Since last week Snowflake went public. Snowflake now is publicly available for trade on the New York Stock Exchange. What this Snowflake IPO means is explained on the Snowflake blog by Frank Slootman, Snowflake’s CEO.
“Frank Slootman completes his trilogy”
With the the Snowflake IPO, CEO Frank Slootman completes his trilogy. Snowflake’s CEO was already enjoying his long earned retirement. In his career he already brought two different IT companies to the public stock exchange. Both Data Domain (2102) as well as ServiceNow (2017) went public under his watch. If you want to know more about that period, you should read his book; Tape Sucks: Inside Data Domain, A Silicon Valley Growth Story.
Still Snowflake manages to convince Frank Slootman to return from his retirement. He joins Snowflake as he thinks Snowflake is; “The next big thing”. Snowflake solves a problem, the IT knows for a long time; processing data at scale.
I wrote on this blog earlier about the origin story of Snowflake. The founders of Snowflake had one goal; “Simply Load and Query Data”. They had to come up with a platform capable to solve the following challenges:
Big Data — Analysing Machine generated (Big) Data. Handling the volume (scale) and variety (structure) of this data.
Cloud — Elasticity and Compute on demand. Simplicity as in Software-as-a-Service (SaaS). Take all the complexity away from the user.
RDBMS — All the good things from the relational database. And… SQL is the language to query data.
They started white boarding in the snowy mountains. This resulted in the Data Platform built from the Cloud; Snowflake.
Why Snowflake is so interesting?
There is a lot of buzz around Snowflake. Is it just another kid on the block or is Snowflake here to stay? I am positive, it’s the latter. Apparently I am not the only one taking the succes of the Snowflake IPO into account.
What makes Snowflake so interesting? There are many interesting features (like data sharing, zero-copy cloning and time travelling amongst others). All these features are built on top of Snowflake’s unique architecture. This architecture consists of 4 different layers.
1 — Storage — There is no limit on storage. Data is not replicated. All users share one single copy, logically separated by databases.
2 — Multi-cluster Compute — Different clusters of compute (virtual warehouses) per workload (BI, EL-T, department, etc.). Elasticity and compute on demand. Automatically or programmatically suspend and resume. Scale up (for performance) and scale out (for concurrency) automatically. Each department accesses the same single copy of data with their own virtual warehouse (compute resource). Each single virtual warehouse sees the same (change to the) data.
3 — Scale out Services — The Brain of the System. Simplicity as in Software-as-a-Service (DWaaS, DataWarehouse-as-a-Service).
4 — Cloud Agnostic — Snowflake is Cloud Agnostic. This means Snowflake can run on three of the largest Cloud providers (Microsoft Azure, AWS and Google Cloud Platform). Snowflake is offered as one equal solution regardless of the Cloud vendor you choose. So whether you choose Azure, AWS or Google, Snowflake offers the same functionality. How things work in the background for that specific Cloud vendor is handled in the background. With Snowflake you can replicate your data around the globe, across different Cloud providers and across regions automatically.
If you want to dig deeper into what makes Snowflake so damn cool, you should check out this blogpost from Nick Akincilar. I would be happy to talk you through the pro’s and the con’s (if any) as well.
Thanks for reading.
Originally published at https://daanalytics.nl on September 22, 2020.