Remco Litjens

Elastic calls in an integrated services network: the greater the call size variability the better the QoS

We study a telecommunications network integrating stream calls and elastic calls that share a given number of available traffic channels. The prioritised speech calls are handled as in a traditional Erlang loss model, while the delay-tolerant elastic calls are served with the remaining (varying) service capacity according to a processor sharing (PS) discipline. The remarkable observation is presented and analytically supported that the expected elastic call holding time is decreasing in the variability of the elastic call size distribution. This phenomenon is demonstrated for both Weibull and Pareto elastic call size distributions. Since the reverse effect is known to hold if elastic calls are served in a first-in first-out (FIFO) manner, an extended model is considered which limits the number of service positions in the elastic calls PS queue and holds excessive calls into a feeding FIFO queue. The impact of the elastic call size variability is demonstrated to depend on the number of service positions in the PS queue, as this determines whether the PS or the (reverse) FIFO effect is dominant. As a consequence of the principal result, network planning guidelines or admission control schemes that are developed based on deterministic or lightly variable elastic call sizes are likely to be conservative and inefficient, given the commonly acknowledged property of e.g. WWW documents to be heavy-tailed. Application areas of the models and results include fixed TCP/IP or ATM networks and mobile cellular GSM/GPRS and UMTS networks.

HSDPA flow level performance: the impact of key system and traffic aspects

We present a flow-level performance evaluation for a umts hsdpa network. The results provide thorough insights into the performance impact of a number of key system, environment and traffic aspects, e.g. terminal location, the presence of intercellular interference, the intrinsic feedback delay in the channel quality indications, soft combining of retransmissions and the applied packet scheduling scheme. The contribution of the identified key aspects in the experienced service quality and spatial fairness is assessed by evaluating gradually more complete scenarios.

Throughputs in processor sharing models for integrated stream and elastic traffic

We present an analytical study of throughput measures in processor sharing queueing systems with randomly varying service rates, modelling a communication link in an integrated services network carrying prioritised stream traffic and elastic traffic. A number of distinct throughput measures for the elastic traffic are defined and analysed. In particular, the differences between the various throughput measures and the impact of the elastic call size distribution are investigated. It is concluded that the call-average throughput, which is most relevant from the user point of view but typically hard to analyse, is very well approximated by the newly proposed so-called expected instantaneous throughput, which can easily be obtained from the systems steady state distribution.

Performance of Cloud Computing Centers with Multiple Priority Classes

In this paper we consider the general problem of resource provisioning within cloud computing. We analyze the problem of how to allocate resources to different clients such that the service level agreements (SLAs) for all of these clients are met. A model with multiple service request classes generated by different clients is proposed to evaluate the performance of a cloud computing center when multiple SLAs are negotiated between the service provider and its customers. For each class, the SLA is specified by the request rejection probabilities of the clients in that class. The proposed solution supports cloud service providers in the decision making about 1) defining realistic SLAs, 2) the dimensioning of data centers, 3) whether to accept new clients, and 4) the amount of resources to be reserved for high priority clients. We illustrate the potential of the solution by a number of experiments conducted for a large and therefore realistic number of resources.

Flow level performance comparison of packet scheduling schemes for UMTS EUL

The Enhanced Uplink (EUL) is expected to provide higher capacity, increased data rates and smaller latency on the communication link from users towards the network. A key mechanism in the EUL traffic handling is the packet scheduler. In this paper we present a performance comparison of three distinct EUL scheduling schemes (one-by-one, partial parallel and full parallel) taking into account both the packet level characteristics and the flow level dynamics due to the random user behavior. For that purpose, we develop a hybrid analytical/simulation approach allowing for fast evaluation of performance measures such as mean file transfer time and fairness expressing how the performance depends on the users location.

Analysis of flow transfer times in IEEE 802.11 wireless LANs

In this paper we present an integrated packet/flow level modelling approach for analysing flow throughputs and transfer times inieee 802.11wlans. It captures the statistical characteristics of the transmission of individual packets at themac layer and takes into account the system dynamics due to the initiation and completion of data flow transfers. In particular, at the flow level the system is modelled by a processor sharing type of queue, reflecting theieee 802.11mac design principle of distributing the transmission capacity fairly among the active flows. The integrated packet/flow level model is analytically tractable and yields a simple approximation for the throughput and flow transfer time. Extensive simulations show that the approximation is very accurate for a wide range of parameter settings. In addition, the simulation study confirms the attractive property following from our approximation that the expected flow transfer delay is insensitive to the flow size distribution (apart from its mean).RésuméNous présentons dans cet article un modèle intégrant les niveaux paquets et flux, qui permet l’analyse du débit utile des flux et de leur temps de transfert dans des réseaux locaux sans filieee 802.11. Le modèle développé prend en compte les caractéristiques statistiques de la transmission de paquets individuels au niveau de la couchemac, ainsi que la dynamique du système liée à l’initialisation et à la terminaison des transferts de flux de données. Plus précisément, au niveau des flux, le système est modélisé à l’aide d’une file d’attente à processeur partagé, refletant ainsi le partage équitable de la capacité de transmission entre les différents flux actifs, tel que réalisé par la couchemac de l’ieee 802.11. Le modèle à deux niveaux, paquets/flux, peut être résolu analytiquement et fournit une approximation simple du débit utile et du temps de transfert de flux. De nombreuses simulations ont montré que, pour une large gamme de paramètres, l’approximation est très précise. De plus, les résultats de nos simulations vérifient une propriété intéressante découlant de notre approximation à savoir que le delai de transfert moyen d’un flux est indépendant de la distribution des tailles (seule sa moyenne compte).

Performance Analysis of Fair Channel Sharing Policies in an Integrated Cellular Voice/Data Network

We study channel sharing in an integrated cellular voice/data network with a finite queue for data call requests that cannot be served immediately upon arrival. Using analytical techniques, a comparison of different fair channel sharing policies is made. As a main result, a closed-form expression is derived for the expected sojourn time (waiting time plus transfer time) of a data call, conditional on its size, indicating that the sojourn time is proportional to the call size. This attractive proportionality result establishes an additional fairness property for the channel sharing policies proposed in the paper. Additionally, as a valuable intermediate result, the conditional expected sojourn time of an admitted data call is obtained, given the system state at arrival, which may serve as an appreciated feedback information service to the data source. An extensive numerical study is included to compare the proposed policies and to obtain insight in the performance effects of the various system and policy parameters.

The impact of QoS differentiation on service quality and system capacity in cellular networks

Mobile network operators are currently preparing the introduction of QoS differentiation mechanisms in their 3G networks, as a means to support service-specific quality requirements, to enable the offering of distinct subscription forms and/or to enhance the resource efficiency and hence network capacity. In this paper we quantify the attainable performance and/or capacity gains from introducing different forms of QoS differentiation in an HSDPA network, considering distinct scenarios of subscription- and/or service-based differentiation. For the case of weight-based QoS differentiation, we demonstrate how the significance of the achievable performance impact and capacity gains depends on the applied differentiation weights, the traffic characteristics, the service mix and the class-specific performance requirements. For scenarios with reasonable traffic mixes and performance requirements, capacity gains of about 20xa0% appear very well possible. For the investigated minimum bit rate-based QoS differentiation scheme, we demonstrate an inherent trade-off between giving high-priority data users preferred access to the shared transport channel, while as a consequence reducing the spectral efficiency of the channel by granting prioritised users access even at times when their channel conditions are unfavourable. Achievable performance gains were primarily observed at cell traffic loads exceeding the ‘dimensioning load’, although for some scenarios capacity gains of up to 8.6xa0% were found, compared to a case without QoS differentiation.