mirror of
https://github.com/ultravioletrs/cocos.git
synced 2026-06-23 04:10:25 +00:00
Sammy Kerata Oina
cef9cbbecd
CI / ci (push) Has been cancelled
* Refactor AgentServer interface and update related implementations; adjust dependency versions in go.mod Signed-off-by: Sammy Oina <sammyoina@gmail.com> * Implement State method for agent.Service and enhance metrics and logging middleware to track state changes Signed-off-by: Sammy Oina <sammyoina@gmail.com> * Update file permission modes to use octal notation in CVMS client and agent main Signed-off-by: Sammy Oina <sammyoina@gmail.com> * Refactor CVMS client and agent main to improve function signatures and variable naming for clarity Signed-off-by: Sammy Oina <sammyoina@gmail.com> * Remove unnecessary fmt import and logging statement in CVMSClient Process method Signed-off-by: Sammy Oina <sammyoina@gmail.com> * Add file-based storage implementation for message persistence in CVMSClient Signed-off-by: Sammy Oina <sammyoina@gmail.com> * Update README and main.go for CVMS service: correct references and improve clarity Signed-off-by: Sammy Oina <sammyoina@gmail.com> * Update README to clarify gRPC host and port descriptions for CVMS server Signed-off-by: Sammy Oina <sammyoina@gmail.com> * Rename sendMessageWithRetry to sendStreamMessage for clarity and consistency in CVMSClient Signed-off-by: Sammy Oina <sammyoina@gmail.com> * Update README to add detailed descriptions for algorithm and dataset paths Signed-off-by: Sammy Oina <sammyoina@gmail.com> --------- Signed-off-by: Sammy Oina <sammyoina@gmail.com>
Cocos AI 🥥
Confidential Computing System for AI
Made with ❤️ by Ultraviolet
Guide | Contributing | Website
Introduction 🚀
Cocos AI is a cutting-edge platform designed to enable secure multiparty computation (SMPC) using Confidential Computing and Trusted Execution Environments (TEEs).
It empowers organizations to collaboratively process sensitive data for AI/ML workloads while ensuring:
- 🔒 Data Privacy: Your data stays encrypted and secure throughout the computation.
- 🛡️ Trust and Integrity: Protected by hardware enclaves with robust remote attestation protocols.
- 🤝 Seamless Collaboration: Multiple organizations can work together without exposing sensitive information.
Features 🛠️
Cocos AI provides essential features for secure and efficient collaborative AI/ML:
- 🖥️ TEE Enablement and Monitoring: Secure VM management for deploying and monitoring workloads.
- 🛡️ Hardware Abstraction Layer (HAL): Built on a hardened Linux kernel, secure bootloader, and minimal root filesystem (minimal TCB).
- 🕵️ In-Enclave Agent and Networking Controller: Essential system software for managing secure workloads.
- 🔒 Encrypted Data Transfer: Asynchronous data transfer and secure result delivery.
- 🛠️ API for Platform Manipulation: Programmatic control for managing workloads.
- ✅ Attestation and Verification Tools: Hardware- and software-supported attestation for integrity assurance.
- 🖱️ Command-Line Interface (CLI): A user-friendly CLI for system interaction.
🚀 Quick Start
Clone the Repository and Build Binaries
git clone git@github.com:ultravioletrs/cocos.git
make
This will generate three binaries:
ls build/
# cocos-agent cocos-cli cocos-manager
Deployment Overview:
- Manager: Deploy on the AMD SEV-SNP host to orchestrate workloads.
- Agent: Build into the EOS-based HAL for secure enclave management.
- CLI: Interact with remote agents to control operations.
📚 Documentation
Comprehensive documentation is available at the official documentation page.
For CLI usage details, visit the CLI Documentation.
Documentation is automatically generated from the docs repository. Contributions to documentation are welcome!
🛡️ License
Cocos AI is published under the permissive open-source Apache-2.0 license. Contributions are encouraged and appreciated!