CryptoNet is a cross-organizational research lab focused on cryptography and protocol design with application in crypto networks - in particular IPFS, Filecoin and Libp2p. CryptoNet is currently 10 full time researchers and engineers, with over 30 active collaborators (from Aarhus, IMDEA, Stanford, Ethereum Foundation and more, see our community) and 6 academic advisors.
Before diving into our achievements, a key goal of this quarter was our public online presence:
- Website: This very website (https://cryptonet.org) is live and running and contains all of our on-going work and available for anyone to learn and contribute. Every project has milestones and call to action (see our list for more)Projects
- Discord: the community interacts in a public discord with 60+ engineers, protocol designers and researchers (join here https://discord.gg/cryptonet)
The following are the list of achievements in Q3 in three different areas of work.
TL;DR
- Proposed 7 Filecoin Improvement proposals that were accepted.
- Launched a line of small composible storage products: Retriev.org for data retrievability, Web3bounty.app for exporting Filecoin to Ethereum.
- Published 8 academic papers at top conferences.
- Proposed a total of 5 new vector commitments, amongst which Caulk received major attention.
Protocols
CryptoNet is the most prolific Filecoin Protocol contributor
Filecoin is a decentralized storage network with over 18 exabytes of available storage capacity. CryptoNet was a main contributors in the early design and development of the Filecoin protocol and continues today to play a key role in the ecosystem.
Improvements to Filecoin Protocol follow a formal process called FIP (Filecoin Improvement Proposal), where community members can propose changes to the core protocol that are then propagated in new software releases.
Just in Q3, our members authored 7 FIP (out of 12 proposed) and audited several, amongst which:
- Auditing highly requested features (FIP36 Sector Multiplier)
- Simplifying the protocol (FIP45 Decoupling FIL+ from Markets)
- Creating new standards (FRC46 Fungible Token Standard)
- Enabling forward compatibility for future changes (FIP41 PreCommit independent of sector content).
- Establish security measures in case of cryptographic breaks (FIP47 Porep Security Policy)
Our members worked on two main efforts:
Complete list of proposed protocol upgrades in Q3: FIP0047 (PoRep Bug Policy), FRC0046 (Fungible Token Standard), FIP0045 (Decouple FIL+ registry from Markets), FIP0044 (Authentication method for Actors), FRC0042 (Calling convention with hashed method name), FIP0041 (Forward Compatibility of precommit). (See
Products
CryptoNet soft launched a line of storage protocols
We assembled the “Onchain Storage” team to create a new line of very small composable storage protocols.
This project aims at:
- Augmenting the Filecoin storage offer with new features (such as data retrieval guarantees and perpetual storage)
- Serving decentralized storage solutions directly via smart contracts instead of traditional web2 apis, so that smart contracts and DAO can program their storage needs (for auto-backup of NFT at minting, crowd-funding storage of important DAO assets)
- Exporting Filecoin storage into other chains, to make over 18 exabytes of storage available for their needs (such as data availability for layer-2s)
Anyone can design, compose storage protocols. We plan to create an ecosystem of products that can all interoparate with each other and can be accessed directly via smart contracts instead of APIs. You can read more about this project in our
Just in this quarter our members went from theory to deployed on the following projects.
"Data Retrievability Insurance"
Retriev.org: A protocol and a web app for last resort retrieval. Retriev provides a cryptoeconomic incentive for storage providers to serve their files when requested, failure to do so repeatedly results in penalties.
Currently live on Ethereum testnet.
"Exporting Filecoin storage in other chains"
Web3Bounty.app: A protocol and a app that allows to store files into Filecoin directly from Metamask with a transaction in Ethereum. This allows for smart contracts to pool assets together to store files into Filecoin. Currently it uses web3.storage, in the future it will store data directly into Filecoin.
"Reputation for IPFS and Filecoin"
Storage Metrics DAO: the project was stuck with the lack of existing network scanning tooling and a post-mortem was produced. This document created momentum and drove key conversations that is leading the formation of teams to build these tools, which we are currently advising (Validation bot, ProbeLab).
Note that the above projects are still in test phase, use at your own risk!
Research
CryptoNet continues to be a prolific research organization
Our researchers and extended network of collaborators have made substantial progress on all of our research efforts (see the our
Amongst our main achievements: 8 papers in Q2-3 only in top academic conference (CCS, Eurocrypt and more, see full list here
- Caulk: Lookup Arguments in Sublinear Time (CCS 2022)
- Linear-map Vector Commitment and their applications (Asiacrypt 2022)
- Encryption to the Future (Asiacrypt 2022)
- Harisa: Succinct Zero-Knowledge Batch Proofs for Set Accumulators (CCS 2022, CESC 2022)
- Z🥝: Zero-Knowledge for Homomorphic Key-Value Commitments: new efficient primitives (SCN 2022)
- MyOPE: Malicious securitY for Oblivious Polynomial Evaluation (SCN 2022)
- What Makes Fiat--Shamir zkSNARKs (Updatable SRS) Simulation Extractable? (SCN 2022)
- On the Impossibility of Algebraic Vector Commitments in Pairing-Free Groups (TCC 2022)
Vector Commitments and Caulk
We proposed 5 new vector commitments this year only: Caulk, Linear-map VC, Muppets, Curve Trees, Harisa from our members.
Caulk is a linkable VC and look-up tables with high impact on: SNARKs proving time for complex circuits (including sha), potential to remove Merkle trees with applications in private cryptocurrencies, stateless blockchains. Caulk could massively reduce the cost of SHA in Filecoin circuits (which accounts for >90% of the Proof of Replication circuit).
Universal Trusted Setup, Super fast SNARKs with Testudo
Testudo is a universal trusted setup SNARK that is prover efficient (meaning that it will massively cut the GPU cost for Filecoin proof production). Initial results promise that it will be more competitive than Halo2 and that it will require minimal new components, since it is based on Groth16 (hence faster to ship).
CryptoNet research grants had great success
We have 4 on-going research grants, two of which led to important milestone in cryptography:
- First post-quantum snark based on lattice with publicly verifiable proof (by Lai et al).
- First functional vector commitment with functional openings with constant-size parameters (by Fiore et al.)
We also launched
CryptoNet also organized 2 Research dinners with highly selected researchers (EthCC and Layer 2 Amsterdam) and over 10+ talks at top academic and community conferences (CCS, Eurocrypt, ZKSummit, ZK Study club).