Bit Rot Detection in Distributed Storage via Syndrome Extraction

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Conventional methods for detecting bit rot in distributed storage devices are inefficient, requiring large network bandwidth and long detection times, as they send all data and parity chunks for remote analysis, whereas existing solutions do not effectively address bit rot detection in distributed storage devices with failure domains.

Innovation Solution

A method and system that receive combinations of data and parity chunks from different stripes within failure domains, create new parity chunks, and identify bit rot by comparing these new chunks to the received parity chunks, reducing network traffic and enabling faster detection.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Reliability

If all data chunks and parity chunks are sent over networks for remote analysis, then bit rot detection can be performed, but network bandwidth consumption increases and detection time increases

Engineering Contradiction:
Improvebit rot detection capabilityVSAvoidnetwork bandwidth consumption
Core Design Contradiction:
ReliabilityVSLoss of energy

Solution Approach 1:

The patent extracts only the essential components needed for bit rot detection by having distributed storage devices compute and send only syndrome values (parities of parities) rather than all data and parity chunks. This extraction principle reduces network bandwidth consumption while maintaining detection capability.

Inventive Principle:
Principle #2Taking out (Extraction)

Solution Approach 2:

The patent applies preliminary action by having distributed storage devices pre-compute syndrome values before transmission. The storage devices perform local computations to generate parity-of-parity values, so that when data is sent to the computing device, only the pre-processed syndrome values need to be transmitted, reducing network bandwidth requirements.

Inventive Principle:
Principle #10Preliminary action

2Reliability

If all data chunks and parity chunks are sent over networks for remote analysis, then bit rot detection can be performed, but detection time increases

Engineering Contradiction:
Improvebit rot detection capabilityVSAvoiddetection time
Core Design Contradiction:
ReliabilityVSLoss of time

Solution Approach 1:

The patent extracts only the syndrome values needed for detection, eliminating the need to transmit and process all original data chunks and parity chunks. This extraction of essential detection elements significantly reduces detection time while maintaining reliability.

Inventive Principle:
Principle #2Taking out (Extraction)

Solution Approach 2:

By performing preliminary computation of syndrome values at the distributed storage devices before transmission, the patent reduces the amount of data that needs to be processed remotely, thereby reducing overall detection time while preserving detection accuracy.

Inventive Principle:
Principle #10Preliminary action

3Reliability

If conventional remote analysis methods are used, then bit rot can be detected, but large quantities of network bandwidth are required

Engineering Contradiction:
Improvedata integrity verificationVSAvoidnetwork bandwidth quantity
Core Design Contradiction:
ReliabilityVSQuantity of substance

Solution Approach 1:

The patent extracts only the critical syndrome values (parities of parities) from the complete data set, transmitting only these extracted elements for verification. This dramatically reduces the quantity of data transmitted over the network while maintaining the ability to verify data integrity.

Inventive Principle:
Principle #2Taking out (Extraction)

Solution Approach 2:

The distributed storage devices perform preliminary computation to generate condensed syndrome values before transmission, reducing the volume of data that must traverse the network while preserving the capability to perform comprehensive data integrity verification.

Inventive Principle:
Principle #10Preliminary action

Data Source

PatentEP3769215B1Systems and methods for detecting bit rot in distributed storage devices having failure domains
Publication Date: 2024.05.01 VERITAS TECHNOLOGIES LLC
  • EP3769215B1 patent drawingFigure 1
  • EP3769215B1 patent drawingFigure 2
  • EP3769215B1 patent drawingFigure 3

AI summary

The disclosed computer-implemented method for detecting bit rot in distributed storage devices having failure domains may include (1) receiving, at a computing device, a combination of data chunks created by adding, for each failure domain storing data chunks, data chunks from different stripes, (2) receiving a combination of parity chunks created by adding, for each failure domain storing parity chunks, parity chunks from the different stripes, (3) creating at least one new parity chunk from the combination of data chunks, and (4) identifying a presence of bit rot in the failure domains when the at least one new parity chunk does not match the combination of parity chunks. Various other methods, systems, and computer-readable media are also disclosed.