目前有許多的動畫不若以往是由二維的影像所組成，而是以三維空間來表現出整個場景，在處理三維場景時，為了要將最後的結果產生在螢幕上，並呈現出逼近真實世界中的效果，我們必須決定螢幕上每一個像素所對應的顏色值，在過去的幾年，已經有許多的繪圖演算法被發展出來，本篇論文所採用的是光跡追踨演算法，由於在整個處理的過程中必須耗費相當大量的時間，若以單一台電腦來計算所有結果，並無法符合經濟效益，因此我們利用目前熱門的電腦叢集系統，藉由集合多台電腦的運算能力，來分擔所有的工作，以縮短整體繪製的時間。
　　本系統是藉由分割畫面，將動畫中每個必須計算的畫面分成數個區塊，並把這些區塊交由後端的每一台電腦去計算，最後再由主電腦組合所有的回傳結果，以完成整段動畫繪製的過程。由於每台後端電腦的效能可能不一致，或者所分配到的區塊其計算複雜度不一樣，因此有些後端電腦可能會出現閒置的狀態，為了進一步增進系統的整體效能，必須讓每台後端電腦隨時處於工作狀態，因此我們加入了負載平衡的能力，一旦後端電腦計算完一區塊後，便向主電腦要求下一個尚未計算的區塊。

　　Many scenes are constructed by 3-D objects in computer animation. The animation is composed of about 25 to 30 frames per second. We must transform 3-D scene to 2-D image on the screen according to the scenario for each frame. Many algorithms have developed to simulate the effect of shadow in real world for rendering 3-D scene. In the thesis, we adopt ray-tracing algorithm to do this work. The ray-tracing calculates the color of pixels on the screen. Unfortunately, it would cost a lot of time to calculate a frame on a computer. Therefore, we use cluster system that groups computers to do the computation.
　　In the purposed system, we divide up the frames on the frontend computer and distribute the partitions among backend computers. Each backend computer calculates the color of pixels on the block and transmits the result to the frontend computer. Once frontend computer recevices all blocks belonging to the same frame, it combines them to a frame. Furthermore, the backend computers may have different performance and the blocks may have different complexity of scene so that some backend computers may wait for the others and be idle. In order to improve the system performance, each backend computer must keep working at any time. We consider the load balance problem and propose a strategy to reach load balance in the system. Experimental results are also presented and analyzed in the thesis.

[1] Arvo James, & Kirk David. 1989. A Survey of Ray Tracing Acceleration Techiques.In: Glassner Andrew S. (ed), An Introduction to Ray Tracing. Academic Press.
[2] Cook Robert L. 1989. Stochastic Sampling and Distributed Ray Tracing. In: Glassner Andrew S. (ed), An Introduction to Ray Tracing. Academic Press.
[3] Foley James D., van Dam Andries, Feiner Steven K., & Hughes John F. 1999. Computer Graphics: Principles and Practice. Addison Wesley. Chap. 15-16.
[4] Glassner Andrew S. 1989. Surface Physics for Ray Tracing. In: Glassner Andrew S.(ed), An Introduction to Ray Tracing. Academic Press.
[5] Donald Hearn, M.Pauline Baker Computer Graphics C version Prentice Hall
[6] 3D Lighting History, Concepts, &Techniques Charles River Media
[7]C. Montani, R. Perego, R. Scopigno Parallel Volume Visualization on a Hypercube Architecture
[8] Phong Bui Tuong. 1975. Illumination for computer generated pictures. Communications of the ACM, 18(6).
[9] Watt Alan, & Policarpo Fabio. 1998. The Computer Image. Addison-Wesley.
[10] Woo Mason, Neider Jackie, Davis Tom, & Shreiner Dave. 1999. OpenGL Programming Guide. Addison-Wesley. Chap. 5.
[11] R. Cook, T. Porter, and L. Carpenter. Distributed Ray Tracing. Computer Graphics (Proceedings of SIGGRAPH 84), 18(3):137–145, 1984.
[12] Edward Angel Interactive computer graphics:A top-down approach with OpenGL Addison-Wesley
[13]Juha Ranta 2002 Ray Tracing
[14] R. Buyya, “High Performance Cluster Computing: System and Architecture”, Vol. 1, Prentice Hall PTR, NJ, 1999
[15] R. Buyya, “High Performance Cluster Computing: Programming and Applications”, Vol. 1, Prentice Hall PTR, NJ, 1999