Targetless Snapshot Replication Data Pointer Table
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
High capacity data storage systems face resource burdens due to the creation and maintenance of traditional snapshot volumes, which require significant memory and resources, whereas targetless snapshots aim to reduce these requirements by using data replication tables with pointers to the original data, but the number of tables needed grows with the number of snapshots.
Innovation Solution
A data storage system employing a single data structure with metadata that maps locations of targetless snapshots on drives, utilizing direct index lookup tables and virtual replication data pointer tables to efficiently represent and maintain snapshots, reducing resource usage by eliminating unnecessary metadata and dynamically sizing metadata structures.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If traditional volume-based snapshots are created, then snapshot functionality is provided, but memory and resource consumption increase significantly
Solution Approach 1:
The patent uses data replication tables with pointers to original data blocks instead of creating actual copies. Each snapshot is represented by a replication table containing metadata and pointers to the source volume data, eliminating the need to allocate separate storage space for snapshot data while preserving snapshot functionality.
Solution Approach 2:
The patent extracts only the essential metadata and pointer information needed to represent snapshots, separating this from the actual data storage. The replication tables contain only the necessary reference information (pointers to original data blocks) rather than duplicating the entire data set, significantly reducing resource consumption.
2Quantity of substance
If targetless snapshots with data replication tables are used, then resource consumption is reduced, but the number of tables to maintain increases proportionally with the number of snapshots
Solution Approach 1:
The patent merges multiple snapshot representations into a single unified data structure. Instead of maintaining separate replication tables for each snapshot, all snapshots are represented within one consolidated structure that contains metadata for multiple snapshots and their corresponding pointers to source volume data blocks.
Solution Approach 2:
The unified data structure serves multiple functions simultaneously: it represents multiple snapshots, stores metadata for all snapshots, and maintains pointers to source data blocks. This multi-functional structure eliminates the need for separate specialized tables for each snapshot while preserving all necessary information.
3Ease of operation
If separate snap volumes are created for each snapshot, then snapshots are accessible to host servers, but significant memory and resources are required to maintain snap volumes
Solution Approach 1:
The patent creates virtual copies of snapshot data through pointer references in the unified replication table rather than physical copies in separate volumes. The pointers reference the actual data blocks on the source volume, providing snapshot accessibility without duplicating data storage requirements.
Solution Approach 2:
The unified data structure acts as an intermediary between the host servers and the actual snapshot data. It provides the metadata and pointer information needed for hosts to access snapshot data while the actual data remains stored in its original location, eliminating the need for intermediate snap volume storage.
Data Source
AI summary
Storage objects and targetless snaps of the storage objects are represented using a system replication data pointer table (SRT), direct index lookup (DIL) tables, and virtual replication data pointer tables (VRTs). The SRT is a system level track-based data structure that stores metadata indicative of the actual (physical layer) allocations for all targetless snapshots in a storage array. The size of the SRT in terms of total entries corresponds to the overall storage capacity of the managed drives of the storage array. Each utilized entry of the SRT includes backend metadata with a pointer to a managed drive and metadata that identifies the associated storage object and track via the VRTs and DIL tables. SRT metadata is created and discarded as backend allocations are utilized and freed so the SRT is a dynamic data structure that can efficiently adjust its size and corresponding memory requirements.