Dynamic Memory Distribution for Virtual Machine Systems
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
In information handling systems, memory module failures necessitate shutting down all virtual machines for replacement, disrupting system operations and requiring extensive downtime.
Innovation Solution
A method for dynamic memory distribution, where unused system physical addresses are reassigned from other virtual machines to the affected VM, allowing memory module space adjustments without shutting down functioning VMs, utilizing a hypervisor to track and recalibrate memory usage.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If all virtual machines are shut down for memory module replacement, then the failed memory module can be replaced, but system downtime increases and productivity decreases
Solution Approach 1:
The patent divides the system physical address space into multiple sections, with each section assigned to a specific virtual machine. When a memory module fails, only the section associated with that module is affected, while other sections and their corresponding virtual machines continue to operate normally. This segmentation allows selective isolation of the failed component without bringing down the entire system.
Solution Approach 2:
The patent implements dynamic reassignment of system physical address sections. When a memory module failure is detected, the hypervisor dynamically reallocates the affected address section from the impacted virtual machine to another virtual machine that can utilize it. This dynamic adjustment enables continuous system operation during memory module replacement, eliminating the need for complete system shutdown.
2Device complexity
If system physical addresses are statically assigned to virtual machines, then address allocation is simple, but memory utilization efficiency decreases when failures occur
Solution Approach 1:
The patent incorporates a feedback mechanism where the hypervisor continuously monitors the health status of memory modules. When a failure is detected, the system responds by triggering the dynamic reassignment process. This feedback loop enables the system to adapt to changing conditions and optimize memory utilization by reallocating address sections from failed to operational modules, thereby improving overall memory efficiency without increasing complexity.
Data Source
AI summary
Dynamic distribution of memory, including identifying memory modules; creating a system physical address (SPA) of the memory modules; assigning, for each virtual machine (VM), a respective section of the SPA to the VM; calculating, for each VM, portions of the respective section of the SPA for the VM that is being used by the VM and that is not being used by the VM; identifying a physical failure of a particular memory module; in response to identifying the physical failure: identifying a particular VM assigned to the section of the SPA associated with the particular memory module that has physically failed; accumulating, for each other VM, the unused portions of the respective SPA for the VM; marking, for each other VM, the unused portion of the SPA for the VM as read-only for the VM; and reassigning a portion of the unused portions of the SPA to the particular VM.