講演抄録/キーワード |
講演名 |
2023-06-02 16:10
Method for sizing backup system on cloud considering burstable network and storage throughput ○Avais Ahmad・Shinichi Hayashi・Kaori Murase(Hitachi) SC2023-8 |
抄録 |
(和) |
(まだ登録されていません) |
(英) |
Organizations wants to protect their data by using data protection services for backup. An infrastructure needs to be deployed for backup services and appropriate VM instance size should be used. It is difficult to choose most optimal infrastructure to reduce cost due to cloud credit mechanism and multiple target systems with variable data sizes backing up with overlapping window time. Cloud Credit mechanism make use of a credit bucket in which credits are accrued when the resource is idle, and credits are consumed on high performance requirement to give high burst speed for some CPU, network or disk. The proposed method overcomes the challenge of estimating the optimal VM size for backup repository when there are multiple data transfer jobs with credit-based mechanism for throughput. The proposed method visualizes a timeseries simulation for transfer of data, available credits and allocated throughput. The simulation checks for each instance type till it finds the optimal instance on which all jobs complete before its window time. User can save approx. 20% of total infrastructure cost each month by implementing this method. |
キーワード |
(和) |
/ / / / / / / |
(英) |
Data Backup / Cloud Credits / Bursting / Simulation / Optimization / Sizing / / |
文献情報 |
信学技報, vol. 123, no. 61, SC2023-8, pp. 42-47, 2023年6月. |
資料番号 |
SC2023-8 |
発行日 |
2023-05-26 (SC) |
ISSN |
Online edition: ISSN 2432-6380 |
著作権に ついて |
技術研究報告に掲載された論文の著作権は電子情報通信学会に帰属します.(許諾番号:10GA0019/12GB0052/13GB0056/17GB0034/18GB0034) |
PDFダウンロード |
SC2023-8 |
|