Managed Node Composition Using Coherence Data Across Compute Sleds
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
In data centers, workloads often underutilize or overutilize available processors, leading to resource wastage and inefficiencies in cloud services due to incorrect resource allocation, resulting in financial and time losses.
Innovation Solution
The implementation of a system that dynamically composes managed nodes with multiple processors from multiple compute sleds, allowing for cooperative execution of workloads across distributed resources, using an orchestrator server and accelerator to manage coherence data and bandwidth, enabling efficient utilization of pooled resources.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Speed
If multiple processors are physically located on the same compute device with shared memory and local bus, then communication speed and coordination efficiency are improved, but resource utilization deteriorates due to underutilization of available processors
Solution Approach 1:
The system segments the compute device into multiple compute sleds, each containing one or more processors. This segmentation allows independent management and allocation of processor resources across different workloads, enabling better resource utilization while maintaining communication efficiency through the interconnect fabric that connects the sleds.
Solution Approach 2:
The patent introduces an intermediary layer (the system architecture with interconnect fabric and coordination mechanisms) that enables processors on different compute sleds to communicate and coordinate as if they were local. This intermediary allows remote processors to be effectively utilized while maintaining communication efficiency comparable to local processors.
2Productivity
If more processors are allocated to a workload to improve execution efficiency, then quality of service is improved, but resource wastage increases due to incorrect resource matching
Solution Approach 1:
The system dynamically composes managed nodes by selectively activating and deactivating compute sleds based on workload requirements. This dynamic approach allows the system to match processor allocation to actual workload needs, improving execution efficiency while preventing resource wastage through on-demand resource activation.
Solution Approach 2:
The patent creates a universal resource pool where compute sleds can be allocated to different workloads as needed. The same physical processors can serve multiple different workloads sequentially or concurrently through virtualization and dynamic composition, eliminating resource wastage while maintaining high execution efficiency across diverse workloads.
3Adaptability or versatility
If processors are distributed across multiple compute sleds to increase available resources, then workload scalability is improved, but system complexity increases due to coordination requirements
Solution Approach 1:
The system uses coordination data structures and coherence protocols that replicate the behavior of local memory access patterns across distributed sleds. By copying the coordination mechanisms used in single-sled systems to multi-sled configurations, the patent enables scalable processor distribution while maintaining manageable system complexity through familiar coordination patterns.
Data Source
AI summary
Technologies for composing a managed node with multiple processors on multiple compute sleds to cooperatively execute a workload include a memory, one or more processors connected to the memory, and an accelerator. The accelerator further includes a coherence logic unit that is configured to receive a node configuration request to execute a workload. The node configuration request identifies the compute sled and a second compute sled to be included in a managed node. The coherence logic unit is further configured to modify a portion of local working data associated with the workload on the compute sled in the memory with the one or more processors of the compute sled, determine coherence data indicative of the modification made by the one or more processors of the compute sled to the local working data in the memory, and send the coherence data to the second compute sled of the managed node.