Analytical Performance Evaluation of Mixed Services with Variable Data Rates for the Uplink of UMTS

Abstract

The contribution of this paper is twofold. First, we introduce a new efficient analytical method for uplink performance estimation in multi-service UMTS networks by taking the dynamic behaviour of the user transmission rate into account. Our performance measures are the average user information rate and the blocking probability of new calls. Next, a traffic management technique is applied to UMTS and incorporated into the analytical approach to specify how resources have to be adequately engineered to meet quantitative performance objectives.For this we investigate the joint behaviour of the admission control characteristics, scheduling, threshold and reservation scheme for UMTS Terrestrial Radio Access Network (UTRAN) in terms of their agreement with Grade of Service (GoS) requirements. The performance of the proposed algorithms is evaluated.

Call-Level Analysis of W-CDMA Networks Supporting Elastic Services of Finite Population

“…2 shows, the DTMC describing the system is irreversible (all arrows are valid -ignore the X's on the arrows, at the moment); therefore, this system does not have a product form solution. By introducing the following three approximations (used also in [3] and [9]), we modify the DTMC, in order to make it reversible and consequently (in the next subsection) determine the state probabilities, q(j), recursively: a) Local balance approximation (LBA): local balance exists between adjacent system states (rate up = rate down). b) Migration approximation (MA): calls accepted with other than the maximum bandwidth requirement are negligible within a space, which is called migration space.…”

“…A single base station (BS) controlling a cell is modeled as a system of certain bandwidth capacity, in the uplink direction. The number of users roaming in the cell or in its vicinity is usually considered infinite [2], [3]; however, due to the limited coverage area of a cell, it is more realistic to be considered finite. Therefore, for each service-class we assume a quasi-random call arrival process (loosely speaking: random arriving calls generated by finite number of sources) [4].…”

“…In [2], only Poisson stream traffic is considered in the uplink of a W-CDMA cell. In [9], the infinite source population version of WF-CDTM is presented (called Wireless Threshold Model (WTM)); a similar model based on the CDTM ( [10]) is found in [3], for different CAC. In [11], a teletraffic model is presented for cellular networks, when they support Poisson elastic traffic (where in-service calls can alter the assigned bandwidth); however, the model does not result in efficient calculations.…”

Performance analysis of CDMA-based networks with interference cancellation, for batched Poisson traffic under the bandwidth reservation policy

“…The latter refers to a link that accommodates Poisson arriving calls of K service-classes with different bandwidth requirements and generally distributed service times; we name this model Erlang Multirate Loss Model (EMLM). In [7], [9] and [13], calls follow a Poisson process, while the models of [6], [10] cover Poisson, Bernoulli and Pascal traffic, based on the model of [16]. In [8], [11] and [12] calls come from a finite number of users.…”

“…In [8], [11] and [12] calls come from a finite number of users. Apart from the different arrival processes considered in [6]- [13], these models also differ in the modelling of local blocking; e.g., the modelling approach of [6], [10] is more complicated compared to [7], as it is shown in [11]. Therefore, herein, we adopt the model of [7] for the local blocking.…”