BalanceAI — Weekly Report
Introduction
Welcome to our weekly update, a summary of the most recent progress and achievements in our project. To access the previous update, please click on the following link: Weekly Report — May 12, 2024
1. We presented the concept of V0 — The Torus (BalanceAI Network — Mainnet)
“The Torus” represents the inaugural phase of BalanceAI Network, signifying the commencement of a convergence between AI and blockchain technologies.
This stage encapsulates the inception of a balanced and symbiotic relationship, as the inaugural phase of our journey unfolds. The Torus sets the base for seamless collaboration, innovation, and the realization of our collective vision.
More information is in the article:
2. V1 Design process has started
We have started working on the design documents for the V1.
- On-Chain Access: The ability for users and/or applications to interact with AI models in a controlled (by Balance AI blockchain data) environment.
- AI Models Wrapped : A wrapper framework serves as a software development kit (SDK) designed to expose AI models via API, streamlining the process of integrating these models into Balance AI Marketplace. This framework automates several crucial aspects related to on-chain access, payments, security, and validation, enhancing the accessibility and usability of AI models in on-chain environment.
3. We have started the migration process to a new domain
We are currently in the process of migrating our website to our new official BalanceAI Network domain. The site will be under construction for a brief period. Coming soon, we will begin publishing updated project information documents.
4. ETH <-> BAI Bridge
Works are underway on the bridge infrastructure (with some community members as the bridge is going to be run independently by the community).
We are working on the Bridge v0 that would feature hardware wallet based authentication to guarantee security of all transaction and minimize any security risks.
We are also researching BEEFY (Bridge Efficiency Enabling Finality Yielder) to fully decentralize bridge services in the v1 version.
https://spec.polkadot.network/sect-finality#sect-grandpa-beefy
https://polkadot.network/blog/the-landscape-of-trustless-bridges-on-polkadot
5. Our V0 Testnet is fully operational
Beta Testnet V0 can be found at: rpc-testnet.balancedao.io
Explorer: https://polkadot.js.org/apps/?rpc=wss%3A%2F%2Frpc-testnet.balancedao.io%2F#/explorer
6. First version of the new documentation is ready
At present, we’re deeply engaged in the meticulous process of documentation. Our focus lies in organizing, detailing, and clarifying vital information that underpins our project. With a keen eye for accuracy and clarity, we strive to ensure that our documentation serves as a reliable reference point for our team members and stakeholders alike. Through this concerted effort, we aim to streamline workflows, enhance communication, and fortify the foundation upon which our endeavors rest.
We are working on publishing the new documentation on our new portal.
7. Audit
We have established a commercial arrangement with the Auditor company (recommended by Astar team). We are currently working with them on auditing. It is a tedious process as we need to make sure V0 is a stable basis for all our next efforts. Work involves re-running all tests, get their feedback, fixing issues and run a full final audit.
Preliminary audit has been completed, the audit plan has been established as follows.
Runtime implementation baseline check. We delve deep into the implementation of BalanceAI to identify any divergence from the design that might render it susceptible to vulnerabilities.
- Attack assessment: Drawing from the previously postulated attack paths, our hacking team identifies implementation flaws. The team endeavors to find ways to successfully execute attacks.
- Hacking vector exploration: In tandem, our experts probe the system from multiple angles. This parallel approach ensures that we holistically cover BalanceAI’s stack, focusing particularly on the runtime logic of the implementation.
- Dynamic testing: We employ dynamic testing methodologies, including coverage guided fuzz testing on the BalanceAI runtime implementation. This approach facilitates the discovery of hidden vulnerabilities by generating a plethora of random inputs to scrutinize how the system responds.
The review covers BalanceAI v0:
A. BAI v0: Network Genesis and Chainspec
B. BAI v0: Network Runtime Configuration
C. BAI v0: Governance Configuration
8. Monitoring infrasctructure
We are working on preparing solid monitoring infrastructure for the V0 Testnet and V0 Mainnet.
The monitoring hub in action with our latest V0 Testnet with 3 validators:
To stay updated with our latest news and developments, be sure to follow us on our official channels.