The Border Gateway Protocol (BGP) is the most commonly used instrument to exchange reachability information between networking nodes in the Internet. Currently BPG operates by propagating only one best path towards neighbouring routers in order to keep the Internet scalable. However this single path propagation comes at the cost of reducing the number of alternative paths available to the autonomous systems (AS). Research attaches great importance to being able to advance BGP to propagate multiple paths towards neighbouring routers.

As there is currently no working implementation of BGP propagating multiple best paths in existence, we summarise requirements and propose a selection of metrics to enable BGP to propagate multiple paths. We develop a set of experiments to verify and validate implementations of multiple path propagation within BGP via testing the protocol against the following factors: topology; number of paths to propagate; different decision processes; and the impact of policies. Last, but not least, we analyse the existing diversity of paths in networking nodes by examining the number of paths; the disjointness of paths; the impact of a possible multiple path propagation; and the influence of commonly used policies on the diversity of paths.

For a high-resolution analysis of simulation models it is often necessary to execute simulation experiments consisting of a large amount of simulation runs. Such simulation experiments may consume substantial computational time, even when a simple model is used, and therefore can often only be successfully computed using a super-computer or by distributing the computation of the simulation experiment on a network of PCs, often referred to as a cluster or a grid.

As part of my thesis, a web based system has been developed to interact with a simulation system, enabling users to start simulations and retrieve simulation results securely through the internet. Users with administrative permission are able to see all simulation activity and to perform administrative tasks.

Furthermore an analysis method will be conceptualised to detect "interesting" regions within the model parameter space, with the goal of increasing the experiment resolution in these regions. In this final presentation, I will describe and present the concepts developed for the application to manage and control simulation experiments and explain the method which was conceptualised to detect interesting regions within the model parameter space.

One major challenge in communication networks is the problem of dynamically distributing load in the presence of bursty and hard to predict changes in traffic demands. Current traffic engineering operates on time scales of several hours which is too slow to react to phenomena like flash crowds or BGP reroutes.

One possible solution is to use load sensitive routing. Yet, interacting routing decisions at short time scales can lead to oscillations, which has prevented load sensitive routing from being deployed since the early experiences in Arpanet.

However, recent theoretical results have devised a game theoretical re-routing policy that provably avoids such oscillation and in addition can be shown to converge quickly. In this paper we present ReplEx, a distributed dynamic traffic engineering algorithm based on this policy. Exploiting the fact that most underlying routing protocols support multiple equal-cost routes to a destination, it dynamically changes the proportion of traffic that is routed along each path. These proportions are carefully adapted utilising information from periodic measurements and, optionally, information exchanged between the routers about the traffic condition along the path.

We evaluate the algorithm via simulations employing traffic loads that mimic actual Web traffic, i.e., bursty TCP traffic, and whose characteristics are consistent with self-similarity. The simulations quickly converge and do not exhibit significant oscillations on both artificial as well as real topologies, as can be expected from the theoretical results.

Der Vortrag behandelt die Analyse und Bewertung von BGP-Daten, die der Lehrstuhl von den Border-Routern des AS SpaceNet bekommt.

Im theoretischen Teil der Arbeit soll geklärt werden, welche Informationen man überhaupt aus BGP-Daten beziehen kann und ob damit Topologie- oder Konfigurationsänderungen nachvollzogen und Fehler bzw. Fehlkonfigurationen erkannt und deren Ursache eingegrenzt werden kann. Zusätzlich zu den Daten der Border-Router, die eine Sicht auf das AS von "Innen" ermöglichen, liefern die Looking Glas Server von z.B. RIPE eine Sicht auf das AS von "Außen", und aus dem Vergleich dieser Sichtweisen lassen sich weitere Erkenntnisse über die Vorgänge im AS beziehen.

Im praktischen Teil soll ein kleines System erstellt werden, welches den Netzadministrator über wichtige, ausgewählte Vorgänge und Veränderungen sofort informiert. Unkritische Ereignisse sollen für den Admin in einer Tageszusammenfassung dokumentiert werden. Ziel ist hierbei die Menge an gesammelten Daten nach geeigneten Kriterien zu filtern und in einer für diesen Zweck sinnvollen Form dazustellen.

AS-path prepending is a common method to influence BGP routing decision. By prepending the local ASN multiple times, ASes can make advertised AS paths appear artificially longer.

Analyzing data from BGP table dumps, we try to shed light on the importance of AS-path prepending in the Internet. Simple experiments are supposed to study how observed prependings can be integrated into interdomain-routing models of the Internet.

For a high-resolution analysis of simulation models it is often necessary to execute simulation experiments consisting of a large amount of simulation runs. Such simulation experiments may consume substantial computational time, even when a simple model is used, and therefore can often only be successfully computed using a super-computer or by distributing the computation of the simulation experiment on a network of PCs, often referred to as a cluster or a grid.

As part of my thesis, a web based system has been developed to interact with a simulation system, enabling users to start simulations and retrieve simulation results securely through the internet. Users with administrative permission are able to see all simulation activity and to perform administrative tasks.

Furthermore an analysis method will be conceptualised to detect "interesting" regions within high-dimensional model response, with the goal of increasing the experiment resolution in these regions.

In this presentation, the infrastructure upon which the web based system is built will be presented. Additionally, an introduction to a method of simulation experimentation, called data farming, as well as an introduction to the analysis method will be given.

Network packets take many ways through the Kernel. Let us take the TCP/IP stack for example. Furthermore (under Linux) there is the PF_PACKET (a raw socket) used to capture packets and the netfilter framework in general used to perform routing and/or packet filtering. Our research group is interested in capturing packet, which are possibly subject to filtering as well, as soon as possible. This work deals with the question, whether the netfilter framework is suitable for this task.

While the first talk was about the way of a packet throught the kernel, and the design of the netfilter framework, this talk will focus on the developement and implementation of the kernel sniffer. Furthermore the problems and necessities of moving a user space application into the kernel are discussed, and the implemented solutions are explained. In doing so, the kernel sniffer was optimised, errors were corrected until the performance reached a comparable level to tcpdump.