Abstract: The class of stream processing applications represents different important processing domains (e.g., image, video, audio, and unstructured data produced by sensors). The project challenge is to enable these applications to run in cloud computing environments rather than running in traditional high-performance processing centers. The cloud computing model enables greater control over the resources and the user can pay-per-use. The elasticity (increase and decrease the use of computational resources) of these environments allows users to save money when stream processing applications support the proper on-demand resource provisioning. This scenario is quite interesting and appropriate to be investigated in these applications since the intensity of the data flow varies and the computational cost of the operation is often irregular for each item/task being processed. Our goal in this project is to investigate different ways of exploiting parallelism and automatic elasticity. In addition, we intend to provide efficient stream parallelism support for a set of applications from this domain in the cloud.