Scaling web Services by offloading expensive computation to a many-core GPGPU using CUDA

Web Services have become the key technology in reaching out to millions of internet users around the world. The success and sustainability of several ecommerce companies hinges on the performance and robustness of their respective websites. In the recent years, the popularity of such e-commerce websites has increased manifold and the number of users is rapidly increasing. In such a scenario, scalability of web services becomes an important aspect. In this paper, we propose a novel approach to enhance the throughput of a web service by utilizing the computing power of general purpose GPUs (GPGPUs).