Showing papers by "Bernard Cousin published in 2018"

Energy-Aware Routing in Carrier-Grade Ethernet Using SDN Approach

Abstract: Soft-defined networking (SDN) is a new approach that enables operators to easily manage all the network elements. In this paper, we address the problem of energy-aware routing in SDN-based carrier-grade Ethernet networks. Our approach is based on turning off network nodes and links to reduce energy consumption, while respecting the rule space capacity for each Openflow switch, and maintaining an allowable maximum link utilization. The problem of identifying the optimal set of network elements to be turned off is NP-hard. We first present an exact model based on an integer linear programming formulation for the problem. Then, we describe a set of first-fit heuristic algorithms suitable for large-sized networks. The exact and heuristic approaches are tested on SNDlib-based instances. Experimentations show the efficiency of both exact and heuristic methods for different network topologies. In particular, our heuristic algorithms are able to achieve a good balance between energy consumption, resource utilization, and network performance.

Combining traffic-shaping methods with congestion control variants for HTTP adaptive streaming

Abstract: HTTP adaptive streaming (HAS) is a streaming video technique widely used over the Internet. However, it has many drawbacks that degrade its user quality of experience (QoE). Our investigation involves several HAS clients competing for bandwidth inside the same home network. Studies have shown that managing the bandwidth between HAS clients using traffic shaping methods improves the QoE. Additionally, the TCP congestion control algorithm in the HAS server may also impact the QoE because every congestion control variant has its own method to control the congestion window size. Based on previous work, we describe two traffic shaping methods, the Hierarchical Token Bucket shaping Method (HTBM) and the Receive Window Tuning Method (RWTM), as well as four popular congestion control variants: NewReno, Vegas, Illinois, and Cubic. In this paper, our objective is to provide a detailed comparative evaluation of combining these four congestion control variants with these two shaping methods. The main result indicates that Illinois with RWTM offers the best QoE without causing congestion. Results were validated through experimentation and objective QoE analytical criteria.

Optical power control in translucent flexible optical networks with GMPLS control plane

Abstract: The continuously increasing traffic of Internet services (cloud services, video streaming, social networks, and recently, Internet of Things services) is leading to huge traffic growth in core optical networks. This traffic evolution is pushing network operators to efficiently exploit their infrastructures in order to postpone, as much as possible, the expensive deployment of new infrastructures. In this respect, the migration from fixed-to flex-grid optical networks was triggered in order to efficiently use optical network capacity, taking benefits from the improved spectral efficiency of flexible transponders. In our previous work [J. Opt. Commun. Netw., vol. 8, no. 8, p. 553, Aug. 2016], we demonstrated that migrating towards flexible networks while keeping in use existing optical amplifiers will cause a power saturation problem over highly loaded links due to the increase in the number of optical channels. To overcome this problem, we proposed in that work a power adaptation process that consists of converting transmission performance margins into optical power attenuation over optical links. However, the realized work considered only a transparent optical network controlled by the generalized multiprotocol label switching (GMPLS) protocol suite. In this paper, we consider the case of a translucent optical network where optical regeneration is required, and thus, the power adaptation process is adapted to this kind of network. A new routing algorithm and protocol extensions are proposed to take into account power and regeneration information in the GMPLS control plane of translucent networks.

ESMRsc: Energy Aware and Stable Multipath Routing Protocol for Ad Hoc Networks in Smart City

Abstract: The mobile ad hoc wireless networks are characterized by the absence of central administration and any network element may be very mobile. There is no fixe element within an ad hoc network. In fact, within these networks, all elements must cooperate so as to create a temporary topology which facilitates communication. To create this topology and carry data, ad hoc networks must use efficient routing protocols.

User Selection in 5G Heterogeneous Networks Based on Millimeter-Wave and Beamforming

Abstract: Heterogeneous Networks (HetNETs) are considered as an effective solution to improve the coverage and system throughput for future cellular networks. The extremely growing mobile market, together with the arising demand for high data rates, motivate us to open a new spectrum related to millimeter waves (mmwaves) while using beamforming that can serve simultaneously a group of users. In this paper, we formulate an optimization problem for HetNETs multi-user selection in a multi-input-multi-output and orthogonal frequency-division multiple access (MIMO-OFDMA) system, aiming to maximize the total system throughput. We solve the problem by applying a modified version of well-known metaheuristic algorithms. The optimal solution is obtained using an exhaustive search algorithm that provides an ideal solution which is complex to be computed. Greedy zero-forcing dirty-paper gZF-DP and zeroforcing selection ZFS algorithms were selected from literature for the sub-optimal solution. In parallel, a water-filling algorithm has been optimized after adding new power constraint and it has been used for power distribution. Hence, we analyze the throughput performance of our systems using throughput metric. The results show that ZFS outperforms gZF-DP algorithm as it achieves higher total throughput, While gZF-DP outperforms ZFS algorithm in the execution time.

Performance evaluation of TcpHas: TCP for HTTP adaptive streaming

Abstract: HTTP adaptive streaming (HAS) is a widely used video streaming technology that suffers from a degradation of user’s Quality of Experience (QoE) and network’s Quality of Service (QoS) when many HAS players are sharing the same bottleneck link and competing for bandwidth. The two major factors of this degradation are: the large OFF period of HAS, which causes false bandwidth estimations, and the TCP congestion control, which is not suitable for HAS given that it does not consider the different video encoding bitrates of HAS. This paper proposes a HAS-based TCP congestion control, TcpHas, that minimizes the impact of the two aforementioned issues. It does this using traffic shaping on the server. Simulations indicate that TcpHas improves both QoE, mainly by reducing instability and convergence speed, and QoS, mainly by reducing queuing delay and packet drop rate.

Two Node-Disjoint Paths Routing for Energy-Efficiency and Network Reliability

Abstract: Carrier grade networks are in general over-dimensioned in order to protect against network resource failures and to handle probable traffic peaks. Such network design exhibits non-negligible energy waste. As well known, during network operation, the traffic load varies remarkably over day hours. Typically, a significant amount of energy saving can be achieved by turning off redundant and underutilized resources. Ensuring a sufficient level of reliability while taking into account energy saving objective is a very challenging task. In this paper, we study the multi-commodity reliable network design for carrier Ethernet networks. Each traffic demand is routed along one working path, and a corresponding backup path computed considering dedicated protection scheme. The primary and backup path must be node-disjoint. We also assume that the links of the carrier grade networks are made of multiple physical cables called bundles. We propose an exact method based on an Integer Linear Programming (ILP) formulation for the two node-disjoint paths with energy-aware routing (TNDP-EAR) problem. The studied problem is known to be NP-hard. In order to solve it efficiently, we propose a heuristic-based algorithm called GreenTNDP. Preliminary experimental results show the effectiveness of our algorithms to solve the problem.

Neural network approach for component carrier selection in 4G/5G networks

Abstract: In order to ensure high transmission rate to a wide range of voice, video and data services, carrier aggregation (CA) has been introduced in the Long Term Evolution Advanced (LTE-A). Moreover, CA is one of the principal enabling technologies for the 5G. Indeed, 4G and 5G can aggregate up to five component carriers (CC), simultaneously, to support a higher bandwidth. In this context, CC selection method is necessary. In this paper, we propose a new CC selection method in order to maximize the global system throughput. Neural network approach is introduced to select the best couple user-CC via a utility function. Simulation results prove that the proposed method outperforms in terms of system throughput, fairness and packet loss rate.