Distributed Logical Unit Architecture for Clustered Storage Consistency
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Maintaining consistency across distributed objects in a Storage Area Network (SAN)-attached clustered storage system becomes challenging as the number of nodes increases, making it difficult to manage and control resources effectively.
Innovation Solution
Implementing a distributed logical unit architecture with a SCSI target instance that coordinates operations across cluster nodes using a cluster transaction manager and quorum manager, ensuring consistent context information is maintained through a global task set and transaction manager, and utilizing a single system image to present a unified view to hosts.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If the number of nodes in a cluster increases to improve scalability, then storage capacity and reliability improve, but maintaining consistent context information across nodes becomes more difficult
Solution Approach 1:
The patent introduces a quorum manager as an intermediary component that coordinates context information updates across all cluster nodes. The quorum manager receives update requests, validates them against the current system image, and distributes approved updates to all nodes, ensuring consistency without requiring direct peer-to-peer coordination between nodes.
Solution Approach 2:
The system implements feedback mechanisms where nodes report their context information state to the quorum manager, which then determines whether updates should be propagated. The quorum manager tracks the system image state and provides feedback to nodes about whether their context information is current, enabling automatic synchronization.
2Quantity of substance
If distributed architecture is used to improve scalability, then storage capacity increases, but resource management and control operations become more complex
Solution Approach 1:
The patent merges the resource management functions of multiple distributed nodes into a unified system image managed by the quorum manager. All nodes present a single, consistent view of storage resources to clients, combining the physical distribution of data with logical centralization of control for simplified resource management.
Solution Approach 2:
The quorum manager serves multiple functions simultaneously: it acts as a transaction coordinator, context information repository, update validator, and system image manager. This multi-functional approach consolidates resource management operations into a single component that can handle various management tasks across the distributed cluster.
3Reliability
If clustered storage system is implemented to improve reliability, then data availability increases, but maintaining a consistent single system image becomes more challenging
Solution Approach 1:
The system performs preliminary validation of context information updates against the current system image before applying them. The quorum manager checks whether proposed updates are consistent with the existing system state and only approves updates that maintain consistency, preventing divergent system images from developing across nodes.
Solution Approach 2:
The quorum manager maintains a master copy of the system image and distributes copies to all cluster nodes. When updates are approved, the quorum manager propagates the updated copy to all nodes, ensuring that every node has an identical copy of the system image, thereby maintaining precision and consistency.
Data Source
AI summary
Consistent distributed storage communication protocol semantics, such as SCSI target semantics, in a clustered storage system are disclosed. The system includes a mechanism for presenting a single distributed logical unit, comprising one or more logical sub-units, as a single logical unit of storage to a host system by associating each of the logical sub-units that make up the single distributed logical unit with a single host visible identifier that corresponds to the single distributed logical unit. The system further includes a mechanism to maintain consistent context information for each of the logical sub-units such that the logical sub-units are not visible to a host system as separate entities from the single distributed logical unit.