NAND flash memory-based storage devices are becoming an attractive storage solution for multimedia storage servers because they afford several advantages including fast read access speed and low power consumption. Multimedia storage devices differ significantly from traditional storage devices in that the user files are accessed in a sequential read manner while metadata are updated frequently. In this paper, we propose a new flash translation layer (FTL) scheme called filtering FTL (FFTL) for nand flash-based multimedia storage devices. The main idea of the FFTL scheme is to filter update requests for metadata and manage them separately from requests for user data. Our experimental results show that the proposed FFTL scheme outperforms existing FTL schemes by dramatically reducing the garbage collection overhead.

NAND flash-based storage systems such as solid-state disk (SSDs) are becoming an alternative secondary storage solution for various applications. In flash-based storage systems, an intermediate software layer called a flash translation layer (FTL) is usually employed to hide the erase-before-write characteristics of the NAND flash memory. In FTL schemes, garbage collection is the major factor limiting the performance of the storage device. In this paper, we propose a new hybrid-level FTL scheme called switchable address translation (SAT), which dynamically switches the address mapping scheme according to the access pattern and the memory usage of the mapping tables. In order to minimize the garbage collection overhead, SAT removes the merge operation which is an innate problem of legacy hybrid-level FTL schemes. Further, unlike previous schemes, SAT utilizes free blocks and controls the execution of garbage collection. Our experimental results show that the proposed SAT scheme reduces the garbage collection overhead significantly as compared to the existing FTL schemes.

Solid-state drives (SSDs) based on NAND flash are widely used for various server applications. Multimedia storage devices differ significantly from traditional storage devices in that the user files are accessed in a sequential read manner while metadata are updated frequently. In this paper, we propose a new flash translation layer (FTL) scheme called the filtering FTL (FFTL) for multimedia storage devices. The main idea of the FFTL scheme is to filter update requests for metadata and manage them separately from requests for user data. This scheme manages frequently updated blocks by using a page-level address mapping scheme. Our experimental results show that the proposed FFTL scheme improves the overall performance by reducing the garbage collection overhead.

In this paper, we present a novel buffer cache management scheme for video-on-demand servers called hybrid buffer cache management (HBM). We carefully argue that multimedia workload may exhibit characteristics other than sequential access pattern. If the access pattern is not sequential, the legacy buffer cache schemes that are grounded on the assumption that the applications sequentially access the multimedia file may not work properly. The HBM automatically detects the reference patterns of each file and intelligently switches between different buffer cache management schemes on per-file basis. According to our experiment, the HBM exhibits better buffer cache hit ratio than interval based caching or LRU, especially when the workload exhibits not only sequential but also looping reference property.