| In the health industry, digital imaging data are stored and processed by a system called Picture Archive and Communication System (PACS). There are two types of architectures that can be chosen to implement a PACS: centralized and distributed. The centralized architecture has only one archive for storing and processing the digital imaging data. While the centralized architecture is currently the most common architecture, it does not provide the required scalability, reliability, and performance when the amount of digital medical images grows and the mobility of the image users increases. A distributed architecture of PACS's has more than one archive to share the load of storing and processing the digital medical image data. The database is separated physically but integrated logically [1]. The distributed architecture can overcome, the limitations of centralized architecture and has advantages of reduced bottleneck, higher reliability and availability, higher modularity and thus higher scalability [1]. To design a distributed PACS, we need to determine the number, the location, and the capacity of the PACS archives. We also need to determine where to store each image file so that we can maximize the archive accessing performance within a certain budget. This thesis provides a methodology that helps PACS administrators to design a distributed PACS. This methodology can also be helpful to analyze the performance of an existing distributed PACS. |
