VM Placement Isolating Physical Processors to Reduce Cache Pollution
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
In multi-processor servers, the co-location of different physical processor types on the same chip leads to cache pollution and performance penalties due to shared cache usage, which complicates the optimization of logical partition placement for virtual machines.
Innovation Solution
A processor placement program builds bit masks for each type of physical processor and identifies solution spaces to isolate each type of processor onto specific containers, ensuring minimal overlap and optimizing their placement across the server system.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If different physical processor types are co-located on the same chip, then device complexity is reduced and resource utilization is improved, but cache pollution occurs and performance deteriorates due to shared cache usage
Solution Approach 1:
The patent segments the physical processors into distinct types and assigns each type to specific container groups identified through combinatorial search. This segmentation prevents different processor types from sharing the same cache resources, eliminating cache pollution while maintaining placement flexibility through the systematic assignment method.
Solution Approach 2:
The patent applies local quality by ensuring that each container group is dedicated to a specific processor type, creating localized homogeneous environments. This means different regions of the server (different container groups) have different specialized processor types, preventing cross-contamination of cache resources while optimizing for specific workload requirements.
2Productivity
If physical processors are isolated by type onto specific containers, then cache pollution is reduced and performance is improved, but the complexity of processor placement optimization increases
Solution Approach 1:
The patent performs preliminary action by pre-identifying container groups that can accommodate specific processor types through combinatorial search before actual processor placement. This advance preparation creates a structured framework that simplifies the subsequent placement process, reducing the apparent complexity while ensuring optimal performance outcomes.
Solution Approach 2:
The patent uses bit masks as abstract representations (copies) of container capabilities and processor requirements. These bit masks simplify the complex placement problem by transforming it into a manageable mathematical representation, making the optimization process more tractable while maintaining the integrity of the physical constraints.
3Manufacturing precision
If combinatorial search is used to identify optimal container groups for each processor type, then placement precision is improved, but computational time increases
Solution Approach 1:
The patent changes the parameter representation by using bit masks to encode container capabilities and processor requirements. This transformation converts a complex multi-dimensional optimization problem into a more manageable form that can be solved more efficiently while maintaining high placement precision through the systematic combinatorial search approach.
Data Source
AI summary
In an approach for isolating physical processors during optimization of virtual machine placement, a server is provided comprising a plurality of containers and a plurality of physical processors. A processor builds a set of bit masks for each type of physical processor required for a logical partition. A processor builds a set of solution spaces based on the plurality of containers and an amount of each type of container of the plurality of containers. A processor completes a combinatorial search of the set of bitmasks and the set of solution spaces. A processor identifies a solution space of the set of solution spaces for the logical partition. The physical and logical configuration of the server is changed based on the solution space for the first logical partition.