Snowflake

Snowflake was founded in 2012 by two former Oracle employees with the goal of offering a pure cloud data warehouse. The Cloud Data Warehousing product, the predecessor of today’s Cloud Data Platform, was released in 2014. The technology was developed from scratch. Snowflake is now in its fifth round of financing and is investing primarily in expanding its competitiveness and sales. As of today, Snowflake is not yet profitable but is showing strong growth and market adoption.

The Cloud Data Platform offers simplicity and flexibility in data delivery for different workloads. An important part of this is the encapsulation of administrative database activities (“Near-Zero Management”) to simplify operations. On the other hand, the utilization of the cloud infrastructure and a special architectural approach offers performance and scalability, and creates the basis for efficient data management for six different workloads. In addition to classic data warehousing, the Cloud Data Platform supports data engineering (building SQL-based data pipelines), data lakes, data science, data application design and data exchange.

The platform can be obtained in the Amazon, Google and Microsoft cloud environments and is cloud-agnostic. Most customers use the Amazon cloud. Technically, the platform uses the cloud provider’s object store as central, fail-safe data storage for all kinds of data types and, based on this, offers cloud services and independent processing nodes for the workloads. The processing nodes can be activated as needed and offer dedicated views of the data. They are equipped with computing power depending on the application. Snowflake uses predefined t-shirt sizes to easily configure the amount of computing power. The cloud services provide functions for management and interaction with the processing nodes, such as security, metadata and the interface for data access. Services, processing nodes and data storage can scale completely independently. In this way, individual workloads, for example, do not affect the performance of other workloads. They run in a separate “processing space” but are based on the same data. Special concepts in the architecture (e.g., micro-partitions) enable parallel read/write operations. The platform is also ACID-compliant and can provide data to all processing nodes immediately after committing without delay. Views of older data are provided for 24 hours, but longer periods are also possible.

Data queries are performed via SQL interfaces. Snowflake is completely ANSI SQL compatible. Data transformations can be implemented using SQL. Partners are available to support with more complex data loading/transformation tasks or analysis and data science requirements.

The Data Exchange – which can support collaboration between departments by sharing data – is an interesting workload. Snowflake enables users to easily make their data available and share it within the company or with external partners using special authorization profiles. These are ultimately views of the central data. A copy of the data is not necessary. The platform takes care of data management and security. Cloud Data Platform licensing is based on actual usage.

User & Use Cases

Of Snowflake’s six workloads, most respondents use the Cloud Data Platform for data warehousing or data lake management. Considering Snowflake’s history, this is hardly surprising. It is interesting to note that 44 percent recognize Snowflake as a data platform that embraces workloads beyond classical data warehousing. This supports its positioning as a central access point to multiple workloads using any kind of data. At first glance, it is not clear how 44 percent of users can claim to use Snowflake technology for data warehouse automation. Snowflake does not support any way to automatically generate, run and maintain data warehouse models and associated data loading routes. In our view, the ability to easily create processing nodes, design SQL pipelines or ‘clone’ content is what is meant here.

Snowflake targets mid-sized companies across all industries. 67 percent of our sample of Snowflake customers come from mid-sized companies with a median of 10 users. But the mean of 62 users indicates there are also larger usage scenarios. An average of 7 technical users works with Snowflake, indicating that it is widely used by business users directly. On the other hand, the fact that 7 people should be needed to run and maintain a fully managed, near-zero management cloud database suggests that there is potential for customers to learn how to leverage Snowflake more efficiently.

Current use

n=27

Number of users using Snowflake

n=27

Number of technical users using Snowflake

n=21

Company size (number of employees)

n=27

Summary of Snowflake highlights

Snowflake achieved 8 top rankings and 28 leading ranks across its three peer groups in this year’s Data Management Survey. Remarkably, Snowflake was rated in the top three
in every KPI in all of its peer groups. This outstanding set of results shows that Snowflake‘s multi-purpose cloud data platform is one of the leading databases on the market. Customers attest
to Snowflake‘s promise to deliver ‘customer satisfaction first‘ through support, innovation, good technology and a fair pricing model. In the ‘Developer efficiency‘ KPI, Snowflake achieved the highest possible score of 10/10, while users rated its ‘Performance’ – which is so important for databases – at 9.9/10.

Developer efficiency – Top-ranked

Peer Group: Global vendors

Innovation power – Top-ranked

Peer Group: Global vendors

Performance – Top-ranked

Peer Group: Global vendors

Product satisfaction – Top-ranked

Peer Group: Global vendors

Time to market – Top-ranked

Peer Group: Global vendors

Want to see the whole picture?

BARC’s Vendor Performance Summary contains an overview of The Data Management Survey results based on feedback from Snowflake users, accompanied by expert analyst commentary.

Contact us to purchase the Vendor Performance Summary

Awards

Snowflake

Peer Groups Analytical database products, Data warehouse technologies, Global vendors (data management)
VendorSnowflake
Number of responses27
ProductCloud Data Platform
Offices26 offices worldwide
Employees1,700
Customers3,400
Revenues (2019)n/a
Websitewww.snowflake.com

Test Drive Snowflake

Go here to get a free trial version of Snowflake:

Get Free Trial

Product Video

Screenshots from Snowflake

Quotes

I have a very high opinion of Snowflake. It is a complete SAAS enterprise data warehouse solution.
CIO/IT management, retail/wholesale/trade, more than 2,500 employees
Extremely innovative. Best migration tool for high-end data platforms (Oracle, Teradata, Netezza etc.).
Line of business employee, utilities, 100- 2,500 employees
The architecture changes the options for how to use a database due to its cloud native functionality and separation of compute versus storage. Opens up whole new use cases. Also very easy to charge back to internal users based on actual usage of the product.
Data steward, telecommunications, 100-2,500 employees
It is the best cloud data management platform.
CEO/board member, consulting, 100-2,500 employees
The amount of capabilities you get for the price they charge is off the charts in comparison to other cloud-based solutions out there.
Data engineer/data manager, consulting, 100-2,500 employees