Due to a faster speed increase of networks than processors, we have today a trend towards the distribution of functionality and workload for network processing on several processing units. For this purpose, network processors are being developed which are special processor architectures aimed for demanding networking tasks such as backbone routing and switching. In this paper, we investigate the possibility of improving the scalability of intermediate nodes by offloading the packet processing workload on the host, and in particular, we present a traffic shaper and packet prioritizer implemented on the Intel IXP2400.