Deadline-Driven Synchronous Replication I/O Handling

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Conventional storage systems face challenges in managing synchronous replication latency, particularly when multiple nodes are involved, as telescoping watchdog settings become complex and difficult to determine due to increased data flow complexity and external network factors, leading to inconsistent performance.

Innovation Solution

A method where an I/O request is assigned a deadline, and nodes calculate a remaining time based on this deadline and network round trip time to determine if synchronous replication should be executed, transitioning to asynchronous replication if time is insufficient, thereby controlling latency without relying on node-specific timeout periods.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Reliability

If synchronous replication is performed in distributed storage systems with multiple nodes, then data consistency and reliability are improved, but system latency and complexity increase due to telescoping watchdog settings and network round trip time variations

Engineering Contradiction:

Improvedata consistencyVSAvoidsystem latency

Core Design Contradiction:

ReliabilityVSLoss of time

Solution Approach 1:

The patent applies dynamics by making the replication mode adaptive rather than static. The system dynamically switches between synchronous and asynchronous replication based on real-time deadline calculations. Each I/O request is evaluated against a deadline threshold, and the replication mode is adjusted accordingly, allowing the system to optimize between reliability and latency on-demand rather than being constrained by fixed telescoping watchdog settings

Inventive Principle:

Principle #15Dynamics

Solution Approach 2:

The patent changes the parameter of replication mode from a fixed state to a variable state. By calculating remaining time based on deadlines and comparing it against thresholds, the system changes the replication parameter dynamically. This allows the same system to operate in different replication modes (synchronous/asynchronous) depending on the specific I/O request characteristics and network conditions, resolving the contradiction between maintaining data consistency and minimizing latency

Inventive Principle:

Principle #35Parameter changes

2Loss of time

If telescoping watchdog settings are used to manage replication timing, then I/O request completion timing is controlled, but device complexity and difficulty of configuration increase

Engineering Contradiction:

ImproveI/O request completion timingVSAvoidwatchdog settings complexity

Core Design Contradiction:

Loss of timeVSDevice complexity

Solution Approach 1:

The patent extracts the timing control mechanism from the complex telescoping watchdog settings and replaces it with a simpler deadline-based calculation. Instead of using multiple nested watchdog timers with complex interactions, the system extracts the essential timing requirement and implements it through straightforward deadline comparisons and threshold checks, significantly reducing configuration complexity while maintaining timing control

Inventive Principle:

Principle #2Taking out (Extraction)

Solution Approach 2:

The patent inverts the traditional approach by instead of setting watchdog timers to ensure completion within time limits, it calculates deadlines based on remaining time and adjusts replication mode accordingly. This inversion simplifies the control logic by working backward from the desired outcome (timely completion) rather than forward from timer settings, making the system easier to configure and maintain

Inventive Principle:

Principle #13The other way round (Inversion)

3Reliability

If synchronous replication is always performed, then data consistency is maintained, but host I/O timeouts occur when network latency exceeds expected thresholds

Engineering Contradiction:

Improvedata consistencyVSAvoidhost I/O timeouts

Core Design Contradiction:

ReliabilityVSObject-affected harmful factors

Solution Approach 1:

Inventive Principle:

Principle #15Dynamics

Solution Approach 2:

The patent applies preliminary anti-action by proactively switching to asynchronous replication when the calculated remaining time indicates that synchronous replication would exceed the deadline. This preemptive measure prevents host I/O timeouts from occurring in the first place, rather than reacting to timeouts after they happen. The system anticipates potential timing issues and takes corrective action beforehand

Inventive Principle:

Principle #9Preliminary anti-action

Data Source

PatentUS11656782B2Global deadline driven local synchronous replication I/O handling and recover

Publication Date: 2023.05.23 DELL PROD LP

AI summary

A method is provided for use in a storage system, the method comprising: receiving an I/O request at an R-node; generating a deadline for the I/O request; generating a C-node command based on the I/O request; transmitting the C-node command and the I/O request to a C-node; calculating, by the C-node, a first remaining time based on the deadline; detecting, by the C-node, whether the first remaining time meets a first threshold; when the first remaining time meets the first threshold, executing the I/O request and transmitting, from the C-node to the R-node, synchronous replication request that is associated with the C-node command; and when the first remaining time does not meet the first threshold, causing the storage system to stop performing synchronous replication and executing the C-node command.