Homoskedastic Mean?

Homoskedastic is an adjective describing a statistical model in which the errors are drawn from the same distribution for all values of the independent variables. 

This is a strong assumption, and includes in particular the assumption in a linear regression, for example,

y = Xb + e

that the variance of the e's is the same for all X's.

The observed variance will differ in almost any sample. But if one believes that the data-generating process does not in principle have greater variances for different values of the independent variable, one would describe the sample as homoskedastic anyway.