Tape Archive Management Using Erasure Coding for Data Access
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional tape archive systems require multiple cartridge tapes for reading and writing, leading to high costs and inefficiencies due to the need for multiple drives and high busy time, especially when accessing small objects or large data sizes, which complicates parallelization and increases access time.
Innovation Solution
A management apparatus that employs a data arrangement determination unit to select between horizontal and vertical data arrangement methods based on data size, allocating smaller data to one tape for efficient writing and larger data to multiple tapes for parallel processing, using erasure coding to ensure redundancy and error correction.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If multiple cartridge tapes are used for reading and writing, then data redundancy and reliability are improved, but access time increases and productivity deteriorates
Solution Approach 1:
The patent segments data into multiple data blocks and distributes them across different cartridge tapes using erasure coding. This allows the system to maintain reliability through distribution while enabling parallel access to multiple tapes simultaneously, thereby reducing overall access time compared to sequential access of multiple tapes.
Solution Approach 2:
The patent introduces a new dimension of parallel processing by enabling simultaneous reading and writing operations across multiple cartridge tapes. This dimensional change from sequential to parallel operations resolves the contradiction by allowing redundancy maintenance without proportionally increasing access time.
2Productivity
If multiple drives are prepared for parallel processing, then productivity is improved, but device complexity and cost increase
Solution Approach 1:
The patent enables existing single-drive tape libraries to perform parallel processing operations by implementing erasure coding and distributed data block management. This allows the system to achieve multi-functionality where a single drive can effectively operate in parallel with itself through intelligent data distribution and recovery algorithms, without requiring additional physical drives.
3Reliability
If data is distributed across multiple cartridge tapes, then reliability is improved, but the number of tapes required increases
Solution Approach 1:
The patent changes the parameter of data organization from traditional replication (where full copies are made) to erasure coding (where data is divided into blocks with redundant parity information). This parameter change allows achieving the same reliability level with fewer tapes, as erasure coding is more space-efficient than full replication while maintaining the ability to recover from tape failures.
Data Source
AI summary
A management apparatus, includes a memory; and a processor coupled to the memory and configured to: manage a library device including one or more drive devices, an accommodation shelf in which accommodation cells that accommodate a plurality of tape media are arranged, and a robot that performs conveyance operation to convey the tape media between the accommodation cells and the drive devices, receive an instruction to write a plurality of pieces of writing data, set, as an erasure coding set, two or more pieces of writing data having a data size less than a threshold among the plurality of pieces of writing data, and allocate the two or more pieces of writing data to one of the plurality of tape media.