SCALABLE ARCHITECTURES FOR INTEGRATED TRAFFIC SHAPING AND LINK SCHEDULING IN HIGH-SPEED ATM SWITCHES

Emerging broad-hand switches must accommodate the diverse traffic para meters and quality-of-service requirements of voice, data, and video a pplications. End-to-end performance guarantees depend on connections c omplying with traffic contracts as their cells travel through the netw ork. This paper presents a leaky-bucket shaper architecture that scale s to a large number of connections with diverse burstiness and bandwid th parameters, In contrast to existing designs, the proposed architect ure arbitrates fairly between connections with conforming cells by car efully integrating leaky-bucket traffic shaping with rate-based schedu ling algorithms, Through a careful combination of per-connection queue ing and approximate sorting, the shaper performs a small, bounded numb er of operations in response to each arrival and departure, independen t of the number of connections and cells, When the shaper must handle a wide range of rate parameters, a hierarchical arbitration scheme can reduce the implementation overheads and further limit interference be tween competing connections. Through simulation experiments, we demons trate that the architecture limits cell-shaping delay and traffic dist ortions, even in periods of heavy congestion, The efficient combinatio n of traffic shaping and link scheduling results in an effective archi tecture for managing buffer and bandwidth resources in large, high-spe ed ATM switches.