Current Proceedings on Technology
Yazarlar: Peter Svec
Konular:-
Anahtar Kelimeler:IPv6 protocol,IPv4 protocol,Measurement methodology,Data analysis,Step forward regression
Özet: Simulation modeling of computer networks is an effective technique for evaluating the performance of network and data transfer with various internet protocols. We compare and describe already realized experiments focused on the comparison of IPv4 and IPv6 protocol. There is a gap in the literature in the evaluation of other transport layers protocols and therefore we focus beside the TCP and UDP also on SCTP that combines possibilities of these two ones. Our aim is to develop methodology which will monitor not only the transmission duration but also the load of communicating nodes. The first step of this methodology is represented by the network test-bed and by defining and detains experiment conditions. In the second step we define options for network traffic generator and place emphasis on the variable representing the time between departures of two succeeding packets and the size of these packets. We also monitor selected load variables of sending and receiving node and connected routers. The time of packet sending and packet size are modeled as independent and identically distributed variables. To analyze collected data we used analysis of variance. We watch the influence of different IP traffic to transmission time so to load of all nodes and routers. Analysis results show that routers are under higher load when we use IPv6; transmission time is lower or equal when use IPv6 regardless used transport layer protocol. The load of communicating nodes if higher or equal one when we use IPv6. If we want to evaluate impact of every examined factor to transfer time we have to apply forward stepwise regression analysis. This analysis allows in each step of regression add independent variables into model until the best model is obtained. Our aim isn’t the best model but just estimation of supply of each examined variable for the variability of dependent variable Time.