Today, Lightspeed is excited to announce our investment in the $250M Series B financing of ClickHouse, an open-source, real-time OLAP database. As database geeks at heart, we’ve seen for several years ClickHouse ‘in the wild’ and had been in awe of its performance and scalability as a database.
There is a saying in the database industry that it takes roughly 10 years or more to get a database to the point where it’s truly robust and reliable at scale. I know from my time running Product at Elastic, that building distributed systems is incredibly hard, and there is a constant stream of bugs, edge cases and tradeoffs, that only rear themselves in production workloads. It is one reason why most new DBs that are launched never achieve widespread success.
Clickhouse’s unusual origins are precisely what allows it to overcome these challenges. It was originally developed at Yandex (the ‘Google of Russia’) and began as a solution for web analytics called Metrica, which is a service used for analyzing web traffic, and #2 in the market behind Google Analytics. In 2009, Alexey Milovidov, a brilliant engineer on the Metrica team, was looking for a database that could create reports on metrics like the number of page views per day, unique visitors, and bounce rate, without aggregating the data in advance. Existing solutions could not support real-time, large datasets, linear scaling, high efficiency, and compatibility with SQL tools. Metrica evolved into ClickHouse, short for “Clickstream Data Warehouse”. In 2016, Yandex open-sourced ClickHouse and broader adoption led to the discovery of new analytics use cases including DNS queries, advertising, operational logs, stock correlation, e-commerce/travel, security/fraud, and web/mobile. CloudFlare and Uber published widely read blogs about their success.
At one Lightspeed portfolio company, they successfully replaced hundreds of MongoDB servers with 1/20 the number of Clickhouse servers, while also improving query response time by 30x. Many of these real-time use cases cannot be effectively served by cloud data warehouses like Snowflake, which by design separate compute and storage. While this separation brought lower costs and ease of operation over traditional databases, it lacks the performance of the ClickHouse architecture.
It was 8 months ago that I had lunch with Aaron Katz, the co-founder/CEO of Clickhouse, who had previously been CRO at Elastic where he scaled revenues from $15M to $500M. He told me about the plan he was hatching to spin-out the Clickhouse IP along with its core engineers into an independent company, in partnership with Yandex. Later, I met co-founders Alexey Milovidov (creator of ClickHouse) and Yury Izrailevsky (who built cloud services at Google and Netflix) who are true database and cloud veterans. It’s a dream team and we’re honored to be part of the journey.
Authors
