Offset Erasure Code Stripes for Correlated Disk Failure Reduction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional erasure code implementations on optical disks are overwhelmed by correlated failures due to standard erasure code stripes including strips from high failure rate regions, leading to increased data loss risks.
Innovation Solution
The method involves selecting erasure code stripes to include at most one strip from high failure rate regions on each storage unit, organizing these strips to be offset from other strips, and mapping high failure rate regions to sequentially numbered strips, while mapping low failure rate regions to additional sequentially numbered strips, thereby reducing the number of high failure rate regions in each stripe.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If standard erasure code stripes are used including strips from the same location in each optical disk, then the erasure code implementation is simple, but the reliability is reduced due to correlated failures from high failure rate regions
Solution Approach 1:
The patent applies local quality by treating different regions of optical disks differently based on their failure characteristics. High failure rate regions are identified and handled separately from low failure rate regions through selective strip placement. The system places at most one strip from high failure rate regions in each erasure code stripe, while allowing multiple strips from low failure rate regions, thereby adapting the erasure code structure to local reliability characteristics of different disk regions.
Solution Approach 2:
The patent segments the erasure code stripe into two distinct components: strips from high failure rate regions and strips from low failure rate regions. By separating these segments and applying different placement rules to each, the system manages complexity while improving reliability. The high failure rate region strips are carefully controlled to avoid correlation, while low failure rate region strips are freely included to maintain storage capacity.
2Reliability
If multiple strips from high failure rate regions are included in the same erasure code stripe, then the storage capacity is maximized, but the reliability is overwhelmed by correlated failures
Solution Approach 1:
The patent applies preliminary anti-action by proactively preventing correlated failures before they can occur. The system预先 identifies high failure rate regions on each optical disk and implements a placement rule that limits at most one strip from such regions per erasure code stripe. This preemptive measure counteracts the tendency toward correlated failures, ensuring that even if multiple disks have high failure rate regions, their failures will not be correlated within the same stripe.
3Reliability
If strips are selected without considering failure rate regions, then the data access time is fast, but the data loss risk increases due to systematic failures
Solution Approach 1:
The patent applies preliminary action by pre-identifying and cataloging high failure rate regions on each optical disk before constructing erasure code stripes. The system performs an initial assessment of disk reliability characteristics, marks problematic regions, and then uses this pre-computed information to guide strip selection. This preliminary characterization eliminates the need for complex real-time analysis during stripe construction, maintaining efficiency while improving reliability.
Data Source
AI summary
A computer-implemented method, according to one embodiment, includes: selecting strips from each storage unit for a given erasure code stripe such that the given erasure code stripe includes at most one strip from a high failure rate region of the respective storage unit, where each of the storage units include high and low failure rate regions. The selected strips are organized such that a number of each strip in the given erasure code stripe is offset from the remaining strips by an amount that is greater than a total number of strips in the high failure rate regions. The organized selected strips are further mapped to form the given erasure code stripe such that the high failure rate regions on each storage unit are mapped to one or more sequentially numbered strips, and the low failure rate regions are mapped to additional sequentially numbered strips.