Note that you don't have to use Cauchy-Schwartz to get the inequality you have written there. However, I don't think that inequality will help you much.
A simple bound for $n$ can be found by finding an upper bound on the LHS of the equation and a lower bound on the RHS of the equation. That way you can easily show that $n>2561$. (Can you figure out how?)
Now, play around with the equation a bit. Does it work if one of the $x_i$'s is $0$? Does it work if all the $x_i$'s are the same? Or if only $1$ is different? Or if only 2 are different? And so on.
Soon enough, you will find a solution and hopefully it involves $n=2562$, otherwise you have to show that the lower values don't work.