A distributed reasoning federation cooperation system based on separation of control plane and data plane
By establishing a distributed inference consortium collaboration system based on the separation of the control plane and the data plane, the problems of data privacy protection and low collaboration efficiency in distributed artificial intelligence are solved, and efficient cross-institutional data collaboration and secure data access control are achieved.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- SUZHOU ABYSS MATRIX TECHNOLOGY CO LTD
- Filing Date
- 2026-04-07
- Publication Date
- 2026-06-19
AI Technical Summary
Existing technologies in distributed artificial intelligence and privacy computing suffer from problems such as coarse security granularity, disconnect between static trust and dynamic behavior, poor compatibility with heterogeneous data, and lack of mandatory control over the data surface, resulting in low efficiency in data privacy protection and collaboration.
A distributed inference alliance collaboration system based on the separation of control plane and data plane is adopted. Through modules such as node registration and measurement, task parsing and orchestration, compatibility analysis and reorganization, remote proof and credential binding, sandbox execution and mandatory verification, trusted transmission and inference calculation, result summary and aggregation feedback, it realizes precise binding and fine-grained control from static code measurement to dynamic data flow.
It enables efficient cross-institutional data collaboration and inference computation while protecting data privacy, ensuring strict constraints on data access scope and secure transmission, and improving system security and collaboration efficiency.
Smart Images

Figure 1 
Figure 2