A hardware isolation-based end-cloud dual-brain collaborative computing power scheduling system

Abstract

This invention discloses a hardware-isolated, edge-cloud dual-brain collaborative computing power scheduling system, relating to the fields of artificial intelligence computing power scheduling and data security technology. The system includes a local brain, a cloud brain, a hardware isolation unit, an identity verification unit, and a task scheduling unit. The cloud brain serves as a centralized computing power hub in the cloud, while the local brain is compatible with all specifications of intelligent computing power carriers. The two establish data boundaries through the hardware isolation unit and complete bidirectional verification through the identity verification unit. The task scheduling unit adaptively allocates tasks based on computing power level, privacy level, and network status. The system supports multi-mode networking, local offline operation, and pure hardware logic driving. This invention features a highly systematic architecture, compatibility with all types of isolation technologies and all specifications of computing power devices, outstanding data management capabilities, and adaptability to various scenarios requiring artificial intelligence computing power deployment.

Description

Technical Field

[0001] This invention relates to the field of artificial intelligence computing power scheduling and data security technology, specifically to a hardware-isolated edge-cloud dual-brain collaborative computing power scheduling system. Background Technology

[0002] Edge-cloud collaborative computing architecture is the mainstream deployment method for large-scale artificial intelligence applications. However, existing computing power scheduling architectures suffer from the following limitations: generic architectural terminology lacks a systematic definition of computing power carriers; incomplete hardware isolation methods and limited data protection dimensions; limited adaptability of computing power carriers, failing to support all specifications of computing power terminals; fixed network topology, not supporting multi-node collaboration and direct terminal connection architectures; and some architectures rely on algorithm operation, lacking sufficient independent hardware operation capabilities. These limitations prevent existing architectures from simultaneously meeting the technical requirements of full-domain device compatibility, multi-dimensional data security, and multi-scenario network adaptation. Summary of the Invention

[0003] Purpose of the invention The purpose of this invention is to overcome the aforementioned limitations in the prior art and provide a hardware-isolated, edge-cloud dual-brain collaborative computing power scheduling system. This system constructs a systematic dual-brain computing power architecture, enabling full-type hardware isolation protection, full-specification computing power carrier compatibility, multi-mode networking adaptation, and independent hardware operation without algorithm dependence.

[0004] Technical solution To achieve the above objectives, the present invention adopts the following technical solution: A hardware-isolated edge-cloud dual-brain collaborative computing power scheduling system includes a local brain, a cloud brain, a hardware isolation unit, an identity verification unit, and a task scheduling unit. The hardware isolation unit establishes the edge-cloud data isolation boundary, the identity verification unit performs bidirectional identity verification of scheduling commands, and the task scheduling unit adaptively allocates tasks based on computing power, privacy level, and network status. The local brain covers all specifications of intelligent computing power carriers, while the cloud brain connects to cloud-based large-scale models and supercomputing clusters. Both support direct network connection and multi-node collaborative network connection. The local brain has offline independent operation capabilities, and the system supports pure hardware logic-driven operation.

[0005] Beneficial effects This invention constructs a systematic edge-cloud dual-brain architecture, covering all types of hardware isolation technologies and all specifications of computing power terminals; it supports multi-mode networking adaptation, improving the adaptability of the architecture to various scenarios; it strengthens data management capabilities through two-way identity verification and data isolation; and it reduces system operation dependence by using the local brain to run offline and pure hardware-driven design, achieving integrated adaptation of computing power scheduling and data security. Detailed Implementation

