As the human civilization relies on more and more advanced - frequently space-based - technologies, various natural factors, which have not been significant in the past, come now into play. The variations of plasma and magnetic field in the surroundings of the Earth - commonly known as the 'space weather' - represents a clear example. Since the main driver of the space weather has to be found in the solar activity - namely CMEs and flares - detailed understanding solar eruptions is necessary for space weather predictions.It is commonly accepted that the formation of current layers and subsequent magnetic reconnection play a key role for the change of magnetic field topology in eruptions and energy dissipation in flares. Nevertheless, many questions remain open in research of this process. Namely, one fundamental problem has not been resolved yet: The issue of energy transport from large to small scales. It is known, that free magnetic energy is accumulated on much larger scales (~1000km) along the current layer formed behind the ejecta than is the typical predicted width of dissipative (kinetic) current sheets (~10m in solar corona). The question arises, how to bridge this enormous scale gap - what is the nature of energy transport from large to the dissipation scales. This issue is closely related to the enigmatic duality between coherent large-scale structures and signatures of fragmented small-scale energy dissipation observed simultaneously in solar flares.In the proposed project we plan to study this fundamental problem using - above all - advanced numerical simulations that extend over a broad range of scales. As a follow-up of our modeling we shall formulate the model-specific results of the simulations in the form directly comparable with observations, namely in radio and X-ray domains, and use the targeted observations as the model tests. The project aims at solar eruptions but its results are relevant for broad-scale magnetic reconnection in general.