This thesis presents a methodology for performance evaluation of broadband access networks. The performance evaluation of networks is a way to identify and analyze how certain features, such as different traffic types or forms of use, can influence the behavior of the network in focus. This way, we can predict how the network
will behave in future situations. The methodology is composed of two approaches: an approach based on measurements and the other based on modeling using Markov processes. The networks analyzed cover the two basic types of access architecture: ADSL2+ (asymmetric digital subscriber line 2+) networks, which are wired networks
that use twisted pair cables; FBWN (fixed broadband wireless network) networks, which are wireless networks based on IEEE 802.16 standard. The measurement approach is focused on network behaves against three situations: transmission of a generic traffic, impact of non-stationary noise in the system, and using network for the transmission of real-time multimedia traffic. The modeling approach, in its turn, is based on predicting the behavior of the network using mathematical models based on Markov processes. The results presented indicate the feasibility of applying this
methodology as form of performance evaluation. The results also make possible the extension of this methodology to other types of broadband access networks such as fiber optic networks, microwave links, and VDSL/VDSL2 (Very-high-data-rate DSL) networks.
|
