As Cloud computing is becoming a mainstream platform, it has also become important to understand the implications on customers' applications or systems when deployed on Clouds. Therefore, simulation tools become a critical requirement that not only evaluate the performance of Clouds but also help in developing Cloud computing further. However, current simulation solutions for Clouds do not support many important application paradigms such as message-passing parallel applications. This limits the usage of these solutions to study the deployment of many scientific applications on Clouds. In this paper, after recognizing the needs and requirements of the Cloud research and development community, we propose a Cloud simulation environment with a scalable network and message-passing application model that allows accurate evaluation of scheduling and resource provisioning policies and thus helps in optimizing the performance of a Cloud infrastructure. Copyright © 2012 John Wiley & Sons, Ltd.