Sometimes you care about how the technical magic is performed behind the curtain. The ‘two french dudes’ and their friends have generously shared their notes with us. Snowflake Paper 2016 http://pages.cs.wisc.edu/~yxy/cs839-s20/papers/snowflake.pdf Paper annotations by Nat Taylor Snowtrail Paper - FiCloud 2023 Smart Query Sampling By Jerry Tang & et al (2023). FoundationDB Paper 2021 https://www.foundationdb.org/files/fdb-paper.pdf Carnegie Mellon Advanced Database Systems 15-721 - 2018 - Ashish Motivala https://15721.courses.cs.cmu.edu/spring2018/slides/25-snowflake.pdf Carnegie Mellon Quarantine Database Tech Talks 2020 - Query Optimization - Jiaqi Yan https://db.

I serendipitously encountered a cool CS paper on memory-efficient data structures for search. Huanchen Zhang gives an excellent talk about it below. Besure to catch the Q&A at the end - there’s a funny exchange. :)

I’m a big fan of Robin Moffatt’s writings and speaking content. I owe him a beer or two from all I have learned just from his prolific work on Kafka Connect. However, when I first encountered his article about configuring a sink to play a topic from the tail, rather than from the head, I didn’t understand why on earth anyone would want to do so. Well, now I do and let me share one possible scenario where this could save the day.

Recently, a random chain of events pushed me to develop a small degree of expertise with Snowflake Kafka Connect sink plugin. While this open-source project has some configuration documentation, and a README, it was not obvious to me how to initiate a build or get one of the tests up and running. After a bit of trial and error, I wrote down the step-by-step instructions below in case I need to do it again later.

macOS Snowsql CLI installation and setup in 3 easy steps: Install brew Install brew via one copy-paste from here Install snowsql binary brew cask install snowflake-snowsql Configure snowsql vi ~/.snowsql/config follow example config file test and launch $ snowsql # default set of configurations $ snowsql -c training # training set of configurations

All I wanted was a graph of real-time broadband utilization of my household. A fully Unifi‘ed network sure looks pretty, but USG just couldn’t do it. Times have changed. Many home networks are wired up for 1.0 Gbps, and today’s open source firewalls have the features I need. Like real-time traffic graphs. My trusty USG served me well for years, but off to the ebay it goes. In its place, I put together some OPNsense hardware and configured it to bypass AT&T Fiber CPE (BW210-700).