🌐 Tracebloc Client

📄 Description

Tracebloc Client is a Kubernetes-based application that runs experiments and communicates results to the Tracebloc backend. It's designed to handle distributed machine learning workloads efficiently and securely.

🛠️ Tech Stack

• Kubernetes
• Helm 3.x
• Azure Service Bus (AmqpOverWebsocket)
• Docker
• Persistent Volume Storage

🚀 Installation & Setup

Prerequisites

• kubectl installed and configured
• Helm 3.x installed
• Access to a Kubernetes cluster

System Requirements

Network Requirements

• One-way communication with Tracebloc backend
• Port 443 open for Azure Service Bus (AmqpOverWebsocket)
• Secure metric and weight file transmission

Cluster Specifications

• RAM: 50 GB (minimum)
• CPU: 20 cores (minimum)

Storage

• Persistent volumes for:
  • Training data
  • Models
  • Weight files

Required Credentials

1. Docker Registry Access:

   • Username
   • Password

2. Client Authentication:

   • Client ID
   • Username
   • Password

📦 Deployment Guide

1. Ensure all prerequisites are met
2. Configure your credentials
3. Follow our detailed deployment guide at: Create Your Client

📜 License

This project is licensed under the MIT License - see the LICENSE file for details.

📞 Support

For additional support or questions, please refer to our documentation or contact the Tracebloc support team.