Ghassane Aniba

Adaptive proportional fairness for packet scheduling in HSDPA

This paper considers packet scheduling in high speed downlink packet access (HSDPA) networks. One of the main features of HSDPA is the capability of tracking fast channel variations and the use of a large set of discrete rate values, which should be used to conduct fast scheduling of packets while ensuring fairness between users. We consider the operating environment where the scheduling is performed in heterogeneous channels. In this case, proportional fairness (PF) scheduling fails to achieve the goal of providing fair throughput to the users. We propose, in this paper, an approach that resolves this shortcoming. The proposed scheduling algorithm, called adaptive proportional fairness (APF) is shown to ensure proportional fairness even under different QoS requirements for users experiencing different channel conditions. Taking into consideration the systems constraints on the available rates, simulation results and comparisons show the high efficiency of our approach compared to PF scheduling.

Adaptive scheduling for MIMO wireless networks: cross-layer approach and application to HSDPA

In this paper, we consider the scheduling problem in multiple-input multiple-output (MIMO) wireless networks. The main important characteristic of an optimal scheduler is to maximize throughput while servicing users in a fair manner. Herein, we formulate MIMO scheduling as a generalized assignment problem (GAP) and propose a general solution for the GAP, namely, a cross-layer MIMO scheduler (CMS), which uses a novel adaptive proportional fairness (APF) mapping approach in conjunction with a new fast transmit antenna selection (FTAS) technique, to determine the set of users to transmit to and the antenna over which the data associated to each user should be transmitted. The proposed scheduler is applied for packet transmission in high-speed downlink packet access (HSDPA), taking advantage of the use of adaptive modulation and coding while coping with the constraints on the maximum number of simultaneous codes a user equipment can support, the limited uplink signalling, and the absence of fast power control. Numerical results show that the proposed CMS provides up to 70% increase in total throughput compared to other scheduling schemes

Cross-Layer Designed Adaptive Modulation Algorithm with Packet Combining and Truncated ARQ over MIMO Nakagami Fading Channels

This paper presents an optimal adaptive modulation (AM) algorithm designed using a cross-layer approach which combines truncated automatic repeat request (ARQ) protocol and packet combining. Transmissions are performed over multiple-input multiple-output (MIMO) Nakagami fading channels, and retransmitted packets are not necessarily modulated using the same modulation format as in the initial transmission. Compared to traditional approach, cross-layer design based on the coupling across the physical and link layers, has proven to yield better performance in wireless communications. However, there is a lack for the performance analysis and evaluation of such design when the ARQ protocol is used in conjunction with packet combining. Indeed, previous works addressed the link layer performance of AM with truncated ARQ but without packet combining. In addition, previously proposed AM algorithms are not optimal and can provide poor performance when packet combining is implemented. Herein, we first show that the packet loss rate (PLR) resulting from the combining of packets modulated with different constellations can be well approximated by an exponential function. This model is then used in the design of an optimal AM algorithm for systems employing packet combining, truncated ARQ and MIMO antenna configurations, considering transmission over Nakagami fading channels. Numerical results are provided for operation with or without packet combining, and show the enhanced performance and efficiency of the proposed algorithm in comparison with existing ones.

Efficient support of IP multicast in the next generation of GEO Satellites

Satellites are expected to have an important role in providing the Internet protocol (IP) multicast service to complementing next-generation terrestrial networks. In this paper, we focus on the deployment of IP multicast over the next generation of digital video broadcasting-based geosynchronous earth orbit satellites supporting multiple spot beams and on-board switching technologies. We propose a new encapsulation scheme optimized for IP multicast, which has two distinct modes enabling two alternative on-board switching approaches: the self-switching and the label-switching. We also detail a set of mechanisms and protocols for ground stations, as well as for the on-board processor to allow an efficient multicast forwarding in this type of environment, while reducing the load of control and data messages in the satellite segment, and building efficient multicast delivery trees reaching only the spot beams containing at least one member of the corresponding multicast session. To integrate satellite links in the terrestrial Internet, we present satellite multicast adaptation protocol (SMAP), a protocol which is implemented in satellite stations to process incoming protocol independent multicast-sparse mode (PIM-SM) messages sent by terrestrial nodes to the satellite system. SMAP helps to update the tables required for the mapping between IP packets and MPEG-2 data segments, their switching on board the satellite, and their filtering at the satellite receivers.

