Object-Based Data Recovery with IO Filter Interception
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current cloud-based object storage systems are not well-suited for continuous data protection (CDP) due to challenges in populating the object store continuously and the complexity of failing over to an object store, which impacts recovery point objectives and application performance.
Innovation Solution
An object-based data recovery system that replicates data in real-time to an object store with metadata for point-in-time recovery, enabling full and version-based recovery by intercepting data through an IO filter and storing it with corresponding metadata, allowing for seamless failover and failback with minimal downtime.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If frequent snapshots are used to achieve lower recovery point objectives, then data protection capability is improved, but application performance deteriorates and storage space utilization is reduced
Solution Approach 1:
The patent uses copy-on-write technology to create snapshots by copying data blocks only when modifications occur, rather than frequently copying entire data sets. This allows multiple snapshots to be maintained with minimal performance impact on the primary storage system, resolving the contradiction between improved data protection and maintained application performance
Solution Approach 2:
The patent implements changed block tracking to identify and protect only the specific data blocks that have changed, rather than treating the entire storage system uniformly. This localized approach reduces the overhead on application performance while maintaining comprehensive data protection capabilities across the storage system
2Quantity of substance
If incremental snapshot methods with changed block tracking are used, then storage space utilization is improved, but processing overhead on primary storage resources increases
Solution Approach 1:
The patent extracts the changed block tracking functionality from the primary storage path to a separate background processing mechanism. This allows incremental snapshots to be created with improved storage space utilization while minimizing the processing overhead on primary storage resources by performing tracking operations asynchronously
Solution Approach 2:
The patent performs preliminary identification of changed blocks before creating incremental snapshots, allowing the system to prepare and organize only the necessary data blocks in advance. This reduces processing overhead during the actual snapshot creation process while maintaining efficient storage space utilization
3Reliability
If full recovery from multiple incremental snapshots is performed, then data recovery capability is improved, but recovery time and computational resources increase
Solution Approach 1:
The patent performs preliminary organization and indexing of incremental snapshot data during the backup process, creating a structured hierarchy of data blocks across multiple snapshots. This preliminary preparation enables full recovery operations to proceed efficiently by allowing the system to quickly locate and reconstruct data across multiple snapshots without excessive computational overhead or recovery time
Solution Approach 2:
The patent segments the recovery process into independent operations that can be performed in parallel, breaking down the complex task of recovering data from multiple incremental snapshots into smaller, manageable units. This segmentation improves both data recovery capability and reduces overall recovery time by enabling concurrent processing of different data segments
Data Source
AI summary
Data recovery systems and methods utilize object-based storage for providing a data protection and recovery methodology with low recovery point objectives, and for enabling both full recovery and point-in-time based recovery. Data generated at a protected site (e.g., via one or more virtual machines) is intercepted during write procedures to primary storage. The intercepted data is replicated via a replication log, provided as data objects, and transmitted to an object based storage system. During recovery, data objects may be retrieved through point-in-time based recovery directly by the systems of the protected site, and/or data objects may be provided via full recovery, for example, within a runtime environment of a recovery site, with minimal data loss and operation interruption by rehydrating data objects within the runtime environment via low-latency data transfer and rehydration systems.