Convergence issue in Mosek

I am formulating a SOCP problem, which is giving the optimal solution for a particular input parameter. But, when I am changing the parameters I am not getting solution with Mosek. Also, using Sedumi I am getting solution but the solution obtained is inaccurate. I believe the problem might be due to some variables poor numerical scaling.
How can I make the proper scaling of the problem? I there any technique to do so in CVX?
Looking forward to a few suggestions.


Read this . Also, read sections 7.3 and 7.4 of the MOSEK Modeling Cookbook