Host-Controlled Service-Level Storage Allocation Across VM Datastores
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing storage resource allocation techniques in information processing systems fail to effectively manage service levels for virtual machines using different datastores, leading to inefficiencies and inability to enforce priority levels across multiple datastores, host clusters, and operating systems.
Innovation Solution
Implementing a host-controlled service level management model that utilizes a service level extraction and enforcement mechanism across multiple datastores, host clusters, and operating systems, allowing administrators to control storage resources based on assigned service levels for virtual machines.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If storage resources are controlled using IO thresholds assigned to each VM within a single datastore, then storage resource allocation can be managed for that specific datastore, but service level control cannot be enforced across multiple datastores used by different VMs
Solution Approach 1:
The patent implements a universal service level management system that functions across multiple datastores, host clusters, and operating systems. The storage appliance serves multiple purposes: it acts as a storage device, a service level enforcement point, and a communication bridge between virtualization environments and physical storage resources. This multi-functionality enables consistent service level control regardless of which datastore or host cluster the VM is using.
Solution Approach 2:
The storage appliance acts as an intermediary component between the virtualization environment (hypervisor/VM) and the physical storage resources. It receives service level information from the virtualization layer, translates it into storage-specific control parameters, and enforces these controls at the storage array level. This intermediary role allows service level management to span across multiple datastores without requiring changes to the virtualization infrastructure.
2Ease of operation
If storage resource control is implemented at the datastore level, then allocation can be managed for individual datastores, but administrators with limited access cannot control storage arrays
Solution Approach 1:
The system enables self-service storage management where administrators with limited virtualization access can still control storage resources. The storage appliance provides a self-contained control mechanism that allows these administrators to define and enforce service levels directly at the storage array without requiring elevated access to the virtual control station. This self-service capability maintains service level reliability while improving ease of operation.
3Adaptability or versatility
If service levels are assigned to VMs using a single virtual control station, then resource allocation can be managed within that control domain, but cross-cluster and cross-OS service level management is not available
Solution Approach 1:
The storage appliance implements a universal service level management interface that accepts service level definitions from multiple virtual control stations and operating systems. It maintains a consistent service level policy database that can be accessed and enforced across different host clusters and OS types, preventing information loss by providing a single source of truth for service level configurations.
Solution Approach 2:
The system segments service level management into distinct functional components: service level definition (at the virtualization layer), service level translation (at the storage appliance), and service level enforcement (at the storage array). This segmentation allows each component to operate independently across different clusters and OSs while maintaining overall consistency, as the storage appliance acts as the central coordination point that harmonizes service level information from diverse sources.
Data Source
AI summary
Methods, apparatus, and processor-readable storage media for implementing host-controlled service levels are provided herein. An example computer-implemented method includes obtaining, at a storage node of a storage system, information comprising one or more identifiers and one or more service levels that are assigned to virtual machines identified by the one or more identifiers; extracting one of the one or more identifiers from an input-output operation received at the storage node; identifying, based on the extracted identifier, the service level that is assigned to the virtual machine that initiated the input-output operation; and controlling storage resources allocated for one or more additional input-output operations from one or more of the virtual machines based at least in part on the identified service level.