Skip to main content

Two awards for the Division of Software and Computer Systems

Hamid Ghasemirahni and Daniel Lundén, the Division of Software and Computer Systems (SCS)
Published May 05, 2022

We talked to Daniel Lundén and Hamid Ghasemirahni who have both recently been awarded for their research.

Daniel won the Distinguished Artifact Award  and Hamid the community award from NSDI

Hamid, tell us a bit about your winning paper.

"This paper focuses on enabling network applications to work more efficiently in a data center by properly ordering packets. Our proposed solution, Reframer, deliberately delays packets for a very short period of time and re-orders incoming packets right before sending them to be processed. In the evaluation, we show a counter-intuitive result: delaying packets for reordering leads to reduced overall delay and up to 84% improvement in terms of throughput. Given KTH’s large push toward sustainability, this is a significant result because ittranslates to a 46% reduction in the number of servers required to process the same amount of load."

What does this win mean for you?

"NSDI is a top conference in networked systems, and this award means a lot to me. I think it also has two important messages for me – firstly, It shows us we are on the right path and working on critical problems for the community, which I believe is the direct outcome of having supportive and experienced supervisors in the NSLab."

"Secondly, it highlights the importance of integrity and honesty in sharing the code and results so that everyone can trust the proposed solution."

What are you working on at the moment?

"Currently, we are continuing the path to extending our work to cover more applications and network functions. Also, during the work on Reframer, multiple problems and suggested solutions popped up, and we put them on a list of works we can do in the future. So, at the same time, we have eyes on these ideas, too."

Daniel, tell us a bit about your winning artefact.

"The artifact and the corresponding paper culminate many years of work on probabilistic programming languages (PPLs) – programming languages used to encode statistical inference problems. Our main contribution is software for highly efficient PPL inference. In particular, our artifact reproduces the evaluation in our paper, which demonstrates up to 6x speedups in execution time compared to the state-of-the-art."

What does this win mean for you?

"All my co-authors and I are happy to receive recognition for our hard work on the software/artifact and hope that it can be helpful to other people in the future. I would also like to emphasize that the paper and artifact were both team efforts. In particular, my co-author Joey Öhman  did extraordinary work on the RootPPL implementation."

What are you working on at the moment?

"I am currently developing a modular framework for constructing and composing efficient statistical inference algorithms for PPLs. In particular, my collaborators and I intend to apply this framework to infer phylogenetic trees (evolutionary trees of species)."

Packet Order Matters! Improving Application Performance by Deliberately Delaying Packets

At the NSDI22 , by Hamid Ghasemirahni, Tom Barbette, Georgios P. Katsikas, Alireza Farshin, Amir Roozbeh, Massimo Girondi , Marco Chiesa, Gerald Q. Maguire Jr., and Dejan Kostić . 

Compiling Universal Probabilistic Programming Languages with Efficient Parallel Sequential Monte Carlo Inference

Distinguished Artifact Award at the 31st European Symposium on Programming ( ESOP 2022 ) by Daniel Lundén, Joey Öhman, Jan Kudlicka, Viktor Senderov, Fredrik Ronquist, and David Broman.