Volume Polarization Consistency Groups Storage Failover
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Polarizing multiple volumes between storage systems can result in data loss and significant recovery times, as well as performance penalties, due to the challenges of managing IO failures and maintaining data consistency in active/active storage configurations.
Innovation Solution
The method involves dividing volumes into consistency groups with defined polarization states, using witness nodes to determine active and passive storage systems, and updating polarization states to manage IO failures at a volume granularity, thereby minimizing recovery time and preserving data consistency.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If polarizing is performed on multiple volumes between storage systems, then data consistency is maintained, but recovery time increases significantly
Solution Approach 1:
The patent segments volumes into consistency groups, allowing polarization to be applied selectively to specific groups rather than all volumes. This segmentation enables faster recovery by limiting the scope of polarization operations to only those groups affected by failures, while maintaining data consistency within each group through group-level polarization states.
Solution Approach 2:
The patent implements partial polarization by allowing only certain consistency groups to be polarized based on their state and failure conditions, rather than polarizing all volumes. This partial action approach reduces recovery time by avoiding unnecessary polarization operations on volumes that are already consistent or not affected by failures.
2Reliability
If polarizing is performed on all volumes, then data loss is prevented, but performance penalties increase
Solution Approach 1:
By dividing volumes into consistency groups, the patent enables selective polarization of only those groups that require it, preventing data loss in critical groups while avoiding performance penalties associated with polarizing all volumes. This segmentation allows the system to maintain high performance by keeping non-critical volumes in an active/active state.
Solution Approach 2:
The patent applies different polarization states to different consistency groups based on their specific requirements and failure conditions. Critical consistency groups can be fully polarized for data protection, while non-critical groups remain in active/active mode for optimal performance, achieving local quality optimization across the storage system.
3Loss of time
If individual volumes are polarized, then recovery time is reduced, but data loss may occur for groups of volumes
Solution Approach 1:
The patent segments volumes into consistency groups and applies polarization at the group level rather than individual volume level. This segmentation ensures that when polarization is triggered, it is applied consistently across all volumes in the affected group, preventing data loss due to inconsistent polarization states while maintaining relatively fast recovery by targeting only the affected group.
Solution Approach 2:
The patent combines multiple volumes into consistency groups and treats them as a unit for polarization operations. By merging volumes into groups with unified polarization states, the system ensures data consistency across all volumes in the group while enabling efficient recovery through group-level state transitions rather than individual volume operations.
4Reliability
If TTL mechanism is used for cluster maintenance, then storage system coordination is achieved, but system complexity increases
Solution Approach 1:
The patent extends the TTL mechanism to operate at the consistency group level, allowing it to serve multiple functions: coordinating polarization states across storage systems, managing failover scenarios, and maintaining cluster integrity. This universal application of TTL reduces the need for separate coordination mechanisms for different volume groups, simplifying the overall system architecture.
Data Source
AI summary
A method, computer program product, and computing system for dividing a plurality of volumes replicated across a pair of storage systems into one or more consistency groups. A polarization state may be defined for each consistency group. An input-output (IO) failure associated with at least one consistency group may be detected. At least a portion of the at least one consistency group may be polarized based upon, at least in part, the polarization state defined for the at least one consistency group.


