Back to Gap Map Model components / Federated learning

Rhino Federated Computing Platform

closed / Overall score: n/a

Rhino Federated Computing's platform for federated statistics, federated learning and federated inference on data that stays behind each custodian's firewall. Custodians connect data across clouds and on-premises sites; custom code runs next to the data in secure containers under role-based access control, with differential privacy, k-anonymization, homomorphic encryption, customer-managed keys and audit logs.

Closed, and on the map as an ADR-005 best-in-class comparator for this category. The `rhino-health` Python SDK on PyPI is a proprietary client for the platform, not the platform, so it is not declared.

Openness

1 high confidence

1.0

license: Proprietary(closed platform
source: closed(no published implementation)

Rhino publishes no implementation of the platform, and even its Python SDK ships under a proprietary license.

https://pypi.org/pypi/rhino-health/json recorded 2026-09-27
PyPI `rhino-health` 2.2.1: classifier "License :: Other/Proprietary License"; "Programmatic interface for interacting with the Rhino Federated Computing Platform".
https://www.rhinofcp.com/solutions/platform recorded 2026-09-27
The platform page: a "secure, scalable software solution for federated learning and collaborative data processing" offered by demo request; no source or self-hosted open edition.

Adoption

not assessed

Rhino reports figures for individual customer networks, such as more than 125 biopharma partners in one pharma network, but no platform-wide count of users or deployments, so no adoption level is recorded.

https://www.rhinofcp.com/ recorded 2026-09-27
Home page: "125+ biopharma partners" and "3 million+ inference runs" in one Top 5 pharma network; a national cancer research alliance; a "450-partner research consortium". Per-network figures, not a platform-wide count.

Capability

4 medium confidence

4.0

Rhino runs production networks across organizations with a broad governance and privacy toolkit, level with Flower. It documents no device or HPC reach and no multi-billion-parameter training figures.

https://www.rhinofcp.com/solutions/platform recorded 2026-09-27
"Federated statistics, federated learning, and federated inference"; "use existing models (including LLMs)"; "privacy-enhancing techniques such as differential privacy, k-anonymization, and homomorphic encryption"; "Custom code is deployed at the source of the data in secure containers"; "RBAC, encryption with customer-managed keys and comprehensive audit logs".

Verified 2026-09-27