@Large Research

The presentation recording of "FootPrinter: Quantifying Data Center Carbon Footprint" is now available at: youtube.com/watch?v=zso1UBCQ…

Kudos to Zebin, Krijn, Nick, and Animesh for winning the best paper award at #ICPE24 with their work on #performance characterization of Linux #storage #schedulers! Well done! @ICPEconf Preprint available at: atlarge-research.com/pdfs/20…

Do you want to evaluate the sustainability of your data center? On Sat. the 11th at #HotCloudPerf (@ICPEconf), Dante Niewenhuis will present Footprinter our tools for supporting data center designers and operators in this task. dl.acm.org/doi/10.1145/36295…

What programming abstractions are sacrificed for simplicity in modern schedulers? And what is the performance cost? On Thu the 9th, 11:00-12:30, @ICPEconf, Sacheendra will present our paper "The cost of simplicity". #icpe #scheduler Preprint: atlarge-research.com/pdfs/20…

Krijn Doekemeijer presented the work from StoNet research group on "ZWAL: Rethinking Write-ahead Logs for ZNS SSDs with Zone appends" at CHEOPS 2024. Preprint (atlarge-research.com/pdfs/20…) and slides (krien.github.io/downloads/ta…)available online!:

Congratulations Jerrit Eickhoff, on completing your MSc thesis on Polka(DOTS) -- differentiated deployment of online game! Well done! The dissertation is available at: repository.tudelft.nl/island…

Very happy to see the community awards in this year's CompSys'23. Congratulations, @jdonkervliet and Maximilian Berger, and their teams and supervisors! (Many thanks to the amazing #compsys23 organizers, Nirvana Meratnia, Qing Wang, and Kuan-Hsun Chen!)

.@tizianodem presented at #compsys23 his work on scheduling task graph on dataflow architectures, tackling space-time tradeoff and pipelining challenges. #dataflow #scheduling @spcl_eth Paper available at: arxiv.org/abs/2306.02730

This work has been developed in collaboration with the SPEC RG Cloud Group and provides FAIR software tools through Github [github.com/atlarge-research/… ]. The tool has been used by over 100 students in various universities around the world. #continuum #cloud #edge #endpoint

This enables a systematic understanding of existing schedulers and can inform operational teams about the practical benefits of adding specific abstractions to their production schedulers, s, e.g., how much can be gained from supporting VM or container migration. #scheduling

Peter-Jan Gootzen, from Animesh Trivedi's research team at VU Amsterdam, and partners at IBM explore the use of DPUs in the cloud, and show both how to do it and also strong performance improvements #compsys23 Paper accepted at #systor23: dl.acm.org/doi/10.1145/35793…