Memory Network Prioritizing Access Requests
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Computing systems face performance degradation and SLA violations due to memory access contention among multiple applications sharing a common memory, leading to inefficient resource allocation and varying service levels.
Innovation Solution
A memory network that prioritizes memory access requests based on application-specific requirements, using a combination of memory address-based and node-based labeling schemes to allocate bandwidth and dynamically adjust priority levels, ensuring differentiated quality of service and meeting Service Level Agreements (SLAs).
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If multiple applications share a common memory without prioritization, then memory resource utilization is maximized, but memory access performance degrades due to contention
Solution Approach 1:
The patent segments memory access requests into different priority classes (e.g., high priority and low priority queues) based on application requirements. This segmentation allows the memory controller to manage different types of access requests separately, ensuring that critical applications receive guaranteed bandwidth while less critical applications share remaining resources, thereby resolving the contradiction between performance and complexity.
Solution Approach 2:
The patent applies local quality by providing differentiated service levels to different memory access requests. Instead of uniform treatment, the system assigns different quality attributes (bandwidth guarantees, priority levels) to different applications or access patterns, allowing high-priority applications to receive enhanced service while maintaining overall system efficiency.
2Reliability
If memory access requests are treated equally, then system simplicity is maintained, but Service Level Agreements (SLAs) cannot be enforced
Solution Approach 1:
The patent implements preliminary action by allowing applications to pre-specify their priority levels and bandwidth requirements before memory access occurs. The memory controller uses these pre-configured parameters to automatically enforce SLAs without requiring real-time manual intervention, thus achieving reliable SLA compliance while maintaining operational simplicity through automated enforcement.
Solution Approach 2:
The patent changes parameters such as priority levels and bandwidth allocation based on application requirements. By dynamically adjusting these parameters according to pre-configured SLA agreements, the system achieves reliable service level enforcement while maintaining ease of operation through automated parameter modification based on application profiles.
3Productivity
If bandwidth is allocated uniformly to all applications, then resource allocation simplicity is maintained, but performance constraints cannot be met
Solution Approach 1:
The patent implements dynamics by enabling flexible bandwidth allocation that adapts to application needs. The system can dynamically adjust bandwidth distribution based on priority levels and SLA requirements, allowing high-priority applications to receive guaranteed bandwidth while low-priority applications share remaining resources. This dynamic allocation mechanism meets performance constraints without requiring overly complex manual configuration.
Data Source
Figure 1
Figure 2
Figure 3
AI summary
In one example, a memory network may control access to a shared memory that is by multiple compute nodes. The memory network may control the access to the shared memory by receiving a memory access request originating from an application executing on the multiple compute nodes and determining a priority for processing the memory access request. The priority determined by the memory network may correspond to a memory address range in the memory that is specifically used by the application.