Proceedings of the International Conference on Computer and Communication Engineering 2008 May 13-15, 2008 Kuala Lumpur, Malaysia 978-1-4244-1692-9/08/$25.00 ©2008 IEEE Mixed Traffic Modeling for Wireless Cellular Network A.B.M.Siddique Hossain 1 , Md. Shahriar Karim 2 , Md. Imdadul Islam 3 1 Faculty, Bangladesh University of Engineering and Technology 2 Technical Staff, Ericsson Bangladesh Ltd. 3 Faculty, Jahangirnagar University, Bangladesh E-mail: shahriar.karim@ericsson.com, shahriar.karim@yahoo.com Abstract Any cell of a network experiences two types of call arrival; new originating call, and handoff call from surrounding cells. If the number of busy users around a cell is finite with an average value of N, then rest of the radio resources of the network can be assigned to support handoff calls. So, for handoff calls, the traffic case resembles a network with limited user traffic (PCT-II), where as rest of the radio resources handle unlimited user traffic (PCT-I). In this paper, a mixed traffic model (combination of PCT-I and PCT-II) has been considered to evaluate different traffic parameters like, probability state and call blocking of a cellular network. For analysis, here, two- dimensional state transition diagram as well as tabular method is considered. I. INTRODUCTION The service area of a mobile network is subdivided into smaller hexagonal parts. According to network terminology, this small hexagonal area is known as “cell” and varies based on their dissimilitude in sizes. For example, a cell can be defined as pico cell, micro cell, macro cell etc. and even as satellite cell based on the area they cover. So, in general, service area of any wireless mobile network consists of a number of cells with various names and sizes. Radio coverage in each cell is provided by installing Base Transceiver Station (BTS), which ensures the availability of radio resources in terms of channels. So, a fixed number of channels (carrier frequency for FDMA, carrier frequency and time slot for TDMA and carrier frequency and code for CDMA case) are available for each cell to support both new originating and handoff calls, and when a call is made, a channel is ceased accordingly to establish the call path. Let us consider a cell where new originating traffic follows PCT-I (Pure Chance Traffic-I, with unlimited users) i.e. M/M/n traffic modeling. Handover arrival in such a cell depends mainly on average number of busy users surrounding that cell. For analysis, it is presumed to have N users in service around a cell i ; therefore, handoff arrival in that cell follows PCT-II traffic (Where both number of users and servers are limited) i.e. M/M/n/K/N traffic modeling for n k = like [3]-[5]. For a network with aforesaid arrangement, mixed-traffic concept is more applicable in analyzing the traffic behavior of the network. In this paper, two-dimensional Markovian chain with PCT-I traffic in x direction (for new arrival traffic) and PCT-II traffic in y direction (for handover traffic) is proposed based on [6]-[7]. The model analyzed in the paper is equally applicable for a circuit switch network as well as for a packet-switched network. However, for a packet switched network, when the service time for a packet is fixed, the model can further be modified to M/D/n/K/N traffic case according to [8]. The rest of the paper is structured as follows. In section II, the proposed model is presented as a two dimensional Markovian chain. Later, generalized equations are derived with the help of cut equations. In section III, an appropriate mathematical example is taken and is explained to justify the whole analysis. Results are enumerated in Section IV along with discussion on obtained results. This section also includes graphical representation of blocking probability for PCT-I and PCT-II traffic model and their comparison. We finish with conclusion, given in Section V. II. METHODOLOGY The proposed model is shown in Figure 1 in a two- dimensional Markovian chain form, where traffic along x axis is PCT-I type with arrival rate λ 2 450