於協同配置資料網格環境中具適應性複本管理的高效能醫療影像儲傳系統之實作(I)

Tunghai University Institutional Repository > 工學院 > 資訊工程學系所 > 國科會研究報告 > Item 310901/10193

Please use this identifier to cite or link to this item: http://140.128.103.80:8080/handle/310901/10193

Title:	於協同配置資料網格環境中具適應性複本管理的高效能醫療影像儲傳系統之實作(I)
Other Titles:	Implementation of a High-Performance PACS System with Adaptive Replica Management in Co-Allocation Data Grid Environments(I)
Authors:	楊朝棟,楊晴雯 Yang, Chao-Tung；Yang, Ching-Wen
Contributors:	行政院國家科學委員會東海大學資訊工程與科學系
Date:	2008
Issue Date:	2011-06-23T05:57:26Z (UTC)
Abstract:	影像診斷學中，醫療影像儲傳系統(Picture Archiving and Communication System, PACS)是一種專門用來儲存、取得、傳送與展示醫療影像的電腦或網路系統。PACS的主要目的在於將醫療系統中所有影像，以數位化的方式儲存，並經由網路傳輸至系統中，供使用者遠端電腦螢幕閱讀影像並判讀。同時也可作為不同影像傳遞交換的工具，隨著軟體及運用程式的進步，將來更可進一步協助醫師進行診斷、教學及醫學研究。而成功的PACS不只須要強大硬體，更依賴完善軟體功能及作業程序。透過日漸成熟的網格計算(Grid Computing)技術，將散佈各地之虛擬組織的資源可以透過網格的概念來調派和集中。再者，利用資料網格(Data Grid)的容錯特性與高可用性，因此可以滿足各種在醫療資訊應用方面的計算與檔案儲存需求。為降低PACS系統之擴建成本與設置第二PACS系統為考量，本計畫為期3?，主要是研發Smart Broker Centric與具有自適性副本管理元件於於協同配置(Co-allocation)資料網格環境中。藉由導入Open Source PACS解決方案，建置於特別設計的網格功能元件之上，期能證實以網格技術來支援PACS系統的可行性，除保有PACS的優勢、實際提昇PACS影像副本交換效率與一個成本較低廉的PACS導入方案。本?度，計畫完成平台架構上所有系統元件的雛型設計開發，平台架構區分Application、Smart Broker、Cyber Abstraction、Grid Middleware以及Fabric等5個階層，其中較高層以使用者為中心，並以Smart Broker作為本架構核心；底層專注於資源整合，並以Cyber Abstraction描述高低二階層之間如何連結。另外設計一個網格知識管理(GKM)模型以支援各元件計算的參數資料輸入。其成果已在APSCC08, IPCADS08與JNCA發表。接下來第1?，計畫整合各階層元件，使用Globus Tookit與雲端計算(Cloud Computing)結合Cross-CA技術，建構跨網格系統的交互驗證機制，提昇計算與儲存資源利用率。實作Co-Allocator元件，該元件能經由Resource Management System (RMS)提供對應資源，並透過GKM的知識基礎，設計出具有認識網格環境現況的能力，即為自適性。為了量測效率，另外建立客戶端應用程式以監控管理Workflow，並且能將工作歷程與結果回饋給GKM。使用Open Source PACS系統為應用實?，進?實驗及測試雛型系統功能及相容性。第2?，根據應用實?的實驗結果，進一步改善各階層元件與雛型應用系統與GKM學習樣本調較。並就我們的PACS系統實際使用情況與真實醫院PACS系統作整體效能比較。 In the first year, we design and implement the prototype, including all components of the platforms. The architecture is composed of five layers: Application, Smart Broker, Cyber Abstraction, Grid Middleware and Fabric. The higher layers are user-centric, and use Smart Broker as the core of the architecture. The bottom layers focus on resource integration, and use Cyber Abstraction to describe the interconnection of the top layers and the bottom layers. Also, a grid knowledge management (GKM) model is designed to facilitate the parameter input of all components. Some research results are published in APSCC08, ICPADS08 and JNCA.In the nest first year, we plan to integrate components of all layers by using Globus Tookit, Cloud Computing and Cross-CA technologies, in order to improve the computing performance and resource utilization. The implemented Co-Allocator component can provide corresponding resources by Resource Management System (RMS), and support grid context-awareness by GKM. For evaluation, we plan to construct client applications for remotely controlling Workflow, and feedback the work history and results to GKM. Taking Open Source PACS for examples, we plan to conduct experiments on functionality and compatibility of the prototype. In the nest second year, we plan to refine the components and the application according to the experimental results. Also, the proposed system will be compared with a real-world PACS of hospitals in terms of overall performance.
Relation:	研究編號：NSC97-2622-E029-003-CC2 研究期間：2008-08~ 2009-07
Appears in Collections:	[資訊工程學系所] 國科會研究報告

Files in This Item:

File	Size	Format
972622E029.pdf	2253Kb	Adobe PDF	651	View/Open

Loading...