Dynamically Composable Computing System Resource Manager
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current data center architectures lack adaptability, configurability, and extensibility to manage increasingly complex and dynamic computing resources and workloads, leading to inefficiencies in resource allocation and performance.
Innovation Solution
The implementation of a dynamically composable computing system that allocates and manages disaggregated computing resources, including memory acceleration logic, to create composite computing nodes that can be configured based on workload requirements, using a resource manager module to coordinate hardware resource sharing and optimization.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If computing resources are statically allocated in traditional data center architectures, then system stability is maintained, but adaptability and configurability to manage complex and dynamic computing resources deteriorate
Solution Approach 1:
The patent implements dynamic resource allocation where computing resources (CPUs, memory, storage, network devices) are allocated and reconfigured based on real-time workload demands. The system transitions from static to dynamic architecture, allowing resources to be composed and decomposed automatically according to service chain requirements and performance conditions
Solution Approach 2:
The patent creates a universal resource pool where diverse computing resources can serve multiple functions and workloads. The same physical resources can be dynamically assigned to different virtual machines and service chains, enabling one resource to perform multiple roles based on demand
2Adaptability or versatility
If computing resources are disaggregated into separate pools, then configurability improves, but system complexity increases
Solution Approach 1:
The patent introduces an intermediary layer (virtualization layer/resource manager) that manages the disaggregated resource pools. This intermediary handles the complexity of allocating and coordinating resources from separate pools (compute, memory, storage, network) to workloads, shielding users from underlying complexity while enabling fine-grained configurability
3Productivity
If virtual elements are ordered to form service chains, then service functionality is achieved, but performance bottlenecks increase due to differing performance limitations of individual elements
Solution Approach 1:
The patent implements feedback mechanisms that monitor performance metrics of virtual elements in service chains. Based on this feedback, the system dynamically identifies bottleneck elements and reallocates resources or reconfigures service chains to balance performance across all elements, ensuring consistent service-level performance
4Productivity
If computing resources are consolidated in traditional architectures, then resource allocation simplicity is maintained, but consolidation density and efficiency deteriorate
Solution Approach 1:
The patent segments computing resources into fine-grained units that can be independently allocated and managed. This segmentation enables high consolidation density by packing multiple workloads onto shared physical infrastructure while maintaining the ability to manage and allocate resources with precision, overcoming the limitations of traditional coarse-grained consolidation
Data Source
AI summary
The present disclosure relates to a dynamically composable computing system. The dynamically composable computing system comprises at least one compute sled including a set of respective local computing hardware resources; a plurality of disaggregated memory modules; at least one disaggregated memory acceleration logic configured to perform one or more predefined computations on data stored in one or more of the plurality of disaggregated memory modules; and a resource manager module configured to assemble a composite computing node by associating, in accordance with requirements of a user, at least one of the plurality of disaggregated memory modules with the disaggregated memory acceleration logic to provide at least one accelerated disaggregated memory module and connecting the least one accelerated disaggregated memory module to the compute sled.