Welcome to the new version of CaltechAUTHORS. Login is currently restricted to library staff. If you notice any issues, please email coda@library.caltech.edu
Published May 11, 2021 | Published
Book Section - Chapter Open

Diolkos: improving ethernet throughput through dynamic port selection

Abstract

In large networked systems, a sudden increase in traffic could slowdown the network significantly, impacting network quality for multiple users. We present Diolkos, a system that leverages smart switches to dynamically re-reroute data flows in response to drops in performance. In contrast to other techniques, our tool predicts the future throughput at each port in a switch if a data flow were to be sent through it, and updates which port should be taken to maximize throughput. We use several techniques to predict network switch performance on a software defined network (SDN) mimicking topologies commonly found in datacenters. Experimentally, we demonstrate the effectiveness of choosing a port to send flows through based on predicted performance. We found that using a distributed predictive technique achieves a 24% improvement over using a traditional heuristic technique.

Additional Information

© 2021 Copyright held by the owner/author(s). We thank our collaborators Wenji Wu, Soren Telfer and Michael Jensen for their help in reviewing this paper. We also thank NVIDIA Corporation for their donation of a TITAN Xp GPU used as part of the development of Diolkos. Part of this work was conducted at "iBanks", the AI GPU cluster at Caltech. We acknowledge NVIDIA, SuperMicro and the Kavli Foundation for their support of "iBanks". We acknowledge support from Caltech's Intelligent Quantum Networks and Technologies (INQNET) research program, AT&T's Palo Alto Foundry and funding support from the U.S. Department of Energy's (DOE) Office of Advanced Scientific Computing Research as part of "Integrated End-to-end Performance Prediction and Diagnosis." This work is partially supported by a DOE/HEP QuantISED program grant, QCCFP/Quantum Machine Learning and Quantum Computation Frameworks (QCCFP-QMLQCF) for HEP, Grant No. DE-SC0019219. This work is partially supported by the U.S. DOE, Office of Science, Office of High Energy Physics under Award No. DE-SC0011925 and DE-AC02-07CH11359.

Attached Files

Published - 3457388.3458659.pdf

Files

3457388.3458659.pdf
Files (977.1 kB)
Name Size Download all
md5:15c04119c463341f694b549c9752f779
977.1 kB Preview Download

Additional details

Created:
August 20, 2023
Modified:
October 23, 2023