My first impression was that this seems related to the S3-FIFO queue data structure that was posted here last year[0], and the personal site of that algorithm's

My first impression was that this seems related to the S3-FIFO queue data structure that was posted here last year[0], and the personal site of that algorithm's main author actually mentions it[1]: > S3-FIFO and SIEVE are being evaluated at several companies, if you are interested in joining the adoption army or if you look for help and want to provide help, find my email on this page. :) Looking into the papers of both algorithms, they basically swapped primary and secondary authors[3][4], with the main author of Sieve being the one to write this blog post. Just sharing this to give some context that could be useful for people who actually know about these topics to assess the claims being made. I don't deal with the problems these data structures aim to solve at my day job, so I'll refrain from commenting on that. But I will say that from an implementation perspective I think the algorithms look elegant in their simplicity, and easy to grok, so at least that part of what they promise is held up. EDIT: Mark Brooker also wrote a blog post about it[4], for those who want to have read the perspective from a third party who has the proper background. [0] https://s3fifo.com/ https://s3fifo.com/ [1] https://junchengyang.com/ https://junchengyang.com/ [2] https://yazhuozhang.com/assets/pdf/nsdi24-sieve.pdf https://yazhuozhang.com/assets/pdf/nsdi24-sieve.pdf [3] https://dl.acm.org/doi/10.1145/3600006.3613147 https://dl.acm.org/doi/10.1145/3600006.3613147 [4] https://brooker.co.za/blog/2023/12/15/sieve.html https://brooker.co.za/blog/2023/12/15/sieve.html