基於網格運算並具備工作負載分享之數位內容智財權管理服務系統

Abstract

分散式運算 (Distributed Computing)允許應用程式透過網路，在不同地方的電腦 上執行，將可不必把運算任務集中在某台電腦，而是以功能任務分派給各地電腦運 算，再傳回結果與回覆任務。隨著系統架構的演進，相較於早期的主從式(Client-Server) 架構，同儕運算(Peer-to-Peer)與網格運算(Grid Computing)架構的分散式系統應用已逐 漸成為企業運算的主流。 本研究提出一個基於網格運算下且具有工作負載分享之數位內容智財權管理服 務系統，此平台符合Grid Portal 的標準，使用Globus Toolkit 與Condor 來作為整個平 台的基礎，並利用Web 2.0 based 介面來提供使用者選擇服務、遞交、控制、管理工 作與即時監控工作進度，目的為提供數位內容智財權管理服務包括： 1、不可視浮水印嵌入與偵測 2、圖片攪亂以增強訊號的強韌性 3、可視浮水印的嵌入 4、影像竄改驗證與還原 5、浮水印攻擊測試 6、使用者自訂工作。 然而為了使得平台具有 QOS 的動態資源協調能力，來降低整體工作執行時間， 在平台中，透過協調機制的工作評估服務，利用電腦效能指標，來預測工作執行時間 並進行有效的工作分割與合併，使平台保有穩定的運算能力。 因此本計畫將充分運用網格運算的特性，有效運用網路服務來對數位內容的使用 作分享、搜尋與管理，同時保護智慧財產權在開放的網路環境下依然能夠保護著作人 的所有權。

Distributed Computing allows applications running on different computers in different location to interact with each other across the network. It can dispatch tasks to remote computers and assemble results without focusing on single computing unit. As the distributed computing environment and technique is advancing, P2P and Grid Computing are gaining attraction of the enterprise computing and become the major computing approach gradually. In this research, we propose a copyright management service system of digital content based on grid computing architecture with load balance sharing mechanism which is in compliance with the Grid Portal standards. The platform will integrate Globus Toolkit and Condor middleware and use web 2.0 to construct the web-based user interface which can speed up the interaction between job submission, monitor control management and DRM services. It will provide six category of digital right management services includes: 1. invisible watermark embedding and extraction 2. image scrambling for the enhancement of signal robustness 3. visible watermark embedding 4. image tamper detection and recovery 5. various watermark algorithms attack tests 6. user defined tasks. In addition to increase the platform computing capability, decrease job transmission overhead, improve the quality of service for dynamic task allocation, a load balance sharing mechanism which will accomplish the job evaluation services by using the performance of computing units to predict job execution time will be proposed in order to achieve the high throughput and sustainable computing power for the platform. Finally, the project will fully apply the characteristics of Grid Computing for the proposed copyright management service system of digital content which will effectively use the service for information sharing, searching and management. In the mean time, the service system will perform a feasible solution for the copyright protection under the open network environment.