Snowflake Architecture and Editions

1. Snowflake Architecture and Editions

Welcome to this course on Snowflake Architecture.

2. Snowflake Functional Architecture

Here, you will build the foundation to becoming SnowPro Core Certified in no time! Let's start with understanding how Snowflake is actually structured.

3. The Problem With Traditional Databases

Let's start by looking at traditional databases, these databases were tightly coupled meaning storage and compute ran on the same physical hardware. This worked fine on a smaller scale, but when data teams grew or teams of analysts were running multiple queries at the same time, you very quickly hit a wall - you couldn’t simply add more compute, you had to scale both storage and compute together, whether you needed it or not.

4. Three Layers: Storage

Snowflake was founded in 2012 with a fundamentally different approach: decoupling these pieces into three separate layers. First, the storage layer where your datasets compressed in a columnar based format on whichever cloud provider you're running on.

5. Three Layers: Compute

Second, the compute layer, which is where queries actually execute, inside what Snowflake calls virtual warehouses.

6. Three Layers: Cloud Services

And finally, the cloud services layer. This layer coordinates everything from authentication to metadata management, query optimization, and transactional control, all before a query ever touches the compute layer. Each layer talks to the other layers. None of them are locked together.

7. Storage and Compute

When you first use Snowflake, it can feel counterintuitive because storage and compute are billed separately. Your data is always there, at a flat ongoing storage rate, but compute only costs you when work is actually happening. A virtual warehouse that's sitting idle costs you nothing in Compute. That changes how you think about resource allocation because you can spin up a dedicated warehouse for a one-off job, let it suspend when it's done, and the cost stops there.

8. Shared Storage Layer

Think about it like this: one storage layer sits at the center with three separate warehouses connected to it - one for the engineering team, one for the product team, one for the marketing team workloads. Same data but with no shared compute. This means, if the marketing team kicks off a large export, the engineering team doesn't feel it. Workload isolation is one of the core reasons organizations choose Snowflake over traditional architectures.

9. Snowflake Editions

Now let’s get into Snowflake Editions. Snowflake offers four types, but the most commonly used and the ones primarily referenced in this course are Standard and Enterprise. Standard is a good fit for most teams getting started, like a startup running internal analytics, for example. Enterprise is where scaling organizations tend to land, such as a retail business running reporting across multiple teams with longer data retention needs. Snowflake also offers a Business Critical edition focused on compliance, think patient records, financial data or anything with a regulatory requirement. They also offer Virtual Private Snowflake which provides a fully dedicated infrastructure - think government agencies or defense contractors where shared resources are not an option. This table has the full feature breakdown.

10. Global Coverage

Lastly, Snowflake runs on all three major cloud providers: AWS, Azure, and Google Cloud which can be deployed across regions globally. Your account lives on one provider in one region, but the platform itself is cloud-agnostic.

11. Prepare for the SnowPro Core Certification

Before we wrap up, I, Emily, a technical curriculum developer at Snowflake, would like to offer you a quick word on how this course fits the bigger picture. This is the first course in the Snowpro Core Certification track on DataCamp, and it lines up with the Snowflake University Platform Skills badge. It is also the first of three courses in a series aimed at getting you ready for the SnowPro Core certification. We are not starting from zero: you should already be comfortable with everyday Snowflake concepts and with writing SQL. If you're completely new to these concepts we recommend you take our Introductory SQL and Snowflake courses, such as Introduction to Snowflake first.

12. SnowPro® Core Certification (COF-C03)

This track has been co-created with Snowflake to help you pass the SnowPro Core certification exam. The courses cover the key skills assessed in the exam, including Snowflake architecture, account management, data loading, performance optimization and collaboration. Complete the track and earn a $50 discount code from Snowflake toward your SnowPro Core exam — a great way to validate your Snowflake expertise.

13. Let's practice!

Over the coming chapters, we will continue to build on the foundations to get you ready to become a SnowPro! Let’s put your knowledge of the Snowflake Architecture to the test.

Create Your Free Account

By continuing, you accept our Terms of Use, our Privacy Policy and that your data is stored in the USA.