SATNAC 2009 Conference Papers
Authors: Philip Foulkes (
Abstract: In the IT industry there is a trend towards the convergence of the transportation of different types of content to single networking technologies. Web, e-mail and VoIP applications all use a common network to transport their data. This convergence leads to lower equipment costs and allows for more future-proof networks to be built. Audio and video production environments tend to use different technologies for the transportation of different types of data. Different types of equipment use varying command and control protocols. There are strict bandwidth, latency and timing requirements that need to be met for transporting real-time audio and video data. IEEE 1394 is a standards-based networking technology that meets these requirements. Ethernet AVB is an up and coming standards-based networking technology that is designed to meet these requirements. This paper discusses the initial phase of an investigation into the interoperability of, and control over, IEEE 1394 and Ethernet AVB networks. A tunnelling application has been developed that allows Ethernet devices to take advantage of the capabilities of IEEE 1394. Control over the parameters of the tunnelling software is performed by an IP based command and control protocol, AESX170.
Authors: Johannes Göbel (
Abstract: Consider a mobile ad hoc network (MANET) where packet transmissions occur between mobile nodes. Such a network requires that the nodes act as relays to form multi-hop routes connecting the origin-destination node pairs that are out of radio transmission range with respect to each other. If the nodes do not to belong the same authority, then the nodes must be given an incentive to spend their resources (battery power and transmission bandwidth) in forwarding packets that originate at other nodes. This can be done by introducing a credit balance for each node, where the nodes use credits to pay for the costs of sending their own traffic, and earn credits by forwarding traffic from other nodes. This paper presents several variants of a credit-based incentive scheme (protocol) to promote collaboration in MANETs. The variants address the fairness of the scheme, namely (1) to assist those nodes that are under-provided with credits: such nodes may lack the credits necessary to transmit their data, and (2) to protect those nodes that are over-provided with credits: such nodes may, to their own detriment, provide too large a proportion of their resources to the community. We first present two basic incentive schemes: the first scheme does not regulate the willingness of the nodes to forward packets on behalf of other nodes; the second scheme contains such a regulatory mechanism. We next focus on protocols whose design lies between the two basic schemes. Both the origin pays and the receiver pays protocols are investigated. Both constant and congestion-related resource prices are investigated. Initial simulation experiments indicate that both the origin pays and the receiver pays protocol with a credit distribution and discounting (redistribution) mechanism is sufficient to afford a reasonably fair allocation of flows in the network.
Authors: Mthulisi Velempini (
Abstract: Multi-channel MAC systems offer high capacity, a prerequisite for the next generation networks. High capacity will ensure that the requirements of delay sensitive and time bounded flows are met. For high capacity to be a reality, a good channel selection, coordination and network connectivity is required. A common control channel approach addresses the connectivity challenge. However, it is perceived to be a bottleneck as it saturates degrading data channels. The saturation problem is not widely studied. This paper investigates saturation levels of both control and data channels. The length and channel occupancy of control and data channels are explored. The analytical results show that data channels saturates ahead of the control channel. The model is premised on short control packets requiring shorter transmission durations; meaning that a control channel is not capacity constrained as compared to data channels.
Authors: Suné von Solms (North West University), Albert Helberg (North West University)
Abstract: We introduce a novel error correction scheme that uses the implicit encoding capability of Random Network Coding. This scheme does not add redundancy to the data prior to transmission, like existing error correcting schemes. Random Network Coding within a large network generates enough redundant information to perform error correction on transmitted data.