Am Dienstag, 27. März 2018, 10:00 Uhr findet im HS T02 ein Gastvortrag statt.
Titel: " From Demand-Aware Networks (DANs) to Self-Adjusting Networks (SANs)“
Vortragender: Prof. Dr. Stefan Schmid
Host: Univ.-Prof. Dr. Robert Elsässer

Traditional communication networks are optimized towards static and demand-oblivious properties such as node degree and network diameter. While such network topologies may be well-suited to serve all-to-all communication patterns, actual communication patterns are often far from uniform but feature much
structure and are often sparse. 
In the first part of this talk, I will explore the design of Demand-Aware Networks (DANs) whose topologies are optimized toward a given traffic pattern.
In particular, I will leverage an interesting connection to coding theory and present an optimal constant-degree network design for sparse communication patterns: the route length in these DANs is proportional to the conditional entropy of the demand matrix, which is asymptotically optimal.
In the second part of the talk, I will discuss challenges involved in designing Self-Adjusting Networks (SANs) which even adapt to changing demands over time. I will show that SANs can be seen as a distributed generalization of self-adjusting datastructures, and will present the design of a particularly interesting type of SAN accordingly: the SplayNet.

Zur Person Stefan Schmid:
Stefan Schmid promovierte 2008 an der ETH Zürich. Über Stationen in Deutschland (TU München und Universität Paderborn 2008-2009 Postdoc; als Senior Research Scientist an den Telekom Innovation Laboratories Berlin 2009-2015), Frankreich (Visiting Professor am CNRS (LAAS), Toulouse 2014), Belgien (Visiting Professor Université catholique de Louvain) und Dänemark (Associate Professor Universität Aalborg, Dänemark 2015-2018) wurde er im Februar 2018 an die Universität Wien berufen (Professur für Communication Technologies).
Seine Forschungsschwerpunkte sind: Kommunikationstechnologien, Robuste verteilte und vernetzte Systeme, Software-Defined Networks und Netzwerkvirtualisierung sowie Netzwerkalgorithmen.