Distributing applications and service applications among cloud systems offer many potential benefits. However, implementing robust, efficient, and flexible distributed applications is more complex than building stand-alone applications. A significant portion of this complexity is due to the fact that developers must consider new design alternatives and must acquire many new skills. The problem of building such applications requires finding and orchestrating appropriate services that are frequently non trivial for a developer. This is due to the very large number of available services on the cloud and the different possibilities for constructing an application from matching services. We need an abstract view of the application using modeling approaches. In addition, we need an ontological description to retrieve and compose automatically this kind of applications. This abstraction allows in one hand the reuse of the elaborated application and on the other hand reduces the complexity and saves the users from the detail of the low level of the environment. Proposed modeling approaches need a standard verification step against required properties to analyze and correct built applications as early as possible in order to avoid any costly maintenance delays due to runtime errors.