No description available.

No description available.

No description available.

No description available.

No description available.

No description available.

"Cloud gaming is a new service based on cloud computation technology which allows games to be run on a server and streamed as video to players on a thin client. Commercial cloud gaming systems, such as Onlive, Gaikai and StreamMyGame remain proprietary, limiting access for game developers and researchers. In order to address these shortcomings, we developed an open source Unity3d cloud-based game system called Uniquitous that gives the game developers and researchers control of system and content. Detailed experiments evaluate performance of three main parameters: game genre, game resolution and game image quality. The evaluation results are used in a data model that can predict in-game frame rates for systems that have not been tested. Validation experiments show the accuracy of our model and allow us to use the model to explore cloud-based games in a variety of system conditions."

As virtualization platforms or cloud computing have become more of a commodity, many more organizations have been utilizing them. Many organizations and technologies have emerged to fulfill those cloud needs. Cloud vendors provide similar services, but the differences can have significant impact on specific applications. Selecting the right provider is difficult and confusing because of the number of options. It can be difficult to determine which application characteristics will impact the choice of implementation. There has not been a concise process to select which cloud vendor and characteristics are best suited for the application requirements and organization requirements. This thesis provides a model that identifies crucial application characteristics, organization requirements and also characteristics of a cloud. The model is used to analyze the interaction of the application with multiple cloud platforms and select the best option based on a suitability score. Case studies utilize this model to test three applications against three cloud implementations to identify the best fit cloud implementation. The model is further validated by a small group of peers through a survey. The studies show that the model is useful in identifying and comparing cloud implementations with regard to application requirements.

A two-dimensional numerical mesoscale model, which ic1udes radiative and turbulent transfers, has been constructed to study the formation, development and dissipation of coastal stratus cloud under an inversion. In the model, the delta-Eddington and emissivity approximations are used for the solar and thermal radiative transfers, respectively. K-theory parameterization is adopted for the turbulent transfer. Ground surface temperature and moisture are predicted using the methods of Deardorff (1977, 1978). This model is applied to a domain which extends 180 km into sea and 240 km inland horizontally and about 2 km from the earth1s surface vertically. A bare, flat soil surface is assumed. As a prelude to the study of the stratus cloud, sea breeze circulations with and without a temperature inversion have been investigated. The model without an inversion yields speeds of the sea breeze front which are close to the values that have been observed under insolation and other meteorological conditions similar to those used in the model. The presence of an inversion causes the depth of sea breeze circulation to be shallower and its inland penetration during the evening hours to be weaker compared to the case without inversion; however, the basic structure of the sea breeze circulation is unaffected by the inversion. The model has been used to study the growth, development and dissipation of stratus cloud under an inversion in the west coast region of the United States. The effects of large scale motions on these processes have also been examined. Cloud parameters such as the times of initial formation and of dissipation inland, the maximum distance of inland penetration, the maximum liquid water path and the cloud-top height are affected in the presence of such large scale motions; for example, both the maximum liquid water path and the cloud-top height are appreciably enhanced - by about a factor of two - when large scale westerly winds (U[subscript g]=5mfs, V[subscript g]=0) are present compared to the case when U[subscript g]=V[subscript g]=0. The cloud parameters predicted by the model are in close correspondence with existing observations in southern California. It is found that the sea breeze circulation is not appreciably affected by the presence of moderate amounts of stratus cloud. While advection plays a dominant role in the horizontal development of the stratus cloud inland, radiative processes (cooling and heating) are observed to govern the vertical growth and dissipation of the cloud layer. Vertical growth is influenced by the rate of radiative cooling at cloud-top. Because of the combined effects of solar and surface heating, the stratus inland is observed to dissipate more rapidly during the morning hours than the cloud over the ocean where surface heating is minimal. / Graduation date: 1984

As mobile network infrastructures continuously improve, they are becoming popular clients to consume any Web resources, especially Web Services (WS). However, there are problems in connecting mobile devices to existing WS. This thesis focuses on three of the following challenge: loss of connection, bandwidth/latency, and limited resources. This research implements and develops a cross-platform architecture for connecting mobile devices to the WS. The architecture includes a platform independent design of mobile service client and a middleware for enhancing the interaction between mobile clients and WS. The middleware also provides a personal service mashup platform for the mobile client. Finally, the middleware can be deployed on Cloud Platforms, like Google App Engine and Amazon EC2, to enhance the scalability and reliability. The experiments evaluate the optimization/adaptation, overhead of the middleware, middleware pushing via email, and performance of Cloud Platforms.