Hypervisor Lifecycle Management in Heterogeneous Clusters
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing lifecycle managers do not support heterogeneous clusters in a datacenter, where hosts may be from different vendors, have different hardware generations or models, or use different hypervisor versions, making it challenging to manage hypervisor lifecycles without affecting running virtual machines.
Innovation Solution
A method of hypervisor lifecycle management that involves a lifecycle manager agent obtaining a desired state document defining the desired software state, including multiple images, and comparing selection criteria from the software policy against hardware information to select and apply the appropriate image to each host in the cluster.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If a lifecycle manager applies the same image to each host in a homogeneous cluster, then the management process is simple and consistent, but the system cannot support heterogeneous clusters with different hardware platforms, vendors, or hypervisor versions
Solution Approach 1:
The patent segments the cluster management approach by creating separate image definitions for different host types. Instead of a single universal image, the system divides images into homogeneous cluster images (for uniform hardware) and heterogeneous cluster images (for mixed hardware), allowing selective application based on host characteristics while maintaining manageable organization through structured image libraries
Solution Approach 2:
The patent applies local quality by tailoring image characteristics to specific host requirements. Each host or host group receives an image customized to its hardware platform, vendor, and hypervisor version, ensuring optimal compatibility and performance while allowing the lifecycle manager to handle diversity through localized customization rather than forcing uniformity
2Reliability
If the lifecycle manager applies updates to hosts, then hypervisor lifecycle management is achieved, but running virtual machines may be affected or disrupted
Solution Approach 1:
The patent implements preliminary action by preparing multiple images in advance with different levels of disruption. The system creates images that can be applied with varying impact on running VMs, allowing administrators to pre-plan update strategies and select appropriate images before applying changes, thus managing the contradiction between update ease and VM stability
Solution Approach 2:
The patent utilizes parameter changes by allowing flexible configuration of image application parameters. The lifecycle manager can adjust parameters such as applicability to specific host types, timing of updates, and disruption levels, enabling fine-tuned control over how images are applied to balance lifecycle management needs with VM runtime stability
Data Source
AI summary
An example method of hypervisor lifecycle management in a virtualized computing system having a cluster of hosts includes: obtaining, by a lifecycle manager (LCM) agent executing in a host of the hosts, a desired state document, the desired state document defining a desired state of software in the host, the software including a hypervisor, the desired state including a plurality of images; comparing selection criteria in a software policy of the desired state document against hardware information obtained from a hardware platform of the host to select an image of the plurality of images defined in the desired state document; and applying, by LCM agent, the selected image to the host.