Fast packet scheduling assuring fairness and quality of service in HSDPA

We consider packet scheduling in high speed downlink packet access (HSDPA) networks in the presence of heterogeneous channels. In this case, proportional fairness (PF) scheduling and its enhanced version, the data rate control (DRC) exponent rule, fail to achieve the goal of providing fair data rates to users. We propose a new scheduling policy that resolves this problem. The proposed adaptive proportional fairness (APF) scheduling is shown to ensure proportional fairness even for users experiencing different channel conditions. The APF algorithm is subdivided into two modules: a short term module, which consists of an enhanced version of the selection criterion adopted in the DRC exponential rule, and a long-term monitoring module, in which we have updating of the control parameters that we introduce to ensure fairness among users. Simulation results and comparisons, provided for the best-effort mode of operation, show the high efficiency of our approach compared to proportional fairness scheduling.

BER Analysis of M-QAM with Packet Combining Over Space-Time Block Coded MIMO Fading Channels

We analyze the bit error rate (BER) performance of M-ary quadrature amplitude modulation (M-QAM) when using space-time block coding (STBC) along with packet combining triggered by automatic repeat request (ARQ) retransmission over multiple-input multiple-output (MIMO) fading channels. Specifically, adopting a log-likelihood ratio (LLR) based approach and considering the 16-QAM case of study, we provide an exact formulation for the aggregate LLR distribution in the case the STBC codeword can be transmitted twice, and derive the resulting BER. For higher number of retransmissions, an approximation of the error function is used to derive the LLR distributions and the systems ensuing BER. Considering different values of combined transmissions and M-QAM with possible constellation rearrangement (CoRe), validation of the proposed BER analytical model through simulations and assessment of the advantages of packet combining are provided for transmissions over additive white Gaussian noise (AWGN) channel and orthogonalized MIMO Rayleigh fading channels with different STBC mappings.

A general traffic and queueing delay model for 3G wireless packet networks

This paper considers traffic modelling and queueing delay estimation of different 3G packet network services. First, a general traffic model for conversational (voice) and streaming (video) services is presented. This model is subdivided into three levels: session level, burst level, and packet level. Based on the proposed model, the statistical behavior of the queueing delay is studied. It is shown that the queueing delay corresponding to voice and video streaming services follows an exponential distribution. Analytical modelling of the probability density function (PDF) of the queueing delay being untractable, we resort to simulated data and provide simple mathematical formulation of the different parameters that characterize the density functions of the different services. Indeed, we present useful equations which could be utilized, directly in network dimensioning, as a reference to satisfy a certain quality of service (QoS) and in the design of radio resource management algorithms.

Dynamic assignment of renewable energy tokens in a collaborative microgrids

Nowadays, electrical energy consumption and energy prices have increased. Thus, the microgrids (μGs) rely on renewable energy receive more attention from consumers. New neighborhoods opt for green philosophy where the most consumed energy comes from renewable generation. Indeed, each residence could be supplied from its own solar and wind generation or from centrally located power plant. However, the latters present many problems in efficiency and reliability. This paper proposes a new approach based on a dynamic assignment of renewable energy tokens (DARET) algorithm to add a smart behavior to μGs. The proposed algorithm allows residential μGs of a small geographic area to dynamically collaborate and share their individual green energy generation in order to supply their overall load. The sharing is dynamically updated in short-term by exchanging data, in terms of individual demands and supplies, between the consumers over wireless links.

Resource allocation in HSDPA using best-users selection under code constraints

We consider the open issue of resource allocation in HSDPA networks for the purpose of enhancing the systems performance both in terms of throughput and fairness while taking into consideration resource constraints specific to the HSDPA architecture. In particular, we propose a two-best user scheduling approach with an optimal power allocation aimed at maximizing data throughput and a selection criterion designed to ensure adaptive proportional fairness between users with different resource requirements and constraints. Compared to the popular carrier-to-interference ratio (CIR) and proportional fairness (PF) methods, the proposed technique, called two-best adaptive proportional fairness (APF) is shown to provide higher performance both in terms of throughput and fairness even when users experience different channel propagating conditions.

Cross-layer design for scheduling and antenna sharing in MIMO networks

This paper formulates the scheduling problem in MIMO networks as a generalized assignment problem (GAP), and advances a new cross-layer design for the scheduling of users and the assignment of their corresponding data to the available transmit antennas. The proposed scheduling and antenna sharing method, referred to as fast transmit antenna selection (FTAS), uses adaptive proportional fairness (APF) mapping as a means to determine the user-antenna assignment that maximizes the network performance both in terms of throughput and fairness. The proposed scheduler is applied in a high speed downlink packet access (HSDPA) network, taking advantage of an inherent HSDPA characteristic, namely, the use of adaptive modulation and coding, while coping with the imposed maximum number of simultaneously supported codes and the absence of fast power control. Numerical results show that our scheduler provides up to 70% increase in total throughput compared to other scheduling schemes applied to HSDPA