Unified iSCSI Target Cluster Switching Fabric
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing storage systems in SAN environments face limitations in data availability and utilization, as they are typically configured to allow only one storage system to service data access requests at a time, leading to potential network overload and reduced availability due to the inability to distribute data access requests across multiple systems effectively.
Innovation Solution
A unified iSCSI target system is configured by interconnecting multiple storage systems through a cluster switching fabric, where each storage system acts as a front end to present a unified target, enabling data access requests to be distributed across multiple disk elements and managed through a replicated database for centralized lun configuration.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If multiple storage systems are interconnected to form a unified iSCSI target, then data availability and network utilization are improved, but system complexity increases
Solution Approach 1:
Multiple storage systems are merged into a single unified iSCSI target, allowing them to present a consolidated view to clients. The cluster switching fabric combines the storage resources of individual systems into a unified address space, enabling data availability across multiple systems while maintaining a simple client-facing interface.
Solution Approach 2:
The cluster switching fabric acts as an intermediary between multiple storage systems and clients. It manages the complexity of inter-system communication and coordinates data access requests, allowing storage systems to function as a unified target without increasing client-side complexity.
2Productivity
If data access requests are distributed across multiple storage systems, then network overload is reduced and availability improves, but the complexity of request routing and management increases
Solution Approach 1:
The unified iSCSI target is segmented into multiple storage systems that can independently service data access requests. The cluster switching fabric divides incoming requests across multiple systems based on data location and availability, improving productivity while managing routing complexity through centralized coordination.
Solution Approach 2:
The replicated database provides feedback mechanisms for centralized lun configuration management. Configuration changes are propagated to all storage systems, enabling coordinated request routing and data distribution without requiring complex manual management of individual system configurations.
3Stability of the object's composition
If a replicated database is used for centralized lun configuration, then configuration consistency across multiple storage systems is improved, but system complexity and data synchronization overhead increase
Solution Approach 1:
The replicated database creates copies of lun configuration data across multiple storage systems. This ensures configuration consistency by providing each system with an up-to-date view of the unified target configuration, eliminating the need for complex manual synchronization while introducing controlled data redundancy.
Data Source
AI summary
A system and method provides a unified iSCSI target using a plurality of loosely coupled iSCSI front ends. A cluster comprising a number of network elements and disk elements exports one or more logical units to iSCSI initiators. Each of the network elements is operatively interconnected with a replicated database which is utilized to share initiator data structures and lun mapping information. An iSCSI initiator may communicate with any of the network elements to access any of the logical units exported by the cluster.