Abstract not found

The transport mechanisms of lnternet have an inherent drawback, which is due to their generality. Internet has to support a broad range of services (data, audio, video, etc.) on top of a unique protocol stack. On the other hand, it is well known, that there is a high potential .for QoS and if/ciency improvement if.joint source and channel coding are used. Utilization of this potential is especially attractive in the case of interactive multimedia transmitted over wireless channels. In this paper we present the Mobile Adaptation System (MAS), which supports the dynamic deployment of boosters .for .flow selective link level packet treatment in wireless Internet access. These boosters use the knowledge of the importance of individual packets, or relations among packets in order to optimise the link layer transmission. FEe discuss our approach using as example a speech property based (SPB) booster for improvement of the perceptual quality of Voice over IP transmission in FEireless LANs.

Abstract. A key early objective of Active Networking (AN) was to support on-the-fly network evolution. Although AN has been used relatively extensively to build application-customized protocols and even whole networking systems, demonstrations of evolution have been limited. This paper examines three AN mechanisms and how they enable evolution: active packets and plug-in extensions, well-known to the AN community, and update extensions, which are novel to AN. We devote our presentation to a series of demonstrations of how each type of evolution can be applied to the problem of adding support for mobility to a network. This represents the most large-scale demonstration of AN evolution to date. These demonstrations show what previous AN research has not: that AN technology can, in fact, support very significant changes to the network, even while the network is operational. 1

Technology trends in recent years have resulted in the rapid and wide-scale deployment of embedded systems as critical components of larger systems such as home appliances, airplanes, ships, and motor-vehicles. Furthermore, traditional desktops and workstations themselves are becoming a federation of embedded systems such as external storage devices connected with USB and FireWire buses. Such ubiquity of embedded systems increases the impact of random and maliciously-induced failures and the need for unforeseen improvements.

Current Internet protocols and network services have been struggling to keep up with the fast evolution from traditional data to today’s multimedia communication technologies and the changing requirements following these advances (for example, support for QoS, multicast, mobility, and security is still lacking in most networks). Active and programmable networking is a step towards enhancing the static and inflexible structures of current networks. As part of the general research trend, this thesis focuses on the design and development of technologies that allow rapid deployment of new functionality throughout the network (for example, customised network services and protocols), which allows network vendors and service providers to respond quickly to the changing requirements and keep up with the fast evolution in network and communication technologies. This thesis starts off with an general introduction into the research area and a description of the basic mechanisms behind active networking. A critical examination of existing active and programmable systems and associated technologies is provided. These results together with previous experiences of developing an active network system

Abstract not found