Virtual Disk Interval Storage for Multi-Stream Data
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Storage systems face performance limitations with multi-stream and multi-user applications due to mechanical latency in electromagnetic disk drives and RAID technologies, leading to inefficient random read/write operations and increased costs with high-end drives like SAS/FC and SSDs, while current caching systems do not differentiate between meta-data and video data, causing performance degradation.
Innovation Solution
The method involves converting random read/write patterns to sequential patterns by using virtual disk intervals and data aggregation policies, where data blocks are mapped from logical to physical addresses to optimize storage and reduce random write/read commands, utilizing a chunk table to allocate physical addresses dynamically and implementing a rotational volume with a cleanup space to improve performance.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If random read/write operations are performed on electromagnetic disk drives, then data can be written to any location, but performance deteriorates due to mechanical latency, seek time, and rotational latency
Solution Approach 1:
The patent segments the storage space into intervals and groups data blocks from multiple streams into sequential intervals, transforming random access patterns into sequential access patterns that exploit the high sequential throughput of electromagnetic disk drives while avoiding mechanical latency penalties
Solution Approach 2:
The patent introduces a storage controller with a cache as an intermediary between the host and the disk drives. The controller receives random write requests, buffers them in the cache, and then issues sequential write commands to the disks, effectively decoupling the random access requirements from the sequential access capabilities of the storage medium
2Productivity
If many high-performance drives (SAS/FC, SSD) are used to improve random read/write performance, then storage performance improves, but cost increases significantly
Solution Approach 1:
The patent changes the access pattern parameter from random to sequential, allowing the use of inexpensive SATA drives instead of expensive SAS/FC or SSD drives. By transforming the workload characteristics rather than upgrading the hardware, the system achieves high performance at lower cost
3Reliability
If RAID5 is used for storage, then data redundancy is provided, but performance deteriorates due to additional read operations required for each write
Solution Approach 1:
The patent performs preliminary actions by pre-calculating and pre-positioning data blocks in sequential intervals before the actual write operation. This allows RAID5 parity calculations to be performed in advance and data to be written sequentially, minimizing the performance impact of RAID5's additional read requirements
Data Source
AI summary
A method for storing data, comprises the steps of: defining one or more intervals for one or more virtual disks, wherein each of the intervals has data; receiving a storage command in a cache, wherein the command having a logical address and a data block; determining a respective interval for the data block corresponding to the logical address of the data block; determining whether the data of the respective interval is to be written to a corresponding storage unit; and receiving a next storage command.