It's absolutely no problem at all. This is the only way you could successfully complete the proof with inference and replacement rules.
Basically, we derive (A&B)v(~A&~B) from A<-->B by means of the equivalence rule.
However, equivalence is a replacement rule. Still part of the basic rules, but more difficult to work with in bigger proofs.
Equivalence basically states that whenever P<-->Q is given, (P&Q)v(~P&~Q) or (P-->Q) & (Q-->P) can replace it... and vice versa because they basically mean the same thing.