Dynamic Partitioning for SAN Management Information Scalability
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional partitioning in storage area networks (SANs) faces scalability issues and backward compatibility challenges due to fixed partition sizes, which grow proportionally with storage capacity and require retroactive updates when SAN configurations change.
Innovation Solution
Dynamic partitioning allows for range-based partitioning of management information, where partitions are defined by a range of storage objects rather than static IDs, enabling adaptive partition sizes that adjust with changes in storage arrays, avoiding the persistence of fixed partition IDs and improving scalability and compatibility.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If fixed partition sizes are used for managing management information, then the partitioning structure is simple and stable, but the partitions grow proportionally with storage capacity causing scalability issues and requiring retroactive updates when SAN configurations change
Solution Approach 1:
The patent implements dynamic partitioning where partition sizes are not fixed but adapt automatically based on the volume of management information generated by storage objects. Each partition's size is determined by the actual data volume it contains rather than a predetermined fixed size, allowing partitions to grow or shrink dynamically as storage capacity and management information volume change, thereby resolving the scalability issue without requiring system reconfiguration
Solution Approach 2:
The system changes the partitioning parameter from fixed size to variable size based on management information volume. By making partition size a dynamic parameter that adjusts according to the actual data volume and storage object count, the system achieves scalability while maintaining simple partition management logic through automatic parameter adjustment rather than manual reconfiguration
2Adaptability or versatility
If fixed partition IDs are persisted with management information, then data consistency is maintained, but retroactive updates are required when SAN configurations change reducing compatibility
Solution Approach 1:
The patent makes partition IDs dynamic rather than fixed, allowing them to change automatically when SAN configurations change. Instead of persisting fixed partition IDs that require retroactive updates, the system generates new partition IDs based on current storage object assignments, eliminating the need for time-consuming retroactive updates and maintaining backward compatibility automatically
Solution Approach 2:
The system performs preliminary assignment of storage objects to partitions before configuration changes occur. By pre-establishing partition assignments based on expected storage object distributions, the system avoids the need for retroactive updates when configurations change, as the dynamic partitioning logic automatically adapts to new configurations without requiring post-change reconfiguration
3Reliability
If large data sets of management information are processed by the management application, then complete monitoring coverage is achieved, but bottlenecks and disproportionate demands on computing resources occur
Solution Approach 1:
The patent divides the large data set of management information into multiple partitions, each containing management information for a subset of storage objects. This segmentation allows the management application to process partitions independently and in parallel, reducing bottlenecks and computing resource demands while maintaining complete monitoring coverage across all storage objects in the SAN
Solution Approach 2:
The system processes management information in partial data sets corresponding to individual partitions rather than processing the entire large data set at once. By handling only the necessary subset of management information for each partition independently, the system achieves efficient processing with reduced computing resource demands while collectively maintaining complete monitoring coverage across all partitions
Data Source
AI summary
A SAN management agent performs range based partitioning according to subranges of a set of storage devices. Partitions expressed as a range of storage objects (e.g. storage devices), based on device Ids, rather than static partition labels, allows the partitions to dynamically change to suit a current SAN configuration and number of storage devices in the storage arrays. Typically, management information is in the form of updates to a particular storage object. Accordingly, each update record in an update set is attributable to a device ID of the storage object to which it pertains. Identification of partitions by a range of storage devices allows the update records to be indexed, or grouped, to a particular partition without persisting a fixed partition ID with the data. The device ID range defines a partition that remains a logical quantity which may change upon SAN reconfiguration by avoiding persisting a fixed partition ID to the update records.