Volume Replication Mode Switching via Write Request Tagging

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Conventional volume replication systems face challenges in switching between synchronous and asynchronous replication modes, leading to lag or glitches in application throughput and potential data loss due to the need to drain pending write requests, especially in network latency-sensitive environments.

Innovation Solution

The system dynamically switches replication modes by processing existing write requests according to their original mode and associating new requests with metadata indicating the updated mode, allowing for seamless transitions between synchronous and asynchronous replication without draining pending requests, thus maintaining application throughput.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Loss of information

If pending write requests are drained before changing replication mode, then information consistency is improved, but application throughput deteriorates

Engineering Contradiction:

Improveinformation consistencyVSAvoidapplication throughput

Core Design Contradiction:

Loss of informationVSProductivity

Solution Approach 1:

The system performs preliminary actions by tagging write requests with mode indicator information before the mode switch occurs. This allows the system to prepare for the transition by already having the necessary information to route requests appropriately, eliminating the need to drain pending requests and maintaining throughput while ensuring consistency.

Inventive Principle:

Principle #10Preliminary action

Solution Approach 2:

The system segments write requests into different categories based on the replication mode at the time of issuance. By dividing requests into those issued in synchronous mode and those issued in asynchronous mode, the system can process them through appropriate pathways, ensuring information consistency without requiring all requests to complete before mode switching.

Inventive Principle:

Principle #1Segmentation

2Stability of the object's composition

If pending write requests are drained before changing replication mode, then system stability is improved, but time consumption deteriorates

Engineering Contradiction:

Improvesystem stabilityVSAvoidtime consumption

Core Design Contradiction:

Stability of the object's compositionVSLoss of time

Solution Approach 1:

The system performs preliminary tagging of write requests with replication mode information before the mode switch. This preliminary action ensures system stability during transition without requiring time-consuming draining of pending requests, as the mode indicator already embedded in each request enables correct processing throughout the transition period.

Inventive Principle:

Principle #10Preliminary action

3Productivity

If replication mode switching is implemented without draining pending requests, then application throughput is improved, but information consistency deteriorates

Engineering Contradiction:

Improveapplication throughputVSAvoidinformation consistency

Core Design Contradiction:

ProductivityVSLoss of information

Solution Approach 1:

The system performs preliminary tagging of write requests with replication mode indicator information before mode switching occurs. This allows continuous processing of pending requests without draining, maintaining application throughput while ensuring each request is processed according to the appropriate mode, thus preserving information consistency.

Inventive Principle:

Principle #10Preliminary action

Solution Approach 2:

The mode indicator embedded in each write request provides feedback information that guides the processing of the request through the appropriate replication pathway. This feedback mechanism ensures that even during mode transitions, each request is handled consistently with the mode intended at issuance time, preventing information inconsistency.

Inventive Principle:

Principle #23Feedback

4Loss of time

If replication mode switching is implemented without draining pending requests, then time consumption is reduced, but system complexity deteriorates

Engineering Contradiction:

Improvetime consumptionVSAvoidsystem complexity

Core Design Contradiction:

Loss of timeVSDevice complexity

Solution Approach 1:

The system segments write requests by tagging them with replication mode indicators, dividing them into synchronous and asynchronous categories. This segmentation enables rapid mode switching without draining requests, reducing time consumption while managing complexity through organized request classification and routing based on the embedded mode information.

Inventive Principle:

Principle #1Segmentation

Data Source

PatentUS11061603B1Systems and methods for switching replication modes in a volume replication system

Publication Date: 2021.07.13 COHESITY INC

AI summary

The disclosed computer-implemented method for switching replication modes in a volume replication system may include (i) in response to deciding to switch from a synchronous replication mode of a volume replication system to an asynchronous replication mode, changing, by a computing device, to the asynchronous replication mode, (ii) associating a new write request to write data to storage, (iii) determining, based on metadata of the existing write request, that the existing write request was issued in the synchronous replication mode, (iv) in response to determining that the existing write request was issued in the synchronous replication mode, processing the existing write request via the synchronous replication, and (v) processing the new write request via the asynchronous replication based on the metadata of the new write request. Various other methods, systems, and computer-readable media are also disclosed.