Targon: The Confidential Decentralized AI Cloud

Targon is a next-generation AI infrastructure platform that leverages Confidential Compute (CC) and Protected pcie (PPCIE) technology to secure the entire stack. By providing a secure execution environment from hardware to application layers, Targon enables verifiable and trustworthy operations across the entire infrastructure in a decentralized fashion.

NOTICE: Using this software, you must agree to the Terms and Agreements provided in the terms and conditions document. By downloading and running this software, you implicitly agree to these terms and conditions.

Overview

Targon provides a comprehensive secure AI infrastructure powered by CC and PPCIE technology:

Docs

For validators, see validator docs

For miners, see miner docs

For mongo-wrapper service, see mongo-wrapper docs

---

Core Security Features

• Hardware-enforced memory encryption and protection
• Secure boot with hardware root of trust
• GPU TEE (Trusted Execution Environment) for isolated execution
• Remote attestation for verifiable computation
• Secure key management and cryptographic operations
• Protected execution environment with memory isolation

AI Infrastructure Capabilities

• End-to-end secure model inference pipeline
• Hardware-level attestation and verification
• Protected model execution with CC or PPCIE isolation
• Verifiable computation through remote attestation
• Secure memory management for AI workloads
• Isolated execution environment for sensitive operations

Current Implementation

• NVIDIA Confidential Compute integration OR
• NVIDIA PPCIE
• Hardware-level security guarantees
• Protected inference execution
• Remote attestation capabilities
• Secure memory encryption
• Isolated compute resources

Future Roadmap

• Secure model training with Confidential Compute or PPCIE
• Protected data processing and storage
• Bare metal access for secure AI workloads
• Comprehensive AI development platform
• Developer-friendly tools and interfaces
• Multi-vendor Confidential Compute / PPCIE support:
  • AMD SEV-SNP integration
  • Additional hardware security technologies

Contribution Guidelines

1. Proposal Requirements

   • Must be discussed on Discord and other community channels
   • Must demonstrate clear technical consensus
   • Must be approved by project maintainers

2. Implementation Process

   • Changes must be thoroughly tested
   • Must include migration plans if needed
   • Must consider backward compatibility
   • Must document security implications

3. Making Changes

   • Follow the existing code style
   • Write clear commit messages
   • Include tests for new features
   • Update documentation as needed

Community Guidelines

1. Communication

   • Be respectful and professional
   • Provide constructive feedback
   • Help others when possible
   • Follow the project's code of conduct

2. Collaboration

   • Respond to review comments promptly
   • Be open to feedback and suggestions
   • Help maintain project quality
   • Share knowledge with the community

Remember: The goal is to improve Targon and Bittensor as a whole. We welcome all contributions that align with this mission and follow our guidelines.