Cloud Snapshot Recovery Using Parallel Temporary Instances

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Data recovery from cloud snapshots following an availability zone (AZ) failure is slow, leading to long latency and potential timeouts due to bandwidth limitations and high latency in accessing region-level storage.

Innovation Solution

Create multiple temporary instances in a secondary AZ to parallelize the copying of recovery data from cloud snapshots to local volumes, then attach these volumes to a storage-controller instance for faster data access.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Reliability

If data is recovered by copying from cloud snapshots to new volumes using a single virtual storage controller, then data recovery is possible, but the recovery process is slow and results in long latency

Engineering Contradiction:

Improvedata recovery capabilityVSAvoiddata recovery speed

Core Design Contradiction:

ReliabilityVSSpeed

Solution Approach 1:

The patent divides the data recovery process into multiple parallel operations by creating multiple temporary instances, each responsible for copying data from different snapshots or different portions of snapshots. This segmentation of the recovery task across multiple instances enables concurrent data transfer operations, significantly improving recovery speed while maintaining complete data restoration capability

Inventive Principle:

Principle #1Segmentation

2Reliability

If a full read of all data from cloud snapshots is performed into new volumes, then fast local access to data is achieved, but the full read process requires considerable time and extends the period of slow data access

Engineering Contradiction:

Improvelocal data access capabilityVSAvoiddata unavailability time

Core Design Contradiction:

ReliabilityVSLoss of time

Solution Approach 1:

The patent performs preliminary actions by proactively copying data from cloud snapshots to local volumes using multiple temporary instances before the primary storage controller fails or before data access is critically needed. This advance data preparation ensures that when failure occurs, the system can immediately switch to using pre-copied local data, minimizing data unavailability time while still achieving fast local access

Inventive Principle:

Principle #10Preliminary action

3Adaptability or versatility

If bandwidth limits on cloud instances are considered, then resource constraints are acknowledged, but the full read cannot proceed faster than the bandwidth limit allows

Engineering Contradiction:

Improvebandwidth utilizationVSAvoiddata recovery throughput

Core Design Contradiction:

Adaptability or versatilityVSProductivity

Solution Approach 1:

The patent segments the data recovery workload across multiple temporary instances, each operating within cloud bandwidth limits. By distributing the total data transfer task across multiple parallel instance-stream connections, the system effectively multiplies the aggregate bandwidth available for recovery operations, achieving higher overall throughput while respecting individual instance bandwidth constraints

Inventive Principle:

Principle #1Segmentation

Solution Approach 2:

The patent combines the bandwidth resources of multiple temporary instances to achieve aggregate throughput that exceeds what a single instance could provide. By merging parallel data transfer operations from multiple instances into a coordinated recovery process, the system overcomes individual bandwidth limitations and achieves higher overall data recovery productivity

Inventive Principle:

Principle #5Merging (Combining)

Data Source

PatentUS20260056848A1Recovering data in cloud-based storage environments

Publication Date: 2026.02.26 DELL PROD LP

AI summary

A technique for recovering data following a failure of a first availability zone (AZ) of a cloud-based system includes creating a plurality of temporary instances in a second AZ of the cloud-based system. The technique further includes copying, by the plurality of temporary instances, recovery data from multiple cloud-based snapshots to a plurality of local volumes of the second AZ such that the temporary instances of the plurality of temporary instances copy respective portions of the recovery data in parallel. The technique still further attaching the plurality of local volumes to a storage-controller instance running in the second AZ.