3237d78ed700402abb774d2628024cae?pvs=4
    • Unsealing vs Storing a second copy https://www.notion.so/pl-strflt/Unsealing-vs-Storing-a-Second-Copy-5f9020e052164765887a5a77cf0fcad2
• People: Leo, Carla, Anca, Matteo with help from Nicola, Irene, Luca
• Link to project:
  • Summary of the work done in 2023/24 https://www.notion.so/pl-strflt/Retrospective-on-PoS-Research-2023-and-Status-0df524fc650a40ab8ec590636fa9f44d

Contact info for the project or project lead

This is on behalf of CryptoNet, whose new re-incarnation will be Public Good Crypto (website coming up soon)

nicola at filoz dot org

[irenegia]

I would like to nominate the “Practical PoRep Improvements WG” which did analysis to understand the pain points for Filecoin Storage Providers and acted on those shipping reports and FIPs that improved the status quo.

• The project name: Practical PoRep Improvements WG
• Impact:
  • SnarkPack (shipped): New sector onboarding method that reduces chain congestion;
  • SnapDeal effort (shipped):
    • A new method for updating a Committed Capacity (CC) sector with data (and deals) without resealing;
    • A blog post (together  with a detailed report and calculator) that helps SPs to understand the different costs of using SnapDeal;
  • Synthetic PoRep (shipped): New PoRep with smaller storage overhead between PreCommit and ProveCommit;
  • NI-PoRep (future work): New PoRep that removes on-chain interaction.
• Links:
  • /~https://github.com/filecoin-project/FIPs/blob/10c12f1585ba0c27a89ed16ab5693253b875587d/FIPS/fip-0013.md
  • /~https://github.com/filecoin-project/FIPs/blob/10c12f1585ba0c27a89ed16ab5693253b875587d/FIPS/fip-0019.md
  • https://filecoin.io/blog/posts/cc-sector-upgrade-guidelines-and-modeling/
  • /~https://github.com/filecoin-project/FIPs/blob/10c12f1585ba0c27a89ed16ab5693253b875587d/FIPS/fip-0059.md
  • /~https://github.com/filecoin-project/FIPs/blob/10c12f1585ba0c27a89ed16ab5693253b875587d/FIPS/fip-0090.md
• Contact info: irene@filoz.org

[Kubuxu]

I would like to nominate the Verifiable Data Aggregation project, which resulted in FRC-0058 standardising Proof of DataSegment Inclusion (PoDSI).

• Project Name: Verifiable Data Aggregation
• Impact:
  Verifiable Data Aggregation enables trust-less onboarding of small pieces of data, an area currently underserved in by Filecoin. Existing storage aggregators integrated it (see Lighthouse), and new applications are built on top of it. It is set to become a basis for L2 data markets, increasing the ease of use of Filecoin and opening new opportunities.
• Links:
  • FRC-0058
  • https://docs.lighthouse.storage/lighthouse-1/filecoin-virtual-machine/section-a
• Contact Info: kubuxu@filoz.org

[Schwartz10]

Project:
Lotus

Description of impact in last 6 months:

• The Lotus team has made countless contributions to the Filecoin core protocol and also maintains the Go Filecoin implementation (Lotus). For GLIF, the protocol research and development has significantly helped to make our defi application possible.
• Lotus has delivered ~15 network upgrades since mainnet liftoff and including implementing, testing and delivering large amount of the FIPs finalized, including but not limiting supporting FVM & FEVM, security fixes, snap deals & etc.

A link to the Github / Twitter:
https://twitter.com/lotus_web3

Contact:
jennifer@filoz.org

[nicola]

Project:
Public Good Crypto (was CryptoNet)

Description of impact in the last 6 months:

• Cryptonet has been a very prolific research and protocol design team in the Filecoin ecosystem
• The protocol part of CryptoNet is going to FilOz, the research part is turning int Public Good Crypto
• This team plans to continue building a large community of researchers and do great research for the success of the Filecoin network (previously worked on SnarkPack, Testudo, Vector Commitments, Proof of Replication)

[gitdata001]

The project name
JiaoZiFS - A version control file system for data lineage & data collaboration.

A description of the positive impact of this project for this specific category over the past six months (Oct 2023 - Mar 2024)

Generative AI is driving increasing interest in storing and making sense of unstructured and semistructured data, along with multimodal data like images, video, and audio.

Soon, previously overlooked data sets like Zoom calls, performance review feedback, customer support tickets, and comments from online forums may provide value—with the corollary that enterprises will want to store them.

In parallel, important signals will emerge from wider domains of the enterprise, beyond traditional centers like product and engineering.

This explosion in the types of data enterprises store, together with the increased heterogeneity of data sources, is another key trend that’s reconfiguring the data infrastructure landscape.

In response to cost pressures and growing data volumes, especially in unstructured and heterogeneous data, enterprises are centralizing their data infrastructure into a data hub.

Long term, we think more and more data will be housed directly in object stores like JiaoZiFS, IPFS, and Filecoin. At the same time, we also foresee data infrastructure becoming more decentralized, with each team within an enterprise operating as a spoke off this data hub.

JiaoZiFS currently supports IPFS as a back-end object storage, you only need to specify this parameter when creating a data repositiory in https://cloud.jiaozifs.com
{"type":"ipfs","ipfs":{"url":"/dns/kubo-service.ipfs.svc.cluster.local/tcp/5001"}}

We deployed IPFS together with JiaoZiFS, you can view these files here:
https://ipfs.jiaozifs.com/ipfs/bafybeidf7cpkwsjkq6xs3r6fbbxghbugilx3jtezbza7gua3k5wjixpmba/#/filesThis also means:
IPFS is used for data storage and data synchronization, and JiaoZiFS is used as a Git-Like data versioning semantic layer.

