ClickHouse December Community Events

The year 2019 ended with a burst of ClickHouse community events in the US and Europe.  There were many great talks about ClickHouse from users exploring new use cases as well as operating at large scale.  Here is a quick rundown.

San Francisco ClickHouse Meetup – 3 December

Cisco sponsored the December ClickHouse Meeting at their Open Late venue in downtown San Francisco.  It’s a great meeting location and we had about 45 attendees, making it the most popular meetup of the year. In addition to networking over pizza and beer, we had two great talks from local users. 

  • ClickHouse Capacity Planning: Methodology for OLAP Workloads — Mik Kocikowski of Cloudflare worked through the steps to size resources for large ClickHouse installations.  Cloudflare ingests a petabyte of data daily. Capacity planning there balances scaling out shards to match ingest rates, then increasing replicas to handle queries on the data.  

  • Unified Data Platform — Pauline Yeung of Cisco worked through three new use cases that she’s exploring related to discovering security threats from DNS log records, analysis of whois records, and monitoring networking tunnels.  ClickHouse is widely used for network management and these use cases provoked a lot of discussion and ideas about efficient implementation. 

Attendees stayed around until the office closed at 9pm discussing their own ClickHouse efforts and sharing ideas.  We would like to give a big shoutout to Cisco for hosting the event. The next San Francisco ClickHouse meetup will be on 3 February at Cloudflare, another great  ClickHouse supporter. 

Moscow ClickHouse Meetup – 11 December

Moscow ClickHouse meetups are traditionally organized by Yandex in a big ‘Extropolis’ conference hall that can welcome 200-250 people. This time it was nearly full at 7pm when the meetup started, but not everybody survived until the last talk ended at a remarkable 11:30pm. It took so long thanks to 7 interesting talks from ClickHouse developers and community users:

  • Alexander Krasheninnikov from Badoo is a frequent speaker at Moscow meetups. This time he opened up a meetup with an interesting talk about time series prediction models in ClickHouse. By the way, Alexander is an official maintainer of the ClickHouse JDBC driver.

  • Artem Zuikov from Yandex ClickHouse team invited everybody to a fascinating journey through the different JOIN types available in ClickHouse. For example, have you ever heard of SEMI JOIN or ANTI JOIN? Please check the slides to learn more.

  • Andrey Konyaev from ArenaData presented an interesting use case optimizing queries with new ClickHouse skipping indices. These are special data structures they can improve query performance enormously when properly used. Andrey demonstrated 15-17X improvement on his benchmark. We plan to publish a research on this topic soon as well.

  • DBeaver team talked about ClickHouse JDBC drivers and compatibility issues they are facing. DBeaver is a nice SQL UI supporting multiple databases, including ClickHouse.

  • Ivan Moskvin from Yandex.Cloud team talked about the architecture of Yandex DataLens — BI SaaS available for Yandex.Cloud. DataLens uses ClickHouse cluster internally in order to materialize intermediate data collected from different sources.

  • Altinity contributed to the meetup through Vladimir’s Chebotaryov’s detailed overview of multi volume storage. Vladimir did a lot of work recently, submitting over 50 PRs polishing this functionality. He talked about use cases and design, and even merged our newest addition — TTL storage policies — in real time!

Traditionally Alexey Milovidov closes a meetup with a ‘what’s new’ kind of talk. It was not an exception this time as well. ClickHouse is evolving so rapidly that it is not easy to catch up with all new features. Alexey reviewed recent additions and talked about the ambitious 2020 development plans. We are going to share those plans in our blog later in December. Meantime, check out Alexey’s slides.

London Open Source Meetup – 16 December

London Open Source Database Practices Meetup was organized by the ProxySQL Team and was held at the Cloudflare office, very close to the breathtaking London Eye and the Christmas market.  Big thanks to ProxySQL and Stacy Rostova for flawless organization as well as to Cloudflare for the venue and beverages. Some 45 people attended the event, which featured three presentations:

It seemed that London is thirsty for news and meet-ups that help users operate large-scale IT systems. Even though quite a lot of people are just beginning use of the technologies presented in the talks, the problem statements were definitely aimed at operations at scale. We look forward to coming back to London soon to quench that thirst. 

Share