Presentation | 2003/7/10 ファイルサーバ利用履歴に基づく分散アプリケーション群とファイル要求列の関連抽出(パターン抽出)(「夏のデータベースワークショップ(DBWS2003)」一般) Norihide SHINAGAWA, |
---|---|
PDF Download Page | PDF download Page Link |
Abstract(in Japanese) | (See Japanese page) |
Abstract(in English) | Large-volume storage systems generally use secondary storage media such as tapes. HSM (Hierarchical Storage Management) softwares are used to keep usability for users by virtualization of secondary storages. However, we must spend times to exchange media and seek data blocks. The purpose of this study is to realize speculative staging technology. It enable to pre-load appropriate files before requests by prediction based on the history of file accesses. This paper discusses use of ICA (Indepent Component Analysis) for derivation of relationships between accessing distributed applications (tasks) and accessed files as the first approach. I examines whether ICA is applicable to the derivation by simulations based on the situation in my research center. And then, I discuss promlems obtained by the examinations. |
Keyword(in Japanese) | (See Japanese page) |
Keyword(in English) | storage / access analysis / ICA (indepent Component Analysis) |
Paper # | DE2003-47 |
Date of Issue |
Conference Information | |
Committee | DE |
---|---|
Conference Date | 2003/7/10(1days) |
Place (in Japanese) | (See Japanese page) |
Place (in English) | |
Topics (in Japanese) | (See Japanese page) |
Topics (in English) | |
Chair | |
Vice Chair | |
Secretary | |
Assistant |
Paper Information | |
Registration To | Data Engineering (DE) |
---|---|
Language | JPN |
Title (in Japanese) | (See Japanese page) |
Sub Title (in Japanese) | (See Japanese page) |
Title (in English) | |
Sub Title (in English) | |
Keyword(1) | storage |
Keyword(2) | access analysis |
Keyword(3) | ICA (indepent Component Analysis) |
1st Author's Name | Norihide SHINAGAWA |
1st Author's Affiliation | () |
Date | 2003/7/10 |
Paper # | DE2003-47 |
Volume (vol) | vol.103 |
Number (no) | 191 |
Page | pp.pp.- |
#Pages | 5 |
Date of Issue |