Ryad Ben-El-Kezadri

ALBA: An autonomic load balancing algorithm for IEEE 802.11 wireless networks

The widespread success of the WLAN technology raises the need for the efficient management of the different access points and the associated wireless devices, especially when multimedia and other QoS demanding applications are employed. In this paper we propose an autonomic load balancing mechanism for WLAN networks and test its applicability for a Voice over WLAN (VoWLAN) use case. The algorithm takes into account the channel utilization, the spatial distribution of wireless devices and the QoS requirements of user applications. The autonomic aspect of our algorithm is related to the use of a knowledge plane, which is a conceptual authority that entertains detailed runtime knowledge about the network state in terms of the load on each access point, as well as the connection speed and application QoS requirements of each wireless device. We use OPNET to evaluate the effectiveness of our management solution and load balancing algorithm. The paper presents the conducted simulations and ends by a critique of the general applicability of our algorithm and suggestions for future work.

Autonomic Management for Capacity Improvement in Wireless Networks

This paper presents an Autonomic Load Balancing Algorithm (ALBA) which enhances resource allocation in Wireless Local Area Networks (WLAN). ALBA offers autonomic management capabilities to access points that ensure the accommodation of more user traffic and an overall network capacity improvement compared to default signal strength based connectivity. Simulations with VoIP traffic over an agglomeration of WLANs show consistent behaviour among neighbouring access points and substantial performance gain in terms of call blocking rate.

XAV: a fast and flexible tracing framework for network simulation

With current network simulators, simulation results are obtained from flat trace files via sequential data access. To speed up simulation output analysis, we propose to store the simulation traces into data warehouse. Our proposed framework relies on a three-layer architecture, with the intermediary database layer serving the trace file post processing services. The trace files are accessed through SQL-like queries which greatly enhances the flexibility of the post processing tools. Furthermore, the database paradigm allows to separate the packets data from the nodes send and receive operations and thus to remove redundant information between simulation events. The global system relies on a new trace file format based on the popular XML language. XML pointers are used to create relationships between data so that post processing tools can navigate into simulation outputs. The tests are carried out under the MonetDB database. The obtained results show considerable response time improvement compared to traditional sequential access mode.

YAVISTA: A Graphical Tool for Comparing 802.11 Simulators

The difficulty to compare network simulators is a major obstacle to the broader adoption of these tools. This problem particularly affects 802.11 simulators because the 802.11 protocol is complex and simulators contain several simplifications that are difficult to evaluate. In this paper, we present a generic visualization system able to interface with the 802.11 protocol implemented in NS-2 and Glomosim. Our tool allows to emphasize the common behaviours and divergences observed between the simulators and the IEEE 802.11 standard. Our contribution includes the definition of a new method to compare simulators and the publication of the most complete list to date of modeling differences relative to NS-2 and Glomosim lower layers.

Proactive Traffic Engineering for IEEE 802.11 Mobile Wireless Networks

As a result of its success, Wi-Fi has been promoted as an access technology to provide full internet connectivity in the municipal area. Supporting application QoS by managing resource provisioning is an important challenge for wireless network operators. This paper proposes a traffic engineering model for managing a Voice over Wi-Fi service in a municipal area with stationary and mobile clients. Our proactive algorithm guarantees the connectivity of the mobile clients while maximiz- ing the number of stationary stations. Simulation results show substantial QoS improvements for mobile clients and a near optimal utilization of the capacity of the wireless access network next to the congested hotspots. Index Terms—IEEE 802.11, VoIP, Traffic Engineering.

XAV: a tracing framework for exploring large network simulation outputs

This paper presents our ongoing works towards visual exploration of large network simulation outputs. Visual exploration allows users to search through simulation traces by using multidimensional representations of the network in an intuitive and interactive way. To speed up trace file exploration, we propose to store the simulation events according a format, namely XAV, that i) exploits the multidimensional nature of users requests and ii) allows quick identification of the packet paths through the network. XAV files are stored in a database to simplify data manipulation. The XAV tracing framework has been implemented in the NS-2 simulator and tested over a wireless ad-hoc network composed of 25 nodes. The performance evaluation shows that XAV enables to extract about 1000 packet paths per second from 100 MB trace files.

Pilotage autonomique de la charge dans les réseaux IEEE 802.11

With the proliferation of the IEEE 802.11 technology as a universal standard for wireless LAN, the management of access points and wireless stations has become essential to provide QoS to the users. In this article, we propose an autonomic piloting system for load management in large 802.11 networks. Our system supports both a reactive and a proactive load distribution algorithm. Load distribution relies on the capacity of the system to locally reorganize the network by dynamically re-allocating stations to the different access points. The autonomic capabilities allow the access points to update their knowledge about the traffic state in their neighborhood. We evaluate our system with VoIP traffic. The evaluation reveals significant gains in terms of overall network capacity.

A flexible channel access model for wireless network interface cards

Wireless network interface cards (WNIC) are currently moving from hardwired devices to autonomic software. This paper proposes a new rationale for the design of WNIC that i) takes into account the need to integrate an ever increasing number of network features, ii) supports the traceability of requirements in system development and iii) allows flexibility at the software architecture level. Our solution relies on the FODA (feature oriented domain analysis) and FORM (feature oriented reuse method) engineering methods to bridge the gap between the systemspsila specification and implementation. As an example, we show how a model of the 802.11 standard based on this method can reproduce the features of a particular channel access protocol or be extended to the 802.11g standard.