English  |  正體中文  |  简体中文  |  全文笔数/总笔数 : 63998/96720 (66%)
造访人次 : 3627097      在线人数 : 409
RC Version 7.0 © Powered By DSPACE, MIT. Enhanced by NTU Library & TKU Library IR team.
搜寻范围 查询小技巧:
  • 您可在西文检索词汇前后加上"双引号",以获取较精准的检索结果
  • 若欲以作者姓名搜寻,建议至进阶搜寻限定作者字段,可获得较完整数据
  • 进阶搜寻


    jsp.display-item.identifier=請使用永久網址來引用或連結此文件: https://tkuir.lib.tku.edu.tw/dspace/handle/987654321/76920


    题名: 雲端資料庫NoSQL技術之探討與開發
    其它题名: Research and Development on Nosql Techniques for Cloud Databases
    作者: 莊博任
    贡献者: 淡江大學電機工程學系
    关键词: Cloud Databases;Social Networks;SQL;NoSQL;Elastic Scaling;Fault Tolerance;Data Consistency;Scalability;Highly Coincident Data
    日期: 2011-08
    上传时间: 2012-05-22 21:54:36 (UTC+8)
    摘要: 常見的傳統關聯式資料庫, 採用 SQL 資料查詢技術,但是近年來由於社群網路和雲端資料 庫的興起,讓資料庫面對的不再只是少量的寫入資料,因此亟需突破容量的限制。以往資料庫的 擴充都是以垂直擴充為主,但現今資料的使用量成指數上升,讓以往的SQL 技術遇上了瓶頸。這 時雲端資料庫方面興起的NoSQL(Not only SQL)技術,提倡以彈性擴充、資料型態自由化來因 應現有之問題及限制,已快速竄升為未來發展應用的大趨勢。NoSQL 的主要架構以Google 提出 的Bigtable 和 Amazon 提出的Dynamo 所代表之分散式架構為基本依據,但現行NoSQL 技術需要 面對如下幾項重要議題: 一、資料一致性:分散式架構下如何保持資料的一致性,讓資料不因龐大的架構而錯亂。主 要以Paxos 的理論為依據架構出一個一致性的系統,現有之系統如Chubby、Zookeeper 即針對資 料一致性的處理而設立,但其中間流程卻可能導致資料處理的延遲,這對處理大量資料的需求而 言,顯然不利,必須加以探討改進。 二、資料備份:過去以備份數越多越完善,現在dynamo 提出之Quorum Protocol 乃依據系統 的需求性來動態調整備份數,但是如果規模過於龐大,面對高併發性的資料,可能會導致頻寬的 瓶頸,所以必須對此協定進行調整改善,以減少其負效應。 三、資料庫的延伸性:現在的NoSQL,部分以單筆數據量來測量系統的效能,但我們必須考 慮,面對資訊爆發時代而言,問題式資料庫所面臨的龐大使用量,可能在超越一個瓶頸後,資料 庫的效能便急遽下降,如何確認一個資料庫的最合適臨界值,以確保其最佳效能,值得深入探討 以為因應。 如上所述,目前 NoSQL 技術雖未臻成熟完備,但是已可預期其將來的廣泛發展應用,此乃 必然趨勢,因為傳統資料庫已無法負荷越來越大量的資料處理。然而現行之NoSQL 架構存在相 當的調整改善空間,值得我們進一步探討與開發,本計畫的主要研究目標,即是針對上述及其他 現有實現NoSQL 架構的可能問題及缺點進行改善或解決,以切合當前雲端資料庫之迫切需求。
    SQL, the data query technique for traditional Relational Databases, can no longer satisfy the rapidly rising processing needs of today’s new computing developments or applications, such as the widely popular Social Networks or the hotly pursued Cloud Databases which need to deal with massive quantities of data, especially the large amount of write-in data. How to maximize the volumes of nowadays databases becomes a critical key issue. Previous databases adopt Vertical Scaling to achieve the goal, but it is now infeasible because SQL will have problems managing today’s exponentially growing data items. A new query technique–NoSQL (Not only SQL)–is thus introduced to meet the current needs in Cloud Computing by using Elastic Scaling and Free Data Format. Based mainly on the distributed structures of the Google Bigtable and Amazon Dynamo, NoSQL will well fit today’s IT developing trend – if the following issues can be fully addressed. (1) Data Consistency: Data consistency is extremely important for distributed systems with massive data, but current consistency maintenance mechanisms, such as Chubby and Zookeeper (based on the Paxos theory), are likely to produce data processing delay due to certain operation steps. (2) Data backup Copy: The Quorum Protocol of Dynamo can dynamically adjust the amount of data backup copies to suit the practical needs, however, when the amount grows excessively huge, highly coincident data may cause bandwidth bottleneck. (3) The Scalability: NoSQL tends to evaluate system performance by Single Amount of Data. This may cause Query Databases to suffer abrupt performance recess after the massive data flow exceeds a certain value. Conducting further investigation into the issue to attain a proper threshold thus becomes necessary. This research plan aims to tackle the above problems and come up with reasonable/favorable solutions. Our ultimate goal will be advancing the fresh NoSQL into a widely applicable data query technique to handle the extraordinarily growing volumes/varieties of data processing in either social networks or cloud computing.
    显示于类别:[電機工程學系暨研究所] 研究報告

    文件中的档案:

    没有与此文件相关的档案.

    在機構典藏中所有的数据项都受到原著作权保护.

    TAIR相关文章

    DSpace Software Copyright © 2002-2004  MIT &  Hewlett-Packard  /   Enhanced by   NTU Library & TKU Library IR teams. Copyright ©   - 回馈