Architecting Data-Intensive SaaS Applications

Chapter 5. Data Sharing

As we saw in Chapter 2, data sharing is an important requirement for data applications. In this chapter we will take a deep dive into this subject and how this impacts data applications.

We’ll start with a discussion of different approaches to sharing data, then move on to design considerations in data applications. Next, you will learn about Snowflake’s architecture, which eliminates the storage costs and overhead of traditional approaches.

In addition to sharing data among different parties, the ability to discover data is also an important element of data applications. For data consumers, this means knowing what data is available and how to get it. For data providers it means ensuring potential customers know about their offerings. You will learn how Snowflake Data Marketplace solves the data discovery problem, building a global data network to drive the data economy.

To provide an example of how data sharing in the Snowflake Data Cloud benefits data application builders, we will conclude with an overview of how Snowflake partner Braze leverages data sharing to drive their business.

Data Sharing Approaches

In this section we will discuss two different approaches for data sharing: sharing by creating copies of the data and sharing references to the data.

Sharing by Copy

The legacy approach to data sharing is to create copies of data to distribute to consumers, as illustrated in Figure 5-1.

Figure 5-1. Sharing through data copy

Data providers export ...