Gilles Vanwormhoudt

Introducing variability into aspect-oriented modeling approaches

Aspect-Oriented Modeling (AOM) approaches propose to model reusable aspects, or cross-cutting concerns, that can be composed in different systems at a model or code level. Building complex systems with reusable aspects helps managing software complexity. But in general, reusability of an aspect is limited to a particular context. On the one hand, if the target model does not match the template point-to-point, the aspect cannot be applied. On the other hand, even when it is actually applied, it is woven into the target model always in the same way. In this paper, we point out the needs of variability in the AOM approaches and introduce seamless variability mechanisms in an existing AOM approach to improve reusability. Our aspects can fit various contexts and can be composed into the base model in different ways. Introducing variability into AOM approaches will turn standard aspects into highly reusable aspects.

GPM: A generic and scalable P2P model that optimizes tree depth for multicast communications

Group communications (real-time and non-real-time) refer to one-to-many or many-to-many communications. On the one hand, multicast is considered as an appropriate solution for supporting group communication-oriented applications (we distinguish IP network multicast from application layer multicast). On the other hand, peer-to-peer model tends to be a good candidate for supporting today Internet applications (e.g. P2P IPTV, P2P VoIP, etc.). In this context, P2P has attracted significant interest in the recent years. This is mainly due to its properties that also make P2P well adapted to today social networks. In this paper, we propose GPM (Generic P2P Multicast): a novel generic and scalable approach, that optimizes multicast tree depth in P2P networks (structured and unstructured), and contributes to control the network overlay latency. For multicast tree construction, the approach we propose is based on a distributed algorithm using a specific data structures (adjacency and forwarding matrixes). GPM model inherits from P2P attributes such as scalability, flexibility and fault tolerance, while taking into consideration the respective characteristics of one-to-many and many-to-many type of applications. We also give a performance evaluation for validation and comparison purposes while considering some main existing application layer multicast protocols. Copyright

MOSQoS: Subjective VoIP Quality for Feedback Control and Dynamic QoS Adaptation.

Supporting QoS for VoIP is a topic that clearly remain of importance. Existing QoS mechanisms for voice over IP need to be adapted. This is essentially due to the high dynamicity and variability of subjective voice quality perception, particularly in some critical conditions (ex. mobile networks). In this paper, we present an architecture, called MOSQoS, that integrates subjective voice quality assessment into a dynamic QoS loop control. This QoS loop control is based on a subjective and dynamic voice quality evaluation (MOS score). It maintains as constant as possible, a predefined target MOS score during the entire voice session. To achieve this approach, we implement PE-Model for subjective voice quality evaluation and combine it with a signaling protocol to dynamically control and optimize network resources such as: queuing allocation and congestion thresholds. Tests and measures obtained through an implementation of the MOSQoS approach, show that results in terms of feasibility and performance, are globally satisfactory.

SIP for e-learning services

Since 1993, we have been involved in the development and deployment of network architecture and protocols for distance learning. The first platform was based on ISDN and H.320 protocols. Now, we implement, test and validate a new Internet based architecture. We present an Internet based platform for e-learning and discuss some of issues concerning accessibility and mobility. To deal with these issues, we propose to integrate SIP into this platform. This integration provides easy support for mobility, service customization, access simplicity, etc., as well as a rapid development and deployment of new services. After a description of the different connection scenarios using SIP, we focus on testing and measuring performances, mainly in terms of service establishment delays. Analysis of results shows that, for low bit rate link access (modem) and wireless (802.11b) access, performance needs to be improved.

A P2P Framework for Decentralized Xconferencing and Its JXTA Implementation

IETF Xcon working group and ITU-T SG 16 specify a generic framework for centralized multipoint conferencing applications. The centralized architecture adopted by these specifications are not well adapted to today open and dynamic networks and have well-know drawbacks. In this paper, we present a P2P approach for supporting conference type services in a decentralized architecture. This approach provides new signaling mechanisms and components for coordinating, managing and controlling a conference system, in a dynamic network of peers. The implementation based on a JXTA middleware, shows that P2P model through its self-organizing characteristic can improve conferencing applications performance by providing a maximal autonomy for a minimal service configuration.

Toward feasibility and scalability of session initiation and dynamic qos provisioning in policy-enabled networks

In this paper we implement and evaluate a new solution for the multimedia session setup with resource allocation in policy-enabled networks that we developed in [1]. Our proposal is based on the use of Session Initiation Protocol (SIP) in the framework of Policy-Based Management (PBM). We specifically evaluate the feasibility and the scalability of such solution in a real environment through experimentation on a test-bed. The latter integrates all the components from session initiation to QoS policy enforcement into network elements. Our results demonstrate both of the relevance and the efficiency of such solution.

A Multiview Framework Driven by Use Cases to Support the Design of Service Components

In the field of software technologies, the adoption of Service Oriented Architecture for IS has been accompanied by the introduction of new service component models which are well-suited to realize services. While these service components offer new capabilities to simplify the construction and maintenance of IS, the issue of their design needs to be further examined as current existing works do not sufficiently bridge the gap between requirements and service component models. In this paper, we present a multi-view framework to support the design of services along the development process. The framework is grounded on the notion of Logical Service Block which represents a service from Use case to assembly model and acts as a pivot between views ensuring the consistency and the trace ability of services at different stages and levels of detail. The proposed framework has been implemented to provide a UML2 tool that helps designers in elaborating views and projecting services towards a specific platform like SCA.

Smart Profile: A New Method for Minimising SIP Messages

SIP (Session Initiation Protocol) is a text-based protocol engineered for high data rate links. As a result, SIP messages size have not been optimised. This is problematic, particularly with the planned usage of this protocol in wireless handsets as part of 2.5G and 3G cellular networks. With low bit rate IP connectivity or signalling channel, the transmission delays for call setup and feature invocation are significant. In this paper, we propose a new approach called ”Smart Profile” where SIP messages are minimised in the access network for performance optimisation and completely reconstructed at the edge. The message reconstruction is based on SIP call profile that contains session attributes extracted from the user and terminal profiles. We discuss about Smart Profile implementation and focus on measuring its performance. Results and comparison with compression method show that Smart Profile solution is globally better.