Zero Copy Integration - Salesforce and Snowflake

What is Zero Copy Integration?

Zero copy integration allows access to data from multiple databases simultaneously without the need to move, copy, or modify anything. This makes data access faster and easier, while also reducing costs and minimizing the risk of errors that typically arise when data is transferred or altered.

Copying data between databases is common, often involving a process called extract-transform-load (ETL), where data is restructured. While ETL is useful for managing enterprise data, physically copying data can be costly, introduce errors, complicate data management, and create delays in synchronization.

How does zero copy integration work? The process varies depending on the platform and whether you're accessing data from a customer data platform (CDP) to a data warehouse or vice versa. In the following examples, Salesforce Data Cloud is the CDP, and Snowflake is the data warehouse, though other platforms could work similarly.

What is a Data Warehouse?

A data warehouse is a secure and efficient place to store and access business-critical data. Traditional data warehouses handle highly-structured data but are often slow and complex. Modern data warehouses like Snowflake, however, can process a wide range of data types quickly and are easier to use. Built on cloud platforms like Amazon and Google, they seamlessly integrate with other systems, including CDPs.

How Zero Copy Integration Works: CDP to Data Warehouse

In this scenario, data from the CDP is shared with the data warehouse (data sharing). Here’s how it typically works:

1. Identify the data in the CDP to share. In Salesforce Data Cloud, this could be data lake objects (clean data), data model objects (structured for business use), or calculated insights objects (like customer lifetime value).
2. Use a point-and-click method to link these objects to the data warehouse (Snowflake).
3. Inside Snowflake, you can run queries that include both Snowflake data and the linked CDP data at the same time.

This process creates "virtual tables," which are like blueprints that point to the data in the CDP. The actual data stays in the CDP, ensuring that any changes made in Salesforce will also appear in Snowflake. Developers and data scientists can use this setup to build AI and machine learning models by combining Salesforce and Snowflake data.

How Zero Copy Integration Works: Data Warehouse to CDP

In this case, the CDP accesses data stored in the data warehouse, a process known as data federation. This is useful because data warehouses often contain vast amounts of data, such as transaction and product data, which can be valuable for calculating customer loyalty or making product recommendations.

Here’s how it works:

1. Salesforce Data Cloud "mounts" tables from the data warehouse as external data lake objects, creating a virtual blueprint of the data.
2. The CDP performs its usual tasks, like identity management and data analysis.
3. The CDP can query both its own data and the data warehouse objects simultaneously through federated queries.

Conclusion

In conclusion, zero copy integration offers a powerful solution for businesses looking to streamline data access and management without the need for costly and error-prone data movement. By allowing seamless communication between customer data platforms (CDPs) and modern data warehouses like Snowflake, organizations can harness the full potential of their data without duplication or delays. This approach not only improves efficiency but also empowers developers, data scientists, and business users to make better, data-driven decisions with real-time insights. As data needs continue to grow, zero copy integration provides a scalable, secure, and effective way to handle large volumes of information across various platforms.