Adaptive Scheduled Periodic Caching for Backup Data Deduplication
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional backup systems, particularly those using solid-state disk drives, are inefficient in identifying and caching the most popular anchors for delta-compression during data deduplication, leading to increased data seek and recovery times and reduced data transfer rates, and are costly compared to magnetic disk drives.
Innovation Solution
An adaptive scheduled periodic caching method that determines the popularity of anchors based on the number of dependent versions and their size, storing the most popular anchors in a solid-state disk location to maximize cache efficiency and reduce disk operations.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Speed
If conventional backup systems use solid-state disk drives for caching anchors, then data transfer rates increase, but the cost increases significantly compared to magnetic disk drives
Solution Approach 1:
The patent applies local quality by differentiating the storage medium used for different types of data. Solid-state disk drives are used specifically for caching popular anchors (high-frequency access), while magnetic disk drives are used for storing less frequently accessed anchors. This selective application of high-performance storage to only the most critical data resolves the contradiction by achieving high data transfer rates for popular anchors without the prohibitive cost of using solid-state drives for all anchor storage.
2Ease of manufacture
If anchors are fetched from magnetic disk during delta compression, then storage cost is reduced, but mechanical latencies increase seek and recovery times
Solution Approach 1:
The patent implements preliminary action by proactively caching popular anchors in solid-state disk drive memory before they are needed for delta compression operations. The system identifies anchors with high popularity (frequently accessed) and pre-loads them into the faster solid-state storage. This eliminates the need for time-consuming mechanical seek operations when these anchors are required, as they are already resident in the high-speed cache.
Solution Approach 2:
The solid-state disk drive acts as an intermediary between the magnetic disk storage and the processing system. Instead of directly fetching anchors from magnetic disk (which causes mechanical latencies), the system uses the solid-state drive as a buffer layer that provides fast access to popular anchors, thereby mediating between the low-cost magnetic storage and the high-performance processing requirements.
3Ease of manufacture
If all anchors are stored on magnetic disk, then cost is minimized, but data deduplication performance decreases due to slow access times
Solution Approach 1:
The patent applies partial action by implementing a hybrid storage architecture that uses solid-state disk drives for only a portion of the anchor storage (specifically, the popular anchors that are frequently accessed), rather than using solid-state drives for all anchors. This partial deployment of high-performance storage provides significant performance improvement for data deduplication operations while keeping costs manageable, avoiding the excessive cost of fully replacing magnetic storage with solid-state storage.
Data Source
Figure 1
Figure 2
Figure 3
AI summary
A system, a computer-implemented method and a computer program product for storing backup data are provided. A data stream is received. The received data stream is deduplicated. A file size of each anchor version of data is determined. A number of delta-compressed version of data dependent on each corresponding anchor version of data is determined. Each computed residence ratio is compared to a threshold residence ratio. An anchor version of data is transmitted for storage in a cache storage. Another anchor version of data is transmitted for storage in a disk storage. A new data zone is processed and content of the new data zone is compared to the anchor version of data stored in the cache storage.