[0006] The specific embodiments of the present invention will be described in detail below: In the hardware-isolated edge-cloud dual-brain collaborative computing power scheduling system described in this embodiment, the cloud brain is deployed on the cloud large model and supercomputing power cluster, serving as the centralized computing power processing hub and undertaking centralized computing power operation tasks; the local brain is an intelligent computing power carrier with independent computing power and privacy data processing capabilities, covering heavy, medium and light-duty equipment of all specifications, and undertaking edge-side task processing and data management. The hardware isolation unit employs one or more combinations of physical isolation, TEE, and security chips to establish a physical data isolation boundary between the local brain and the cloud brain, enabling data partitioning and transmission control. During the initiation and transmission phase of computing power scheduling commands, the identity verification unit performs two-way authentication with a global trusted root for both the scheduling requester and the data receiver; only after successful verification can the scheduling command be executed. The task scheduling unit collects the local brain's computing power level, the privacy level of the tasks to be processed, and the network connectivity status in real time, and adaptively allocates tasks to the local brain or the cloud brain for execution. Point-to-point communication links are established between multiple local brains to achieve cross-device computing power resource sharing and collaborative scheduling; a single lightweight local brain can directly establish a communication link with the cloud brain, independently forming an edge-cloud dual-brain collaborative architecture. When the network connection is interrupted, the local brain switches to independent operation mode, autonomously completing all task processing and closed-loop management of privacy data; after the network connection is restored, the local brain and the cloud brain only synchronize non-sensitive operational data to complete the architecture state synchronization. This system can operate independently of artificial intelligence algorithms, completing the entire process of computing power scheduling, identity verification, and data management based on pure hardware logic circuits. This embodiment has no limitations on device computing power specifications or application scenarios, and can be adapted to the computing power deployment needs of multiple scenarios, including civilian, commercial, and industrial applications. Attached Figure Description

[0007] Figure 1 This is a schematic diagram of the overall system architecture of the present invention; Figure 2 This is a schematic diagram of the computing power scheduling and network operation process of the present invention; Figure 3 This is a schematic diagram illustrating the full-specification local brain grading adaptation of the present invention.

Claims

1. A hardware-isolated edge-cloud dual-brain collaborative computing power scheduling system, characterized in that, It includes a local brain, a cloud brain, a hardware isolation unit, an identity verification unit, and a task scheduling unit; The hardware isolation unit is used to establish a data physical isolation boundary between the local brain and the cloud brain; The identity verification unit is used for two-way trusted identity verification before the transmission of computing power scheduling instructions; The task scheduling unit is used to allocate tasks to the local brain or cloud brain for execution based on computing power level, data privacy level and network status.

2. The hardware-isolated edge-cloud dual-brain collaborative computing power scheduling system according to claim 1, characterized in that, The hardware isolation unit adopts any one or more combinations of physical isolation, Trusted Execution Environment (TEE), TrustZone, independent security chip, logical isolation, and virtualization isolation.

3. The hardware-isolated edge-cloud dual-brain collaborative computing power scheduling system according to claim 1, characterized in that, The local brain is an intelligent computing power carrier with computing power and privacy data processing capabilities; the cloud brain is a centralized computing power hub deployed in cloud-based large models and supercomputing clusters.

4. The hardware-isolated edge-cloud dual-brain collaborative computing power scheduling system according to claim 3, characterized in that, The local brain includes enterprise mainframes, enterprise servers, home mainframes, home servers, intelligent robots, vehicle-mounted computing units, smartphones, smartwatches, smart bracelets, and portable smart terminals.

5. The hardware-isolated edge-cloud dual-brain collaborative computing power scheduling system according to claim 1, characterized in that, Multiple local brains support point-to-point communication, enabling cross-device computing power collaboration and resource sharing; a single local brain can communicate directly with the cloud brain, forming an independent end-to-cloud dual-brain collaborative architecture.

6. The hardware-isolated edge-cloud dual-brain collaborative computing power scheduling system according to claim 1, characterized in that, When the network is interrupted, the local brain independently completes task processing and data management; after the network is restored, the local brain synchronizes non-sensitive data with the cloud brain.

7. The hardware-isolated edge-cloud dual-brain collaborative computing power scheduling system according to claim 1, characterized in that, This system can achieve computing power scheduling and local closed-loop management of privacy data based on pure hardware logic, and can run independently without artificial intelligence algorithms.

Images