Skip to content

Instantly share code, notes, and snippets.

@aviflax
Last active April 9, 2019 11:33
Show Gist options
  • Save aviflax/cd894c83a523d421127d7b071ff71c94 to your computer and use it in GitHub Desktop.
Save aviflax/cd894c83a523d421127d7b071ff71c94 to your computer and use it in GitHub Desktop.
Evaluating Druid for our Data Warehouse

Evaluating Druid for our Data Warehouse

Druid

Imply

Prominent People

  • Fangjin Yang is one of the creators of Druid and CEO of Imply

Articles

Talks

  • "Druid: Powering Interactive Data Applications at Scale" by Fangjin Yang at Strange Loop in September 2016
    • Excellent talk that was our first introduction to Druid and first made us think it might be able to replace our data warehouse
  • Building Interactive Data Applications at Scale by Fangjin Yang and Vadim Ogeivetsky, at Strata in Feb 2015
    • The first half of this talk is an earlier version of the first half of the above talk, so it’s not necessary to watch the first halves of both talks
    • The second half, by Vadim Ogeivetsky (who was working at Metamarkets at the time and is now at Imply) is a fascinating look at getting data out of Druid and similar systems for the purposes of visualizations.
    • Ogeivetsky introduces a high-level conceptual model for such queries that is decoupled from the specific model of Druid or any other data storage system. He also introduces a library/framework for translating high-level queries into lower-level Druid queries: Facet.js.
    • Facet.js looks extremely compelling, but unfortunately it looks like it’s not available right now. It seems that Imply forked Facet.js and renamed their version to “Plywood”. But Metamarkets, who owned Facet.js, shut down Facet.js and sent GitHub a DMCA takedown notice for Plywood, so the GitHub repo is currently unavailable. The code is still available via NPM though, so we could probably use it even while the IP battle is ongoing.

Misc

  • Imply has a community mailing list they call Imply User Group
    • It’s set up as “private” for some reason but anyone can request to join and they seem to approve most requests quickly
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment