An overview of Databricks, the company
Databricks was founded in 2013 by seven researchers at the University of California, Berkeley.
This was the time when the world was learning how the Meta, Amazon, Netflix, Google, and Apple (MANGA) companies had built their success by scaling up their use of AI techniques in all aspects of their operations. Of course, they could do this because they invested heavily in talent and infrastructure to build their data and AI systems. Databricks was founded with the mission to enable everyone else to do the same – use data and AI in service of their business, irrespective of their size, scale, or technological prowess.
The mission was to democratize AI. What started as a simple platform, leveraging the open source technologies that the co-founders of Databricks had created, has now evolved into the lakehouse platform, which unifies data, analytics, and AI in one place.
As an interesting side note, and my opinion: To this date, I meet people and organizations that equate Databricks with Apache Spark. This is not correct. The platform indeed debuted with a cloud service for running Apache Spark. However, it is important to understand that Apache Spark was the enabling technology for the big data processing platform. It was not the product. The product is a simple platform that enables the democratization of data and AI.
Databricks is a strong proponent of the open source community. A lot of popular open source projects trace their roots to Databricks, including MLflow, Koalas, and Delta Lake. The profile of these innovations demonstrates the commitment to Databricks’s mission statement of democratizing data and AI. MLflow is an open source technology that enables machine learning (ML) operations or MLOps. Delta Lake is the key innovation that brings reliability, governance, and simplification to data engineering and business intelligence operations on the data lake. It is the key to building the lakehouse on top of cloud storage systems such as Amazon Web Service’s Simple Storage Service (S3), Microsoft Azure’s Azure Data Lake Storage (ADLS), and Google Cloud Storage (GCS), as well as on-premises HDFS systems.
Within the Databricks platform, these open source technologies are firmed up for enterprise readiness. They are blended with platform innovations for various data personas such as data engineers, data scientists, and data analysts. This means that MLflow within the Databricks Lakehouse platform powers enterprise-grade MLOps. Delta Lake within the Databricks Lakehouse platform powers enterprise-grade data engineering and data governance. With the Databricks SQL product, the Databricks Lakehouse platform can power all the business intelligence needs for the enterprise as well!
Technologies and Trademarks
Throughout this book we will refer to trademarked technologies and products. Some notable examples are Apache Spark™, Hive™, Delta Lake™, Power BI™, Tableau™ and others that are inadvertently mentioned.
All such trademarks are implied whenever we mention them in the book. For the sake of brevity and readability, I will omit the use of the ™ symbol in the rest of the book.