Sebastian Schellenberg

Implementation and validation of an address resolution mechanism using adaptive routing

In disaster scenarios, communication systems usually consist of heterogeneous nodes and damaged infrastructure. Communication is important for rescue teams and victims as well but a serious problem because normal network systems like wired or mobile radio Internet could be unreliable or simply not available. To deal with these problems, much effort has been spent to mobile ad hoc networks (MANETs) and their specialties. Those networks are usually unsteady and highly mobile. Therefore, classical network services like resolution of names to their regarding local IP addresses is a big challenge. To overcome the problem of name resolution, we introduced a new approach for consistent and efficient name resolution using adaptive routing techniques. In this paper, we present the implementation for a Linux-based platform using the Click Modular Router and the validation of our testbed using the Network Simulator 3.

Movement patterns for mobile networks in disaster scenarios

Node mobility is one essential feature of mobile networks, especially for delay tolerant networks, where packet delivery is only possible through node movements. In mobile ad hoc networks, mobility adds new challenges to the protocol design. Therefore, realistic movement models are required to evaluate protocol performance. Usually, random node movement is considered for that but this kind of movement is not applicable for first responders in disaster scenarios. Their movements are driven by tactical mission requirements and highly depend on the type of disaster. In this paper, we analyze two different movement scenarios to derive common features as well as differences and present how to use these results to generate generic and at the same time realistic movements for first responders.

Impact of Traffic and Mobility Patterns on Network Performance in Disaster Scenarios

First responders show specific movement and traffic patterns that are closely related to their current task during a disaster relief mission. In order to develop efficient communication systems both aspects should be considered. This is important especially if Delay Tolerant Networks (DTN) are used, because they are exploiting the mobility of the nodes in order to deliver messages in scenarios with intermittent connectivity. In a disaster scenario, such a network is formed by the communication equipment of first responders. In this paper, we evaluate different mobility models and traffic generation options and show their impact on the performance of the epidemic routing protocol. Our results show that the correlation between traffic and mobility in disaster scenarios has a significant impact on the delivery ratio and experienced delay, respectively

A Realistic Underlay Concept for Delay Tolerant Networks in Disaster Scenarios

During disaster relief missions, it is essential to provide reliable communication between the different rescue teams and to the victims as well. Since a disaster is likely toe Mobile Ad hoc Networks (MANETs) are the best choice to reestablish communication. However, this type of networks has a major drawback, it can only cover relatively small areas. In large disaster sites, this results in creating frequent partitions of the network as rescue teams spread over the area. In order to handle these disruptions, delay tolerant networks could be used by employing mobile nodes as data ferries between different partitions. These ferries provide the same functionality as border nodes interconnecting different MANETs. In this paper, we present our concept of delay tolerant border nodes and a realistic mechanism to resolve names in such disaster scenarios. This approach enables robust communication throughout the disaster area while supporting heterogeneous technologies and different organizational policies.

A Computationally Inexpensive Battery Model for the Microscopic Simulation of Electric Vehicles.

The transition from classic combustion engine vehicles to electric vehicles is a major step to reduce worldwide CO2 emissions. In order to correctly and efficiently investigate impacts on the electric grid and the dimensioning of charging infrastructure, or to explore new technologies for a further increase of the driving range, realistic simulation models are required. In this paper, we present an accurate yet computationally inexpensive battery and kinematic model to be used in microscopic traffic simulation to help study the performance of thousands of electric vehicles. Our model also supports recuperation and range extender modules while only relying on the vehicles speed and a set of constant predefined parameters. It can therefore be easily coupled with current sophisticated traffic simulators. We show its applicability and correctness regarding the State of Charge and power flows using comprehensive real-life experiments.

Hybrid spectrum sharing in dynamic spectrum access networks

In this work, inefficient spectrum usage problem which will lead to spectrum scarcity in future communications has been addressed. According to Federal Communications Commission, spectrum sharing is a technique to efficiently utilize the spectrum. We analyse dynamic spectrum access concepts such as overlay spectrum sharing and underlay spectrum sharing. Our overlay and underlay analysis have been made with the use of five and eight state continuous time Markov chains respectively. We derive the steady state probability for these states and also calculate the throughput of the aforementioned spectrum sharing schemes with the use of Shannons channel capacity. In addition to that, specifically, we propose a hybrid spectrum access scheme which combines overlay and underlay spectrum sharing schemes and employ them to improve the system throughput by efficiently using the spectrum. Our approach, by using continuous time Markov chain, reduces the complexity in modelling the dynamics of primary and secondary user. We have also shown numerical results to validate the performance of the proposed scheme. Our scheme lays foundation for further works in the field, and our scheme could be enhanced in future works.

Secure and QoS-aware communications for smart home services

In this paper, we present the combination of two concepts addressing important aspects for smart home services: a new quality of service concept and a new integrated concept for a smart home security system. We assume that every flat in a house has its own smart device to perform local services and act as an interface for the different sensors and actors inside the flat. These smart computers are interconnected using a plain switched Ethernet infrastructure. All different types of sensors and actors communicate via different types of standards. Based on the classification of each service, we present a solution to prevent congestion situations inside the network. This scheme is not only applicable to the presented scenario, it could also be applied to handle similar problems within classical computer networks and many other fields. These advantages are realized within the affirmation of security of the whole system. All communication channels, all information and user-data and all kinds of access to the system are secured by a comprehensive security architecture.

Routing-based and location-aware service discovery in Mobile Ad-hoc Networks

In disaster scenarios, a stable and reliable communication is essential for coordinating rescue operations. As the communication infrastructure could be hit, e.g. by earthquakes or hurricanes, mobile ad hoc networks (MANETs) can be used for connection establishment. As such networks should be decentralized to increase robustness, some usual services are no longer available. In our previous publications, we presented our name resolution framework based on an adaptive routing system. In this paper, we show a service discovery mechanism based on this system.We show, how this mechanism can easily be improved with the use of location information, which is an important issue for service availability.

IP-based access to sensor networks enabled by a transparent proxy server

The protocol of a sensor network is optimized for its application field, but to benefit from these networks it is necessary to make sensor data available for other networks, too. Therefore, this paper deals with the realization of an interconnection between wireless sensor networks and the Internet. For the first realization of the system, a sensor network based on the IEEE 802.15.4 standard and ZigBee was used. The interface for internetworking is realized as a new gateway-based approach using a transparent proxy server offering a web-based service for the user to query or to change the status of a special sensor network node. The concept of the developed system is independent of the sensor network system and can be used for many network types to establish an interface to the Internet and additionally to provide intercommunication between these technologies.

Buffer Management for Reliable Multicast over Delay Tolerant Networks

Flooding, earthquakes, and other disasters are recent phenomenons which unfortunately occur more and more often. After such occurrences, the infrastructure is likely damaged, but rescue teams need to communicate with each other to plan the rescue missions. Mobile Ad hoc Networks (MANETs) are an efficient and cost-effective opportunity to establish communication, if no working infrastructure is available. However, they come with several challenges, e.g. Interruptions due to partitioned MANETs. Delay Tolerant Networks (DTNs) are a possible solution to overcome this problem. Furthermore, rescue missions require reliable communication between different rescue teams and among their team members. In order to efficiently reach all members of a team, multicast is promising. However, guaranteeing that all recipients receive a DTN multicast message is difficult and might consume too much buffer space in the DTN nodes. In this paper, we present our approach of a DTN-based multicast protocol that provides efficient buffer management without neglecting the required reliability. We explain our algorithm to efficiently free the node buffers. Our simulation results will show that our algorithm increases the delivery ratio for different group sizes.