多層化之分散式天文資料索引儲存系統;Hierarchical Distributed Index and File System for Astronomical Observation Data

NCUIR > College of Electrical Engineering & Computer Science > Graduate Institute of Computer Science and Information Engineering > Electronic Thesis & Dissertation > Item 987654321/65616

Please use this identifier to cite or link to this item: http://ir.lib.ncu.edu.tw/handle/987654321/65616

Title:	多層化之分散式天文資料索引儲存系統;Hierarchical Distributed Index and File System for Astronomical Observation Data
Authors:	黃郁哲;Huang,Yu-che
Contributors:	資訊工程學系
Keywords:	分散式系統;環狀架構;空間索引;Distributed System;Ring Structure;Spatial Index
Date:	2014-07-16
Issue Date:	2014-10-15 17:06:20 (UTC+8)
Publisher:	國立中央大學
Abstract:	在天體觀測中，星體的資料是龐大且是隨時增加的，所以在保存資料上需要非常大量的儲存設備來支援。假如在資料的儲存上缺少一套有效率的管理方法，那麼當研究人員需要使用資料時，就必須親自確認所需的資料位於那些機器上，再逐一的到這些機器上把資料撈取出來，這種行為費時又不切實際。有鑑於此，本論文開發一套多層化的分散式天文資料管理系統。此系統首先使用多層次三角網格來建構出天空模型，來給予星球一個編號來當作資料索引。由於這些編號都是有順序性而且連續的，本論文將伺服器連結成環狀的架構，讓每一個伺服器都負責一段連續但不重複的資料索引值範圍，並使用這些範圍來確認所要保管的資料，另外每二個相鄰的伺服器其所管理的資料索引值範圍都是連續的。而採用環狀架構的原因為其擴充性、容錯性高以及高可用性。此種架構為系統帶來了良好的基礎。為了加速資料搜尋的效率，此系統必須運行多種不同的服務，因此會大幅的增加機器的負擔，為了要降低機器的負荷，我們將以下三個功能獨立到三層不同的環狀架構。它們分別負責1.將使用者輸入的搜尋條件轉換成系統搜尋命令2.建立資料索引來找出使用者所需的資料以及該資料的存放位置3.儲存星球的實體資料。將這些功能分開運行之後，除了可以降低單台機器的負擔，往後任一個層次運行的功能需要進行調整時，即可直接替換而不需要重新建構整個系統。 ;The data of star observation is huge and increase continuously. Therefore it needs strong hardware to store these data. If these data lacks of a efficiently manage methodology, users will need to find and get data themselves, it’s not a good and efficient way. In order to solve this problem, this paper proposes a hierarchical design distributed file system to manage these astronomical observation data. First, this system uses Hierarchical Triangular Mesh to compute a key for every star, the values of these keys are sequential, so this paper links the servers in the system to form a ring structure, and let every server manages an unrepeated range of keys, range that every two adjacent servers manage are also continuous. Besides, The reason why adopting ring structure is due to its scalability, fault tolerance and availability, these advantages build the good foundation of the system. For increase the efficiency of query, the system needs execute multiple services. To reduce the load of servers, this paper builds three level of the ring structure and allocates these services to different layer. These layers are 1. Query Transaction, 2. Data index, 3. Storage. This design not only reduces the load, but also makes these services independent.
Appears in Collections:	[Graduate Institute of Computer Science and Information Engineering] Electronic Thesis & Dissertation

Files in This Item:

File	Description	Size	Format
index.html		0Kb	HTML	372	View/Open

社群 sharing

Loading...