This paper discusses the dimensioning of buffers and the bandwidth all
ocation for data traffic in the ATM network. Data traffic is notorious
ly complex and bursty, making such dimensioning a difficult task. Howe
ver, the COMBINE project, when dimensioning their InterWorking Units (
IWUs), adopted a Poissonian packet arrival model, based upon the argum
ent that burstiness at timescales higher than that of a packet arrival
are a problem to be tackled by flow control at higher layers. This pa
per presents experimental results from the COMBINE testbed that show t
hat this hypothesis was justified and that good TCP goodput was obtain
ed based upon this dimensioning approach, due to TCP's ability to adap
t to network congestion. However, it is also shown that it was the TCP
algorithm that was ultimately responsible for controlling the packet
loss ratio in the network and not the bandwidth allocation or buffer s
ize. The results highlight the importance of taking into account the m
utual influence between the ATM layer and the transport layer congesti
on control algorithms.