Halos formed in the standard \Lambda cold dark matter framework should follow an universal mass density profile and fit a well defined mass-concentration relation .
Lensing analyses of clusters with a large Einstein radius seem to contradict this scenario , with the massive cluster Abell 1689 being often claimed as a notable example of a highly over-concentrated halo .
Shape and orientation biases in lensing studies might be at the basis of this disagreement between theory and observations .
We developed a method for a full three-dimensional analysis of strong and weak lensing data .
Surface density maps estimated from lensing are de-projected to infer the actual triaxial structure of the cluster , whose mass distribution is approximated as an ellipsoidal Navarro-Frenk-White halo with arbitrary orientation .
Inversion is performed under competing a priori assumptions , integrated in the method thanks to Bayesian statistics .
We applied the method to A1689 .
Whatever the considered priors on shape and orientation , both weak and strong lensing analyses found the halo to be slightly over-concentrated but still consistent with theoretical predictions .
We found some evidence for a mildly triaxial lens ( minor to major axis ratio \sim 0.5 \pm 0.2 ) with the major axis orientated along the line of sight .
Exploiting priors from N -body simulations , we found mass M _ { 200 } = ( 1.3 \pm 0.4 ) \times 10 ^ { 15 } M _ { \odot } and concentration c _ { 200 } = 10 \pm 3 for the weak lensing analysis of Subaru data , M _ { 200 } = ( 1.7 \pm 0.3 ) \times 10 ^ { 15 } M _ { \odot } and c _ { 200 } = 6.1 \pm 0.9 for the strong lensing analysis of multiple image systems , and M _ { 200 } = ( 1.3 \pm 0.2 ) \times 10 ^ { 15 } M _ { \odot } and c _ { 200 } = 7.3 \pm 0.8 for the combined weak plus strong analysis .