likelihood function for random effects mode‏l

#1
Can someone give hints on how to formulate a likelihood function for random effects model?
y_ij=mu_i+b_j+e_ij, where i=1, 2 are the mean before and after.j=1, 2, ...n, b_j is random intercept with N(0, 1) distribution. eij~N(0,1).

Should each subject be considered iid? :confused:

Thanks.