Storage Array API for Containerized Remote Replication
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current container orchestration platforms and CSI standards do not provide methods for containerized applications to access replication services, leading to data unavailability and inability to replicate data across storage arrays, especially during intensive workloads or disaster events, resulting in lack of load balancing and disaster recovery capabilities.
Innovation Solution
A storage array-based interface, such as an API, is provided to enable containerized applications to manage and monitor data replication services directly from a node cluster, allowing for remote replication and disaster recovery services by configuring replication metadata, establishing endpoints, and performing replication actions like failover and failback.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If containerized applications use current container orchestration platforms and CSI standards, then applications can be deployed and managed, but applications cannot access storage array replication services directly
Solution Approach 1:
The patent introduces an intermediary component (storage array interface or adapter) that bridges containerized applications and storage array replication services. This intermediary translates container orchestration platform requests into storage array-specific replication commands, enabling applications to access replication services without direct integration complexity.
Solution Approach 2:
The patent creates a universal interface layer that works across different container orchestration platforms (Kubernetes, Docker Swarm) and storage arrays. This multi-functional interface handles multiple operations (replication, snapshots, cloning) through a single standardized mechanism, improving adaptability without increasing complexity.
2Reliability
If data replication services are not available to containerized applications, then system complexity remains low, but data availability and disaster recovery capabilities are lost
Solution Approach 1:
The patent enables containerized applications to self-configure and manage their own replication services through the provided interface. Applications can independently initiate replication, configure parameters, and monitor status without requiring complex external management systems, thus improving reliability while minimizing integration complexity.
Solution Approach 2:
The patent allows replication services to be pre-configured and prepared in advance through the interface. Replication relationships, policies, and parameters can be established before disasters occur, ensuring data availability is already in place when needed, rather than requiring complex real-time setup during crises.
3Adaptability or versatility
If remote replication across storage arrays is not enabled, then system configuration is simpler, but load balancing and disaster recovery capabilities are unavailable
Solution Approach 1:
The intermediary interface simplifies the configuration of complex remote replication and disaster recovery settings by providing abstracted, high-level commands. Users configure replication through standardized interface parameters rather than dealing with low-level storage array specifics, making disaster recovery capabilities accessible without proportionally increasing operational complexity.
Data Source
AI summary
Aspects of the present disclosure relate to enabling storage array-based remote replication from containerized applications operating on one or more node clusters. In embodiments, a host executing one or more operations from a node cluster is provided an interface (e.g., an application programming interface (API)) to a storage array. Additionally, the host can be delivered resources to manage and monitor the storage array to perform one or more data replication services directly from the node cluster and via the interface. Further, data replications services are triggered in response to instructions issued by the host directly from the node cluster and via the interface.