Performance Analysis of the Uplink of a CDMA Cell Supporting Elastic Services

Abstract

Abstract.We consider a single cell of a multi-service CDMA network, in which some of the service classes are explicit rate controlled. We call these elastic service classes. The instantaneous bit rate of elastic services is dynamically adjusted between a minimum and maximum value such that the system always remains work conserving. We develop a Markov model that allows us to study the impact of such state dependent (dynamic) rate control on the class-wise blocking probabilities and the first two moments of the holding times. We conclude that dynamic (state dependent) rate adjustment decreases the class-wise blocking probabilities and only moderately increases the expectation and the second moment of the time spent in the system.

On Scheduling and Interference Coordination Policies for Multicell OFDMA Networks

“…These two quantities represent a trade-off since more admitted sessions imply lower per-session throughput and thereby longer file transfer times. This trade-off in a WCDMA environment has been investigated by Altman in [6] and subsequently by Fodor et al in [7].…”

On the Tradeoff Between Blocking and Dropping Probabilities in CDMA Networks Supporting Elastic Services

“…The basic CDMA uplink equations that serve as a starting point for this paper are described in details in [3] and [4]. In this section we summarize these results and refer to these references for the derivation of them.…”

“…The power P i that is received at the base station from the mobile terminal for session i must fulfill (see [4]): In practice,Ψ is defined such that the noise rise in the system stays under some predefined threshold, typically less than 7dB. In the single class case it means that the number of admitted sessions must fulfill:…”

“…We build on the model developed for elastic traffic in previous work [4] and extend it with allowing for a state dependent soft blocking and capturing the fact that sessions are sometimes dropped. The main assumption that we make is that the session drop probability is connected to the load of the system.…”

Flow Level Performance Comparison of Packet Scheduling Schemes for UMTS EUL

“…Analytical studies on EUL performance capturing both the packet and flow level dynamics of the system are rare. Interesting references here are [4] and [11]. In particular, in [11] flow level performance metrics are analysed for two (rate-fair) scheduling disciplines assuming that the transmit powers of all mobiles are sufficient to reach the maximum bit rate.…”