Federated Learning: Self-managed On-premise or SaaS?

Scaleout offers federated machine learning solutions that maintain data ownership and privacy without central data pooling. Our federated learning framework, FEDn, supports machine learning across distributed datasets like data silos or mobile devices, respecting privacy and cybersecurity requirements. Our SaaS model simplifies access to federated learning technology, enabling users to focus on ML development rather than infrastructure management, especially beneficial during early project development.

FEDn can be deployed on user-specific hardware or virtual private clouds, managed by the user or Scaleout, offering flexible security options for advanced security needs like IP-sensitive ML models. Our goal is to simplify the adoption of federated learning technology for all users.

Deployment Models

FEDn is available in various deployment models to suit different project stages:

• SaaS (Software as a Service): For managing projects and uploading ML code and models without extensive DevOps involvement. Ideal for early-stage projects, including pilots and proof-of-value phases.
• Self-managed: FEDn can be deployed on private clouds or on-premise for organizations with strict cybersecurity needs, allowing full control over deployment, security, and privacy.

Capabilities

• Multi-tenant and Single-tenant Options: Supported to meet diverse organizational needs.
• Complex Operations Management: Scaleout handles server aggregation, data storage, user authentication, network configuration, and system monitoring, reducing the operational burden on users.

Conclusion

Scaleout's FEDn framework follows cloud-native principles for seamless deployment across major cloud providers. Whether you choose our fully managed SaaS or self-managed solution, FEDn offers flexibility and control for various federated learning scenarios. Start with our SaaS model and scale to a self-managed model as your needs evolve, maintaining performance and adaptability.