We present new exact-repair regenerating code constructions that lead to improved achievable  bounds on the storage/repair bandwidth tradeoff region in distributed storage systems.