Uniform Host Attachment via Extended Identifier Ranges
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current data storage systems face challenges in efficiently managing and replicating data across multiple hosts and data storage systems, particularly in maintaining consistent access paths and identifier ranges for stretched volumes in metro cluster configurations, which can lead to identifier collisions and inefficient data access.
Innovation Solution
The technique involves configuring data storage systems with unique identifier sets for target ports and port groups, using normal and extended identifier ranges, and managing access path states to ensure consistent exposure of stretched volumes to hosts, allowing for efficient data replication and access across multiple systems.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of manufacture
If multiple data storage systems use the same identifier range for target ports and port groups, then configuration simplicity is maintained, but identifier collisions occur causing access conflicts
Solution Approach 1:
The identifier space is segmented into normal identifier range (0-4095) and extended identifier range (4096-65535). Data storage systems can be configured with either normal or extended identifier sets, allowing multiple systems to coexist without collisions while maintaining manageable configuration complexity through this structured division.
Solution Approach 2:
The patent introduces an additional dimension to the identifier space by adding extended identifier ranges beyond the traditional normal range. This dimensional expansion allows systems to differentiate their identifiers without changing the fundamental identifier structure, resolving collisions while preserving compatibility.
2Reliability
If stretched volumes are exposed to hosts over multiple paths with different identifier sets, then data access redundancy is improved, but path management complexity increases
Solution Approach 1:
The data storage system is designed to handle multiple identifier sets (normal and extended) through a unified target port and target port group structure. This universal approach allows the system to manage multiple access paths with different identifier configurations without requiring separate management mechanisms for each path type.
Solution Approach 2:
The target port group acts as an intermediary layer between the host and multiple target ports with different identifier sets. By grouping target ports and managing them collectively, the system simplifies path management while maintaining the ability to expose stretched volumes over multiple paths with different identifier configurations.
3Reliability
If extended identifier ranges are used for target ports and port groups, then identifier collision prevention is achieved, but identifier management complexity increases
Solution Approach 1:
The patent changes the identifier parameter range from the traditional normal range (0-4095) to an extended range (4096-65535) when needed. This parameter change allows systems to avoid identifier collisions by using a larger address space, while the underlying management mechanisms remain compatible with existing systems through the dual-range approach.
Data Source
AI summary
Two data storage systems, DS1 and DS2, may be initially configured with identifiers for target ports and target port groups. Subsequently, the two system may be combined into a cluster including a stretched volume configured from the volumes V1 and V2, respectively, on DS1 and DS2, where V1 and V2 are exposed to the host as the same logical volume, L1, over multiple paths from DS1 and DS2 to the host. V1 may have a normal attribute indicating target ports and port groups of DS1 have associated identifiers as specified in an initial configuration when reporting information regarding L1 to the host. V2 may have an extended attribute indicating that target ports and port groups of DS2 have associated extended identifiers determined using a first extended value and using identifiers from an initial configuration when reporting information regarding L1 to the host.