The main challenge of this application execution environment is to deal with the large variety of resources available at Grid level and their changing conditions over time. In order to deal with this challenge, the Grid level service managing the application will have access to all the information about the availability of such resources, how busy they are (currently and planned) and the costs related to executing applications there.

Another challenge on which the overall design of the XtreemOS services will focus is that the cost to obtain the system information and take the appropriate decisions has to be orders of magnitude less than in current middleware-based systems. This factor will make the crucial difference between our proposal and the existing Grid infrastructures.

Additionally, as the operating system is Grid-aware, the accuracy of the information will be much higher, resulting in a better scheduling control, and allowing implementation of better scheduling policies. Having full control of all nodes, XtreemOS will have the chance to implement VO-wide scheduling policies, enabling co-allocation of applications across multiple sites. It will also be a challenge to take advantage of this accuracy to efficiently offer the Grid level services in a more transparent way to the user.