NDE (Nondestructive evaluation) presents a way to characterize various properties of materials, components, or structures without damaging the objects. X-Ray techniques including radiography and CT (Computed Tomography) are of great importance because of their capability in detecting interior defects as well as surface flaws. In recent years, more and more interest is focused on micron-level features. A consequence is the rapidly growing data volume of both original CT data and 3D images. This data handling capability at both data acquisition stage and in the processing part is largely limited by computer power available, including CPU processing and data transfer. In this thesis a design for a high resolution CT system based on amorphous silicon detector and a microfocus X-Ray tube is presented. The implementation of the solution involves diverse research areas including network programming, parallel program developing, file format design and converting, physical system calibration, algorithm development, and graphical interface development. The crucial part of the approach is a 64-node Linux cluster. Utilizing the computing power of the cluster, the image reconstruction process is expedited tremendously. Moreover, the capacity of system memory and disk space makes it suitable to serve as temporary CT data storage through the fast Ethernet network connections, which in turn makes it possible to handle large dataset (multi-gigabyte) from a normal data acquisition PC without lengthening processing time. Keywords: X-Ray Computed Tomography, high resolution, network programming, parallel programming, Linux cluster