In the seek of more computing power, two sources of complexity are to face. On one hand, it is possible to aggregate a large amount of computing power (and storage) at the price of very complex resources such as grids. On the other hand, such available computing power allows to imagine more complex applications such as code coupling applications to achieve more realistic simulations. Component models appear as a solid foundation to handle simultaneously both sources of complexity. However, component models need to provide adequate abstractions to offer a simple programming model while enabling high performance on any kind of resources. This paper reviews several abstractions dedicated for scientific applications: data sharing between components, master-worker relationships, parallel to parallel component communications and collective communications among components.