This way you can use JiaoZiFS as the model registry to checkpoint the model in the training iteration. IPFS is used to synchronize model data sets between core data center and near/far Edge.

Longer term, we believe each spoke will operate as a generative-AI-enabled agent, with the ability to pull data in real-time and answer queries autonomously, all while ensuring proper governance.

A link to the project’s GitHub or Twitter

/~https://github.com/GitDataAI/jiaozifs

Contact info for the project or project lead

info@gitdata.ai

[taoshengshi]

The project name
Mount - An Asset Creation,Issuance and Governance Protocol for Data DAOs

A description of the positive impact of this project for this specific category over the past six months (Oct 2023 - Mar 2024)
https://docs.google.com/document/d/1ROCr5kEWvhweIrSOGqn393BTOD-5YZ1pmxVdfB0d9mY/edit#heading=h.htydihn1kc6i

A link to the project’s GitHub or Twitter
/~https://github.com/datadao-wg/mount-protocol

Contact info for the project or project lead
KEN Labs and taoshengshi01@gmail.com

[momack2]

The project name
The Filecoin Virtual Machine

A description of the positive impact of this project for this specific category over the past six months (Oct 2023 - Mar 2024)

• FIPs 30, 31, 32, 48, 49, 50, 54, 55, 57 (which together implement user programmability via FVM and FEVM)
• The Filecoin Virtual Machine introduced user programmability into the Filecoin network - unlocking a massive community of builders and applications that make Filecoin much more rich, capable, and upgradable than before
• The FVM was the main building block which unlocked DeFI lease/loan markets (unblocking SP growth and access to collateral), L2 networks building on Filecoin, and on-chain storage solutions like Lighthouse, Eastore, and many others which unlock rich storage solutions

A link to the project’s GitHub or Twitter

• /~https://github.com/filecoin-project/ref-fvm
• https://twitter.com/fvmdev
• fvm.filecoin.io

Contact info for the project or project lead
@raulk @trruckerfling @momack2 @Stebalien

[jennijuju]

The project name

Filecoin Protocol Opportunity WG

A description of the positive impact of this project for this specific category over the past six months (Oct 2023 - Mar 2024)

This WG focuses on researching, proposing, reviewing, and auditing new FIPs, from early ideation/research to prototypes of new improvements to the Filecoin protocol, spec, etc.

Collectively they have proposed >30 FIPs, to provide some features adopted or prototyped over the past 6 months:

• PoRep features (FIP0059, FIP0090)
• FVM enhancements (FIP0071, FIP0072, FIP0073, FIP0075)
• User programmability (FIP0076, FIP0083, FIP0062)
• Security and chain scalability (FIP0060, FIP0061)

A link to the project’s GitHub or Twitter

Contact info for the project or project lead

@jennijuju @nicola

[jennijuju]

Note: This WG is not yet "official." The goal is to acknowledge those who actively and consistently participate in Filecoin protocol development and motivate more developers to do the same.
A couple of us took a brief look at the proposed and finalized FIPs and FIP discussion and would like to thank @anorth, @ZenGround0, @nicola, @irenegia, @lucaniz, @Kubuxu, @Stebalien, @jennijuju, @arajasek, and @steven004 for contributing to this effort. We are likely missing some folks—please feel free to nominate those developers who have constantly contributed and impacted this front!

[taoshengshi]

@hunjixi - former core developer of go-filecoin (venus)

[kkarrancsu]

Dear Filecoin Community,

The FIL-RetroPGF organizers have spent the last week reviewing the nominations submitted in each of the eligible categories. We used the criteria outlined here to determine which projects would be invited to apply in this round. Our assessment of eligibility is based on our best interpretation of the information provided in the nomination but tended to bias towards inclusivity, deferring judgement to badgeholders as much as possible.

The following page contains a list of projects that are being invited to apply to be eligible for FIL-RetroPGF-1 funding.

Applications are open from 8 Apr 2024 to 22 Apr 2024. We encourage you to submit your application here as soon as possible!

Best regards,
FIL-RetroPGF Organizers

[kkarrancsu]

Dear Filecoin Community,

This is a gentle reminder that the application submission deadline is fast approaching. If you are a project invited to apply and have not already done so, please submit your applications here by Apr 22.

We have updated the Round 1 - Invited Applications page to include a POC. Additionally, we have published the following guidelines on how to approach the application process and included an FAQ at the bottom.

Best Regards,
FIL-RetroPGF Organizers

[kkarrancsu]

Dear Filecoin Community,

We have extended the deadline to submit your applications. If you haven’t already done so, please submit your application here by 12PM UTC Apr 24, 2024. There will be no further extensions.

Best regards.

[kkarrancsu]

Dear Filecoin community,
Thank you for your continued patience. The organizers have finished tallying the results and computing the allocations for each project, according to the scoring rules that were previously set forth. Please visit the following resources for the results:

• FIL-RetroPGF-1 Results overview
• Results Spreadsheet
• Deep dive into the results.

We will contact project POCs in the coming days to coordinate distribution. A huge thank you to the entire Filecoin ecosystem for coming together to make this first round successful. Be on the lookout for Round 2!

[kkarrancsu]

Dear Filecoin Community,

Thank you for your continued patience. We are now in the next phase of FIL-RetroPGF-1—funds distribution. The Open Impact Foundation will coordinate distribution and contact projects that have been allocated funds from the FIL-RetroPGF-1 process.

Please be on the lookout in your inbox for emails coming from hello@openimpact.foundation

Best Regards,