By fitting a flexible stellar anisotropy model to the observed surface brightness and line-of-sight velocity dispersion profiles of Draco we derive a sequence of cosmologically plausible two-component ( stars + dark matter ) models for this galaxy .
The models are consistent with all the available observations and can have either cuspy Navarro-Frenk-White or flat-cored dark matter density profiles .
The dark matter halos either formed relatively recently ( at z \sim 2 \dots 7 ) and are massive ( up to \sim 5 \times 10 ^ { 9 } M _ { \odot } ) , or formed before the end of the reionization of the universe ( z \sim 7 \dots 11 ) and are less massive ( down to \sim 7 \times 10 ^ { 7 } M _ { \odot } ) .
Our results thus support either of the two popular solutions of the ‘ ‘ missing satellites ’ ’ problem of \Lambda cold dark matter cosmology – that dwarf spheroidals are either very massive , or very old .
We carry out high-resolution simulations of the tidal evolution of our two-component Draco models in the potential of the Milky Way .
The results of our simulations suggest that the observable properties of Draco have not been appreciably affected by the Galactic tides after 10 Gyr of evolution .
We rule out Draco being a ‘ ‘ tidal dwarf ’ ’ – a tidally disrupted dwarf galaxy .
Almost radial Draco orbits ( with the pericentric distance \lesssim 15 kpc ) are also ruled out by our analysis .
The case of a harmonic dark matter core can be consistent with observations only for a very limited choice of Draco orbits ( with the apocentric-to-pericentric distances ratio of \lesssim 2.5 ) .