Bernd Lamparter

Charging support for ad hoc stub networks

Wired equivalent security is difficult to provide in ad hoc networks due to high dynamics, wireless link vulnerability, and the requirement of complete decentralization. Especially mutual authentication among nodes in a self-organized mode of ad hoc network, where a single definition of trust is hard to build, is very difficult. This leads to several loopholes in previously established security architectures and intrusion detection systems if employed to ad hoc networks. Also, typical to ad hoc routing major security lacks are unfair participation of nodes in the system, and frequent attacks to the routing protocols by ad hoc nodes that do not want to forward foreign packets but use other nodes to forward their own packets. Contrary to detection-based approaches to tackle this situation, we propose a motivation-based approach, which does not require mutual authentication. We enable this by providing (a) a realistic architecture of ad hoc access network to an Internet Service Provider, (b) a workable business model for charging in this architecture and (c) a necessary security protocol to implement the charging scheme. For the latter we give a formal validation.

Lightweight QoS-support for networked mobile gaming

In this paper, we present an approach to provide Quality of Service (QoS) for networked mobile gaming. In order to examine the QoS requirements of mobile games, we ported a simple real-time game called GAV (GPL Arcade Volleyball) to a PDA and performed several traffic measurements over both GPRS and UMTS networks. We show that due to high end-to-end delay and delay jitter, real-time games are not supported by GPRS. While UMTS improves both delay and jitter, it still does not match the requirements of real-time games. The key reason for this problem is that overprovisioning, as it is used to allow real-time games in the Internet, is very expensive in mobile networks. At the same time, QoS classes for mobile networks are not tailored to real-time games. In order to reduce delay and jitter for this application class, while still accounting for the very bursty nature of real-time game flows, we propose to use a combination of statistical multiplexing and QoS guarantees. The general idea is to aggregate multiple game flows and perform reservation for that aggregate. As a theoretical background, we use a queuing system based model. Through simulation of a sample network with the traffic data generated by GAV, we validate our assumptions and demonstrate the performance and characteristics of our approach.

On digital signatures in ad hoc networks

Cryptographic primitives need to be carefully evaluated when being applied in components ensuring enhanced protection aims. This work analyses the performance of RSA- and ECDSA-based digital signature schemes in the context of multi-hop ad hoc networks. Our work shows that contrary to a single wireless hop scenario with restricted clients in the wireless network and a powerful server in the fixed network, the choice of an appropriate signature candidate is not as obvious as for the aforementioned architecture. Our performance analysis considers three partially interdependent axes (a) the required security level, (b) the device restriction and (c) the protocol specification with the required security relationships, and the assumed traffic flow. We take (c) into account by analysing two charging protocols for wireless multi-hop scenarios. By generalising our observations, we carefully derive some recommendations to strengthen the security engineering process for dependable distributed systems over the wireless network. Copyright

X-Movie: Transmission and Presentation of Digital Movies under X

We describe a system for storing, transmitting and presenting digital films in a computer network. The hardware used in the system is standard hardware, as found in typical workstations today; no special hardware is required. The movies are shown in a window of the X Window System. This allows full integration with the classical components of computer applications such as text, color graphics, menues and icons. The X-MOVIE system is based on color lookup table technology. We present a new algorithm for the gradual adaptation of the color lookup table during the presentation of the film.

XMovie: architecture and implementation of a distributed movie system

We describe a system for storing, transmitting, and presenting digital movies in a computer network. The hardware used in the system is standard hardware, as found in typical workstations today; no special hardware is required, but if available it can be used to provide better performance. The XMovie system has several innovative features. First, it contains a new algorithm for the gradual adaptation of the color lookup table during the presentation of the movie to ensure optimal color quality on low-end workstations. Second, it is a multistandard system supporting the compression techniques MPEG, Motion JPEG, and a newly developed extension to the well-known Color Cell Compression method. Third, it contains AdFEC, a new adaptable forward error correction method for our movie transmission protocol.

