Medical Imaging Compute Allocation for Latency-Sensitive Reconstruction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing medical imaging systems face inefficiencies in managing local compute instance (LCI) resources, leading to underutilization and increased costs due to suboptimal task allocation and reliance on cloud-based processing, which can be time-consuming and costly.
Innovation Solution
A resource allocation system dynamically manages tasks across a network of medical imaging systems, prioritizing local compute instances (LCIs) for latency-sensitive operations and redistributing tasks based on availability and latency, using a centralized controller to optimize resource use and fallback models for network failures.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Power
If cloud-based processing is used for medical imaging tasks, then computational power and processing capability are improved, but data transmission time and network dependency increase
Solution Approach 1:
The system segments processing tasks into two categories: latency-sensitive tasks executed locally on LCIs and non-latency-sensitive tasks executed remotely on CCI. This segmentation allows the system to utilize cloud computational power for suitable tasks while maintaining fast response for time-critical operations, thereby resolving the contradiction between computational power and data transmission time.
Solution Approach 2:
The resource allocation system acts as an intermediary that dynamically assigns tasks between local and cloud compute instances based on task characteristics and current system state. This intermediary layer enables optimized task distribution, allowing the system to leverage cloud resources without incurring unnecessary transmission delays for latency-sensitive operations.
2Productivity
If cloud-based processing is used for medical imaging tasks, then processing capability is improved, but system costs increase
Solution Approach 1:
The system dynamically adjusts task allocation between local and cloud resources based on real-time conditions such as LCI availability, task priority, and network status. This dynamic approach ensures that cloud resources are utilized only when necessary and beneficial, optimizing processing capability while minimizing the energy costs associated with cloud computing and data transmission.
Solution Approach 2:
The resource allocation system changes operational parameters by monitoring LCI utilization metrics and adjusting task assignment strategies accordingly. When LCIs are available and suitable, tasks are assigned locally to reduce costs; when cloud resources provide significant processing benefits, tasks are assigned to CCI. This parameter-based dynamic adjustment resolves the contradiction between processing capability and system costs.
3Loss of time
If tasks are reassigned from cloud to local compute instances, then data transmission time is reduced, but LCI availability and workload capacity become constraints
Solution Approach 1:
The system implements feedback mechanisms that continuously monitor LCI availability, workload capacity, and task completion status. Based on this feedback, the resource allocation system dynamically adjusts task assignment decisions, reassigning tasks between local and cloud instances as conditions change. This feedback-driven approach allows the system to reduce data transmission time when LCIs are available while adapting to LCI capacity constraints when they occur.
4Reliability
If fallback mechanisms are implemented for network failures, then system reliability is improved, but system complexity increases
Solution Approach 1:
The system performs preliminary actions by pre-positioning fallback capabilities and maintaining awareness of both local and cloud resource capacities before failures occur. The resource allocation system is designed with built-in fallback logic that automatically activates when network failures or LCI unavailability are detected, eliminating the need for complex post-failure recovery mechanisms and reducing overall system complexity while maintaining high reliability.
Data Source
Figure 1
Figure 2
Figure 3
AI summary
Methods and systems are provided for increasing an efficiency of use of computing resources of a plurality of connected imaging systems, based on adaptive prioritization of tasks within and between compute instances (108, 112, 122, 132, 142, 152, 162) and centralized task assignment. When a scan is not being performed using an imaging system, computing resources of the imaging system may be operated in a cooperative mode, where the computing resources may be advantageously used to perform processing tasks for other imaging systems of the connected imaging systems. Computing resources may be allocated based on compute time (606), data transmission time (604), and availability of compute instance (612). In particular, cloud-based AI models (800) used by an imaging system may be integrated more efficiently. For example, local versions of such models may be cached, for operation in the event of a failure of external network connections.