Data-warehouse company Snowflake went public this week, reaching an extraordinary market value of $70.4 billion, the largest IPO for a software company ever. Snowflake offers cloud-based data warehousing, whose data storage and query engine contain techniques pioneered in CWI’s Data Architectures group. One of the group’s PhD graduates who developed that technology, Marcin Żukowski, is co-founder of Snowflake.
Behind one of the most remarkable IPOs ever lies the story of CWI researchers who, a decade earlier, laid the foundations for a revolutionary new way of data management. The PhD thesis of Marcin Żukowski, which he published in 2009, contains the blueprint of modern analytical database technology. Żukowski and his thesis advisor Peter Boncz brought this technology to market in 2008, via the CWI spin-off VectorWise. After VectorWise was acquired, and realizing the huge potential of data management in the cloud, Żukowski went on to co-found Snowflake in 2012.
Already managing 250 million gigabytes of data for 1,300 customers worldwide, Snowflake is regarded is one of the most promising data companies in the world. This week, Snowflake made its debut on Wall Street. Snowflake’s stock opened at more than double its listing price and then soared in early trading afterwards. This resulted in the biggest tech IPO of the year, and the largest IPO for a software company ever, with a market value of $70.4 billion
Core technologies from CWI
In its data storage and query engine, Snowflake uses two technologies called vectorized query execution and lightweight compression methods in its columnar data storage. Both were techniques invented at CWI, to make data analysis for more efficient than current methods.
Among all analytical data management products, both techniques are now widely used. For instance, Databricks, one of CWI's collaborators who last year announced a $100M investment in Amsterdam, recently announced to be switching to vectorized execution as well in their new Delta Engine.
The market for cloud-based data storage is growing tremendously, as it offers many advantages for companies and other organizations that want to manage large amounts of data. Cloud storage frees them from the burden of maintaining their own data infrastructure, and keeping it safe and up to date.
Analytical data services providers in the cloud not only care about performance, but also cost, as storage and queries are charged to their customers by the amount of used resources. As such, the adoption of the widely recognized efficient techniques developed at CWI, are quite important for an analytical data management service in the cloud.
Żukowski’s PhD research at CWI was supervised Peter Boncz and Martin Kersten from CWI’s Database Architectures group.