Persistent Memory Pool Regulation for Runtime Volatile-Nonvolatile Switching
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Cloud computing platforms lack a mechanism to provide both volatile and non-volatile memory extensibility on a single memory class such as persistent memory (Pmem) without downtime, limiting their ability to adapt to varying workload demands.
Innovation Solution
A system that dynamically regulates volatile and non-volatile memory pools on persistent memory (Pmem) by grouping memory modules into different pools (VMP, DMP, NMP) and adjusting their sizes based on runtime workload requirements, using a cloud resource manager (CRM) and OS/hypervisor coordination.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If persistent memory is used for both volatile and non-volatile memory functions, then memory extensibility is improved, but the ability to dynamically adjust between modes without downtime is limited
Solution Approach 1:
The system dynamically adjusts memory pool configurations at runtime based on workload requirements. The cloud resource manager can modify the size and characteristics of volatile and non-volatile memory pools without requiring system downtime, enabling continuous adaptation to changing computational demands while maintaining persistent memory functionality.
2Stability of the object's composition
If memory pools are fixed at instantiation, then system stability is improved, but runtime adaptability to varying workload demands deteriorates
Solution Approach 1:
The patent implements dynamic memory pool regulation where the cloud resource manager can adjust volatile and non-volatile memory pool sizes at runtime based on monitored workload characteristics. This maintains system stability through controlled adjustments while enabling adaptability to varying computational demands without requiring complete reconfiguration or downtime.
Solution Approach 2:
The system employs feedback mechanisms where the cloud resource manager monitors workload requirements and uses this information to dynamically adjust memory pool configurations. This closed-loop control enables the system to maintain stability while adapting to changing conditions by making informed, incremental adjustments based on actual performance data.
3Productivity
If separate hardware resources are allocated for volatile and non-volatile memory, then performance is improved, but device complexity and operational costs increase
Solution Approach 1:
The patent makes persistent memory universally usable for both volatile and non-volatile memory functions through software-defined memory pools. By using a single hardware resource type (persistent memory) that can be dynamically allocated to serve different memory requirements, the system maintains high performance while reducing hardware complexity and operational costs compared to maintaining separate dedicated hardware resources for each memory type.
Data Source
AI summary
A system for managing non-transient memory in a cloud computing environment, comprising a plurality of data processors configured to cooperatively provide a cloud computing environment, a persistent memory pool system configured to interact with each of the plurality of data processors to identify persistent non-transient data memory devices at each of the data processors and a plurality of memory pools created by the persistent memory pool system, wherein each of the plurality of memory pools has a designated function.