Snapshot Mechanism for Thinly Provisioned Memory Volumes
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing snapshot mechanisms in computer memory are inefficient in terms of data and metadata space usage, performance, and flexibility, particularly in thinly provisioned systems and those employing data deduplication, as they often require extensive metadata and can lead to prolonged snapshot creation and deletion processes, impacting recovery point objectives.
Innovation Solution
A snapshot mechanism that only stores metadata for changed parts of the address range, allowing for a thinly provisioned snapshot hierarchy with efficient mapping and indirection layers, enabling fast snapshot creation and deletion, and optimizing metadata usage based on user writes, with granular mapping and deduplication techniques.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of manufacture
If traditional snapshot mechanisms store metadata for the entire address range, then snapshot creation is simple and fast, but metadata space consumption becomes excessive and defeats the purpose of thin provisioning
Solution Approach 1:
The patent divides the address space into multiple segments or ranges, and creates separate mapping tables for each segment. Only segments that contain actual data require metadata entries. This segmentation allows the system to avoid storing metadata for the entire address range, reducing metadata space consumption while maintaining snapshot functionality.
Solution Approach 2:
The patent implements partial metadata storage by only creating mapping table entries for address ranges that actually contain data, rather than pre-allocating metadata for the entire address space. This partial action approach reduces metadata overhead while still providing complete snapshot coverage for all actually-used addresses.
2Quantity of substance
If snapshot mechanisms use fine granularity mapping, then data space efficiency improves, but metadata overhead and processing complexity increase
Solution Approach 1:
The patent segments the address space into manageable ranges, creating mapping tables at an appropriate granularity level. This segmentation balances data space efficiency with metadata management complexity by grouping addresses into logical units that are neither too coarse nor too fine-grained.
Solution Approach 2:
The patent dynamically adjusts the granularity of mapping tables based on the actual data distribution and access patterns. By changing the parameter of mapping granularity adaptively, the system optimizes between data space efficiency and metadata processing complexity according to specific workload characteristics.
3Speed
If snapshots store complete metadata for all address ranges, then read operations are fast and simple, but snapshot deletion becomes prolonged and impacts recovery point objectives
Solution Approach 1:
The patent extracts and removes metadata entries for address ranges that are no longer needed or have been released. By selectively removing only the necessary metadata entries during snapshot deletion rather than processing the entire address space, the system reduces deletion time and improves recovery point objectives while maintaining read performance.
Solution Approach 2:
The patent performs partial metadata processing during snapshot operations by only touching the metadata entries that are actually needed for the specific address ranges being modified. This partial action approach reduces the time overhead for snapshot creation and deletion while maintaining fast read performance for accessed data.
4Adaptability or versatility
If the system creates snapshots frequently to improve backup flexibility, then backup coverage improves, but metadata overhead accumulates and consumes excessive space
Solution Approach 1:
The patent merges or consolidates metadata from multiple snapshots, particularly for common address ranges that haven't changed. By combining metadata entries across snapshots and eliminating duplicates, the system enables frequent snapshot creation for backup flexibility while preventing metadata overhead from accumulating excessively.
Solution Approach 2:
The patent discards redundant metadata entries that are no longer needed or can be inferred from other snapshots, and recovers space by removing duplicate mappings. This allows the system to maintain frequent snapshots for backup flexibility while managing total metadata overhead through selective discarding of unnecessary entries.
Data Source
AI summary
A memory management system for a thinly provisioned memory volume in which a relatively larger virtual address range of virtual address blocks is mapped to a relatively smaller physical memory comprising physical memory blocks via a mapping table containing entries only for addresses of the physical memory blocks containing data. The memory management system comprises a snapshot provision unit to take a given snapshot of the memory volume at a given time, the snapshot comprising a mapping table and memory values of the volume, the mapping table and memory values comprising entries only for addresses of the physical memory containing data. The snapshot is managed on the same thin provisioning basis as the volume itself, and the system is particularly suitable for RAM type memory disks.