A Demand Analysis for Peer-to-Peer Computational Resource SharingBachelor Thesis

5 April 2022, by Johanna Reichert

Abstract:
Simulations require a large amount of computing power. One approach to meet the additional demands is distributed computing. Our vision is computational resource sharing to take place peer to peer, with end-user devices running the computations. This thesis investigates the interest in such a system and derives requirements for it. For this purpose, a mixed methods approach with interviews as qualitative and a survey as quantitative data collection method is used. The participants of the studies are scientists using discrete event simulators. Up to now, the idea of the peer-to-peer system is limited to the offloading of simulations running with the network simulator OMNeT++. The investigation shows that there is a need for additional computing power when simulating with OMNeT++, but also when using other simulators. In addition, the need for a social overlay becomes apparent. Namely, this should give users the freedom to self-regulate how much of their resources they want to share and with whom.

Supervised by:
Prof. Dr. Janick Edinger, Prof. Dr. Eva Bittner