This document contains a proposal for managing the diverse set of technologies require to implement decentralize data unions using Ancon Protocol v1.0 and named Decentralized Verifiable Data Agent, where like Verified Credentials agents, you have a network of agents, where an agent can be a smartphone, IoT, browser, server or edge application.
It describes separating Ancon Protocol v1.0 into layers, which separate concerns and aspects such as consensus and proofs, storage and data, and computation and jobs.
We will explain the next version of the protocol, which includes removing the dependency of Graphsync, incorporating SQLite APIs, Indexed DB, VFS and Waku (a libp2p library). Additionally we will include separating the Dag Storage Nodes, Indexers and Merkle Trees consensus in Tendermint or Polkadot.
For the last 40 days the team has been working on implementing a use case for protocol v1 with great results and confirmations of the protocol choice of technology stack.
However we left some pendings protocol features for later, which we assume was not going to be a blocker for implementing dapps as the initial plan. The following items are major paint points to consider:
- Distributed Data Query
- Incomplete Dag syncronization
- Partial suppport for offchain computation
- Dag Indexing, publishing and privacy
Previously, we propose DAG Chain Redux as a way to consolidate indexing using smart contracts events. While this will work in server side scenarios it does not cover client side or light client nodes like browsers or smartphones. It also does not have any compatibility with for example enterprise ethereum trusted offchain computing specification.
Currently Ancon Protocol v1 only synchronizes with IPFS pinning services. While Ancon Protocol is compatible with Graphsync it will only work for server side scenarios because the current Graphsync is only available with Go Language SDKs.
Ancol Protocol v1 + DAG Chain Redux in its current iteration is implemented using JEXL which is an expression language for javascript. Ideally a complete feature will have to include QuickJS compile WASM artifacts and will have to be run in a specific client side node.
Besides storage and computation a succesful data union must include: Data indexing, Data Publishing, Security and Privacy controls which are nice to haves to the core protocol features.
As per ADR-040, Cosmos is migrating towards SMT (Sparse Merkle Trees) using Celestia SMT. This change makes newer ICS23 Proof storage DB agnostic.
V1.5 takes major inspiration from Foundation DB, Static torrent website with peer-to-peer queries over BitTorrent on 2M records and A future for SQL on the web. Also we will take ideas from Datasette publishing feature.
-
Implement a concrete Typescript class from blockstore-core interface, using AbsurdSQL. This blockstore becomes a fsstore-like storage, with Waku as networking layer and JS IPLD and multiformats as DAG block layer.
-
Permanent sync and large assets of data using adapters to Arweave, Filecoin and/or Subspace using $ANCON + cross chain liquidity bridges
-
Implement SQLite queries with gas consumption and estimation
-
Implement block sync with Waku (ideas: Litestream)
-
Implement dual layer data-execution module: a DAG layer (immutable) and a DB layer (read-only), where (a) Any insert or updates are signed and created as DAG blocks, and a linkeable and versioned JSON data model is stored in the DB Layer. (b) Read-only queries are executed on DB layer (c) GraphQL is used as execution layer and lower level QuickJS Smart Contracts for other use cases.
- Manages account keys, DID compatible with Ed25519 crypto suite
- Relayers will in v1.5 query the state commitments from Ancon Node chain, with the incentives ocurring when posting root blocks from Ancon Node chain to Ancon relayers
- We propose, a gasless chain, that only stores state commitments separate from the state storage, taking inspiration from ADR-040.
- DAG blocks commit transactions and relayers submit roots to protocol smart contracts
- Messages encrypted with recently announced
HPKEfor secure forward secrecy. - Add BBS+ for selective disclosure
This feature is left as a nice to have, it mostly focuses on creating Docker container like dapps, using the DAG Chain Redux either with Ancon Protocol v1.0 or v1.5 Data Agent Library.
A Verifiable Document DApp is a JSON schema similar to an ERC-721 Metadata and a Verified Credential / Presentation, in this case a Verifiable Document, used to create vendored (kiosk-like) dapps. A JSON defines the details and app metadata, signed as a payload and stored as a DAG block (DID verifiable).
{
name: "Selling 5 lottery tickets for Sunday 13 March 2022",
description: "5 lottery tickets of # 55",
link: "https://tensta.did.pa/v0/file/baguqeerar56ywt7p3qbbf6wiqgja5ybvkblz4wx37vyybthr2jl65d657jha/",
service: "urn:dapp.ancon.did.pa:lottery"
}Where link contains the data and service the dapp that renders the data dapp. The Data Dapp must implement the following interface:
async fetchData(), gets the link metadataasync renderServiceTemplate(), gets the UI template for the specific service
UI templates must be contained in an HTML component, similar to a VueJS component which contains HTML, Code and Stylesheets. Recommendation is to create Verifiable Document DApps per user flow (Eg Enroll User, Swap token, Mint NFT).
DApp Provenance information can be display using the DAG Block response signature.
A Verifiable Document DApp must always start user interaction from a QR or shareable link.
Verifiable Document DApps requires a V1.0 server side DAG Chain redux handler or a V1.5 Data Agent Library, this component maintains a DAG Chain Redux using web workers/server side code.
Web 3.0 Onchain events ---> DAG Chain Redux indexer(V1.0 server side code or V1.5 Web Workers)
Where the reducer is implemented as JEXL or a QuickJS Smart Contract (V1.0). For V1.5 Data Library, SQLite engine with GraphQL interface must be used as it takes care of data sync, publishing and security.
- Offloads server nodes indexing to edge devices, together with Waku and Ancon Data Agent library, allow indexing to happen close to the owner, maintains privacy, and only publish public datasets or collections to data marketplaces (like eg NFT marketplaces).
- Allows for Ancon Protocol v2.0 - Content Provenance and Authenticity to be fulfilled, by creating proof of android or ios with SafetyNet or DeviceCheck, maintaining access control and origin signing of content, like images, audios or videos created with smartphones or IoT devices.
- Perfects the use case for ad-hoc networks of data unions, and preserves security, privacy and integrity by separating state commitments (proofs) from state storage (DAGs).
In this whitepaper we described Ancon Protocol v1.5, which integrates v1.0 paradigm shift of verifiable, offchain metadata and vector commitments proof for cross chain scenarios, with features to expand into data unions and trusted offchain computing, while also being open for futures roadmap related to content provenance and authenticity for Web 3.0 data economies.
Rogelio Morrell,Kendall Kant,Ricardo Menotti for Industrias de Firmas Electronicas SA , 2022.