ABSTRACT One of the principal goals of seismology is to infer the nature of an earthquake source from observations of seismic ground motion. This work shall discuss the seismic source both in the 2D finite-fault and in the point-source approximation. By inverting 3-component accelerograms the rupture history and the slip distribution for the Mw 6.3 earthquake occurred in central Italy on April 6, 2009 are determined. The method of linear programming is used for the inversion and the simplex method is applied to solve the linear programming problem (Das and Kostrov, 1994). All known parameters, such as crustal structure and station distribution are kept fixed and a large-enough fault area is considered. Physical constraints such as the positivity of the slip rates on the fault and a pre-assigned seismic moment are used to stabilize the solution. Using synthetic data with a checkerboard slip distribution shows that the obtainable spatial resolution is around 2 km. Observed records acquired from local stations of the national strong-motion network are inverted. Only data from rock stations distributed uniformly around the fault at epicentral distances less than 80 km are used. The accelerograms are filtered at 1 Hz and about 15 seconds of the signals are modelled. The obtained slip distribution shows a single major asperity and is in agreement with other similar studies of the L’Aquila earthquake. The main event of L’Aquila is used to validate a stable and automatic procedure implemented by SeiSRaM group (Dep. of Mathematics and Geosciences, University of Trieste) for the SE Alps transfrontier network to estimate in real time the seismic moment, moment magnitude and corner frequency of events recorded by broad-band velocimeters and accelerometers. The procedure has two steps: the first one consists in an interface with the Antelope system (a software that manages the network) from which pre-processed waveforms are retrieved. The second step consists in estimating the seismic moment and the corner frequency by spectral analysis. The S-wave train is identified through an automatic picking procedure of Antelope software or, if that procedure fails, through the estimates arrival times based on the travel-time. The transversal component of motion is used to minimize conversion effects. The analyzed frequency window is selected on the basis of the signal-to-noise ratio (SNR). The source spectrum is obtained by correcting the signals for geometrical spreading and intrinsic attenuation. For the latter, different relationships are tested for frequency-dependent Q value in order to characterize the anelastic proprieties of the seismic region. Source spectra for both velocity and displacement are computed and, following Andrews (1986), the seismic moment and the corner frequency are estimated. The procedure is successfully validated using the recordings of some recent strong earthquakes like Carnia 2002 (Mw=4.9), Bovec 2004 (Mw =5.1), Parma 2008 (Mw =5.4) and Aquila 2009 (Mw =6.3) and the recording of some minor events in the SE Alps area for which independent seismic moment and Mw estimates are available. Since one year the procedure is applied to events recorded by the National Accelerometric Network (RAN). The agreement between moment magnitudes estimated by the SeiSRaM procedure and the INGV local magnitudes is very good.