Fault-Tolerant Data Storage Partitioning

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Large-scale data storage systems face challenges in fault tolerance and recovery due to the risk of hardware failures, which can lead to long downtime and potential loss of data, especially when data and metadata are spread across numerous devices.

Innovation Solution

The system partitions a large data collection into multiple independent, fault-contained partitions, allowing for localized failure management and quick recovery of affected partitions using a restore-at-replica process, where each partition includes metadata for the entire file system, enabling self-identification and independent operation.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Quantity of substance

If data is stored as a single large collection across many storage devices, then storage capacity and deduplication efficiency are improved, but fault tolerance deteriorates and recovery time increases

Engineering Contradiction:
Improvestorage capacityVSAvoidfault tolerance
Core Design Contradiction:
Quantity of substanceVSReliability

Solution Approach 1:

The patent divides the large storage collection into multiple independent partitions, each capable of operating autonomously. When a failure occurs, only the affected partition needs to be recovered while others remain accessible, thus maintaining fault tolerance without sacrificing storage capacity. This segmentation allows the system to scale storage while preserving reliability through localized failure containment.

Inventive Principle:
Principle #1Segmentation

2Quantity of substance

If data is stored as a single large collection, then deduplication efficiency is improved, but recovery time increases

Engineering Contradiction:
Improvededuplication efficiencyVSAvoidrecovery time
Core Design Contradiction:
Quantity of substanceVSLoss of time

Solution Approach 1:

The storage system is divided into partitions that can be independently recovered. When a failure occurs, only the affected partition needs to be restored rather than the entire collection, dramatically reducing recovery time while maintaining deduplication efficiency within each partition.

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

The patent implements restore-at-replica capability where backup copies of partitions are maintained in advance. When a partition fails, the system can quickly restore it from the replica without needing to recover the entire collection, reducing recovery time while preserving deduplication benefits.

Inventive Principle:
Principle #10Preliminary action

3Quantity of substance

If more storage devices are added to increase capacity, then storage scalability is improved, but likelihood of failure increases

Engineering Contradiction:
Improvestorage scalabilityVSAvoidfailure probability
Core Design Contradiction:
Quantity of substanceVSReliability

Solution Approach 1:

By organizing storage into independent partitions, the patent allows the system to scale by adding more devices while containing failure risk within individual partitions. Each partition operates independently, so failures in one partition do not propagate to others, enabling scalability without proportionally increasing system-wide failure probability.

Inventive Principle:
Principle #1Segmentation

Data Source

PatentUS8713282B1Large scale data storage system with fault tolerance
Publication Date: 2014.04.29 EMC IP HLDG CO LLC
  • US8713282B1 patent drawing
  • US8713282B1 patent drawing
  • US8713282B1 patent drawing

AI summary

According to one embodiment, a large scale data storage system with fault tolerance is described. In one example, a system includes a plurality of storage partitions, each partition including a portion of a large scale data store, and metadata corresponding to the portion of the large scale data store. At least one partition includes namespace metadata for the large scale data store including namespace data for a portion of the large scale data store outside the at least one partition.