XOR Chain Topology for Distributed Data Protection

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Cloud-based data storage systems with geo-replication and XOR techniques face excessive network traffic load on single zones during disaster recovery, which can overwhelm backend zones and lead to inefficient workload distribution.

Innovation Solution

Implementing a chain topology for XOR chunk creation and recovery, where each frontend zone generates and updates XOR chunks, distributing processing and network traffic evenly across zones, and storing the final XOR chunk in a backend zone that does not host user data.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Reliability

If conventional XOR implementation is used with geo-replication, then data protection is achieved, but excessive network traffic load concentrates on single backend zones during disaster recovery

Engineering Contradiction:
Improvedata protectionVSAvoidnetwork traffic load
Core Design Contradiction:
ReliabilityVSObject-generated harmful factors

Solution Approach 1:

The patent divides the XOR chunk creation process into multiple segments, with each frontend zone responsible for creating and updating XOR chunks for its own data. This segmentation distributes the network traffic load across multiple frontend zones instead of concentrating it on a single backend zone, resolving the contradiction between maintaining data protection reliability and reducing harmful network traffic concentration.

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

The patent changes the architectural dimension by moving XOR chunk creation from backend zones to frontend zones. Instead of having backend zones centrally manage XOR operations, the system distributes this function across the frontend zone dimension, where each zone independently creates and maintains its own XOR chunks, thereby distributing network traffic across multiple zones rather than concentrating it.

Inventive Principle:
Principle #17Another dimension (Dimensionality change)

2Ease of operation

If backend zones store XOR chunks, then data recovery is enabled, but workload distribution becomes uneven and backend zones get overwhelmed

Engineering Contradiction:
Improvedata recoveryVSAvoidworkload distribution
Core Design Contradiction:
Ease of operationVSProductivity

Solution Approach 1:

The patent implements self-service by enabling each frontend zone to independently create, update, and manage its own XOR chunks without requiring backend zone intervention. This self-service approach allows frontend zones to handle their own data recovery operations using locally stored XOR chunks, eliminating the need for backend zones to process recovery requests and thereby achieving even workload distribution while maintaining ease of data recovery.

Inventive Principle:
Principle #25Self-service

3Quantity of substance

If XOR chunks are centralized in backend zones, then capacity efficiency is maintained, but network traffic overload occurs during disaster recovery

Engineering Contradiction:
Improvestorage capacity efficiencyVSAvoidnetwork traffic overload
Core Design Contradiction:
Quantity of substanceVSObject-generated harmful factors

Solution Approach 1:

The patent segments the centralized XOR chunk storage model into distributed storage across multiple frontend zones. Each frontend zone stores and manages its own XOR chunks locally, eliminating the single point of concentration in backend zones. This segmentation maintains overall storage capacity efficiency while distributing network traffic across multiple zones, preventing traffic overload during disaster recovery operations.

Inventive Principle:
Principle #1Segmentation

Data Source

PatentUS10754845B2System and method for XOR chain
Publication Date: 2020.08.25 EMC IP HLDG CO LLC
  • US10754845B2 patent drawing
  • US10754845B2 patent drawing
  • US10754845B2 patent drawing

AI summary

One embodiment is related to a method for creating a redundancy data chunk for data protection with a chain topology, comprising: transmitting a data chunk of a first frontend zone of a data storage system to a second frontend zone of the data storage system; creating a redundancy data chunk at the second frontend zone of the data storage system based on the data chunk of the first frontend zone and a data chunk of the second frontend zone; passing the redundancy data chunk onto one or more subsequent frontend zones of the data storage system from the second frontend zone, wherein at each subsequent frontend zone the redundancy data chunk is updated based on the received redundancy data chunk and a data chunk of the respective subsequent frontend zone, and wherein the redundancy data chunk is passed through each subsequent frontend zone exactly once; and at a last subsequent frontend zone of the data storage system, forwarding the updated redundancy data chunk to a backend zone of the data storage system for final storage.