Feasibility of Supporting Real-Time Traffic in DiffServ Architecture

Abstract

Since Integrated Services architecture is not scalable, it seems the only solutions for Quality of Service (QoS) architecture in the Internet are Differentiated Services (DiffServ) or its variations. It is generally understood that networks with DiffServ architectures can guarantee the end-to-end delay for packets of the highest priority class, only in lightly utilized networks. We show that, in networks without loops, the delay bounds for highest priority packets exist regardless of the level of network utilization with DiffServ architecture. These bounds are quadratically proportional to the maximum hop counts in heavily utilized networks; and are linearly proportional to the maximum hop counts in lightly utilized networks. We argue that, based on the analysis of these delay bounds in realistic situations, DiffServ architecture is able to support real time applications even in large networks. Considering that loop-free networks, especially the Ethernet networks, are being adopted more than ever for access networks and for provider networks as well, this conclusion is quite encouraging. Throughout the paper we use Latency-Rate ($\mathcal{LR}$) server model, with which it has been proved that FIFO and Strict Priority schedulers are $\mathcal{LR}$ servers to each flows in certain conditions.