Kubernetes Storage Targets Bridging Cloud Block and On-Premises I/O
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Legacy virtual machine and hypervisor approaches are lagging in migrating to the Kubernetes environment, hindering the adoption of container-as-a-service platforms, and there is a need for improved technologies to provide cloud-native storage and self-service tools with centralized data management.
Innovation Solution
Deploying a cluster node controller as a cloud-native executable container that functions as a storage target, utilizing a storage target interface, and transforming virtualization system components into cloud-native executable containers to interface with Kubernetes environments.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If legacy virtual machine and hypervisor approaches are used, then existing virtualization infrastructure can be maintained, but migration to Kubernetes environment is hindered and adoption of container-as-a-service platforms is slowed
Solution Approach 1:
The patent transforms the virtualization control plane from a legacy VM-based architecture to a container-native architecture by changing the fundamental parameters of how controllers are deployed and managed. This involves converting controller functionality from hypervisor-dependent VMs to Kubernetes-compatible containers, enabling the system to adapt to modern container-as-a-service platforms while maintaining core virtualization capabilities through software-defined approaches.
Solution Approach 2:
The patent segments the virtualization control plane into independent, containerized controller components that can be deployed and managed separately within the Kubernetes environment. This segmentation allows each controller function to be encapsulated in its own container, facilitating gradual migration and reducing the complexity of transitioning from legacy VM-based architectures to container-native approaches.
2Reliability
If cloud-native executable containers are deployed as storage targets, then dependencies on public cloud providers are reduced and performance is enhanced, but deployment complexity increases
Solution Approach 1:
The patent creates universal storage target containers that can operate independently of specific cloud providers while maintaining compatibility with Kubernetes environments. These containerized storage targets implement standardized interfaces and protocols, enabling them to function as drop-in replacements for cloud-provider-specific storage services. This universality allows organizations to reduce dependencies on public cloud providers while leveraging the portability and scalability of container technology.
3Productivity
If virtualization system components are transformed into cloud-native executable containers, then scalability is enhanced, but technical challenges of migration increase
Solution Approach 1:
The patent transforms static, VM-based virtualization controllers into dynamic, containerized components that can be easily scaled, updated, and managed within Kubernetes. The containerized controllers leverage Kubernetes' native capabilities for dynamic resource allocation, automated orchestration, and horizontal scaling. This dynamic approach enables the virtualization system to adapt to changing workload demands in real-time while simplifying the migration process through standardized container images and deployment manifests.
Data Source
Figure 1A1
Figure 1A2
Figure 1B
AI summary
Methods, systems, and computer program products. For implementing a container-attached storage facility. A cloud-resident containerized control module is situated in a Kubernetes Pod. A plurality of storage devices are organized into a common access space and made accessible by the cloud-resident containerized control module. In operation, and responsive to receiving a storage I/O request referencing a portion of storage that is addressable via the common access space, the cloud-resident containerized control module redirects the storage I/O request to a storage device of the common access space. Some instances of storage I/Os are redirected toward a cloud-provided block storage facility. Some instances of storage I/Os are redirected to storage devices situated in an on-premises environment.