Register now After registration you will be able to apply for this opportunity online.
This opportunity is not published. No applications will be accepted.
User-centric data traffic engineering with karma
You are excited to host your friends for a long awaited movie night, only to get disappointed by low resolution streaming and buffering interruptions. Wouldn’t it be nice if you could ask your internet service provider (ISP) to guarantee fast service on your special night? You could perhaps buy a temporary bundle of fast internet, but previous experience has shown that monetizing access to the internet data highways is deemed highly unfair. Motivated by the recent development of karma mechanisms, the aim of this thesis is to investigate a data traffic engineering (TE) scheme which can tailor to the time-varying demands of the end users in a fair and efficient manner.
Keywords: Data networks, traffic engineering, game theory, karma mechanisms, net neutrality
Data traffic engineering (TE) refers to the problem of finding a bandwidth allocation in a data network that optimally satisfies user demands without exceeding data link capacities. Traditionally, user demands are taken as exogenous, truthful inputs to the TE problem. But with data-intensive content (e.g., high quality video) becoming the mainstream, it is not clear how peak congestion periods, in which simultaneous high bandwidth demands cannot be satisfied by the network capacity, are currently handled. Regardless of whether all users experience the same bandwidth reductions, or some users experience more reductions than others, the current methods are agnostic to the true private needs of the users. Karma mechanisms are promising to fill this gap, since they can be used to elicit private preferences in a fair manner.
Data traffic engineering (TE) refers to the problem of finding a bandwidth allocation in a data network that optimally satisfies user demands without exceeding data link capacities. Traditionally, user demands are taken as exogenous, truthful inputs to the TE problem. But with data-intensive content (e.g., high quality video) becoming the mainstream, it is not clear how peak congestion periods, in which simultaneous high bandwidth demands cannot be satisfied by the network capacity, are currently handled. Regardless of whether all users experience the same bandwidth reductions, or some users experience more reductions than others, the current methods are agnostic to the true private needs of the users. Karma mechanisms are promising to fill this gap, since they can be used to elicit private preferences in a fair manner.
Your task is to a) review the state of the art in data traffic engineering (TE), b) understand the current karma mechanism, with focus on its application to the related problem of vehicular traffic management, c) develop a tractable formulation for a karma-based TE scheme, d) demonstrate the efficacy of your approach in simulation experiments using standard network benchmarks, and e) (optional) implement and showcase your solution on a real model network.
Your task is to a) review the state of the art in data traffic engineering (TE), b) understand the current karma mechanism, with focus on its application to the related problem of vehicular traffic management, c) develop a tractable formulation for a karma-based TE scheme, d) demonstrate the efficacy of your approach in simulation experiments using standard network benchmarks, and e) (optional) implement and showcase your solution on a real model network.
Please apply directly to the posting on SiROP. We look forward to your application.
Ezzat Elokda (elokdae@ethz.ch), Dr. Georgia Fragkouli (gfragkouli@ethz.ch), Dr. Romain Jacob (jacobr@ethz.ch)
Please apply directly to the posting on SiROP. We look forward to your application. Ezzat Elokda (elokdae@ethz.ch), Dr. Georgia Fragkouli (gfragkouli@ethz.ch), Dr. Romain Jacob (jacobr@ethz.ch)