About the impact of co-operation approaches for Ad hoc networks

Although the need for protocol extensions dealing with cooperation issues for ad hoc routing protocols is undoubtedly accepted, it is still a question of belief whether such protocol extensions really intensify participation within the ad hoc network and benefit the community. This statement is applicable to both motivation-based (e.g. [1], [2]) and detection-based approaches (e.g. [3], [4]). It is our intention to give a quantitative estimation of the added value of such co-operation based approaches. We present results based on two assumptions: First, the co-operation approach uses an underlying on-demand routing protocol (DSR, AODV, etc.) and in particular tries to combat malicious behavior of nodes in the forwarding phase. More precisely, although a node behaves well in the routing phase, it simply drops packets in the forwarding phase. Second, the network topology is low-densely filled with nodes, meaning that each node’s transmission range does not reach more than a few other nodes (approximately four). In such topologies, the loss of one node often cannot be handled by diverting traffic. This work particularly shows in which range any cooperation approach increases the probability of the destination to receive data addressed to it. Note, that for the measurement of the effect of a co-operation approach, it needs not to be shown how high participation is in the presence of it. It is instead required to evaluate the increase of throughput to the destination compared to solely communicating without such an approach. The statistical participation model we imply is both simple and meaningful. It is suitable to evaluate the increase in participation in the presence of any co-operation approach. It is fully characterized by the ad hoc nodes’ behavior: 1. For each ’transmission event’, each node uniquely decides to either forward all traffic or to drop all traffic, i.e. if a node once decides to send a bundle of packets it indeed sends all these packets. 2. the ratio of forwarding or dropping traffic is uniformly distributed over all nodes of the ad hoc community and thus it is equal for all. Note, that our participation model in particular includes the case where nodes may participate in the route discovery phase while dropping traffic in the forwarding phase. Our model also takes the nodes’ internal states into account. We feel that such states are mainly driven by a device’s remaining battery power but also by its owner’s random behavior. Depending on its internal state, a node may at some point in time decide to forward data, whereas it drops packets at other times. II. Effect of Participation Intensification

eXtended Color Cell Compression - A Runtime-effient Compression Scheme for Software Video

Multimedia applications require a compression and decompression scheme for digital video. The standardized and widely used techniques JPEG and MPEG provide very good compression ratios, but are computationally quite complex and demanding. We propose to use an extension to the much simpler Color Cell Compression scheme as an alternative. Our extension includes the use of variable block sizes, the reuse of color index values from previously encoded blocks, and Huffman encoding of the stream of blocks. We present experimental results showing that our scheme provides much better runtime performance than MPEG, at the cost of a slightly inferior compression ratio. It is thus especially suited for software videos in high-speed networks.

Estimating the value of co-operation approaches for multi-hop ad hoc networks

Abstract In multi-hop ad hoc networks, each node may be involved in forwarding data acting as a router. Unfortunately, in civilian scenarios with usually an open community, it is not obvious why nodes should forward foreign data and waste their battery power, CPU cycles, etc. Some work has already been done in proposing approaches to intensify participation in ad hoc networks. However, it is more or less a question of belief whether these approaches really benefit the community. This work gives a rational estimation of the effect of approaches that intensify participation in ad hoc networks. We present essential results assuming in particular low-dense networks that use on-demand routing protocols by applying some fundamental statistics to the problem.

Application layer issues for digital movies in high-speed networks

Digital movie system require a new generation of communication protocols. They are very demanding in terms of bandwidth, delay jitter, end-to-end delay, and multicast support within the network. We concentrate on application layer issues for digital move transmission. After introducing an application layer architecture, we present a novel, adaptive forward error correction scheme for movies, combining high correction probability with low overhead. We also present an efficient compression scheme for software movie systems, with better runtime properties than standardized movie compression algorithms.

X-MOVIE: Digitale Filmübertragung und Darstellung im X-Window-System

Diese Arbeit beschreibt ein System zur Speicherung, Ubertragung und Darstellung von Bewegtbildsequenzen in einem digitalen Rechnernetz. Dabei kommen ausschlieslich Hardwarekomponenten zum Einsatz, die in typischen Netzen aus Arbeitsplatzrechnern und Servern vorhanden sind; spezielle Hardware ist nicht erforderlich. Die Darstellung der Filme erfolgt in Fenstern des X-Window-Systems, um eine weitgehende Integration der Bewegtbildsequenzen mit den klassischen Bestandteilen einer EDV-Anwendung, wie zum Beispiel Text, Farbgraphik, Menus und Symbole (Icons), zu erreichen. Der Schwerpunkt der Arbeit liegt bei den Darstellungsformaten und Ubertragungsprotokollen fur digitale Bewegtbildsequenzen.