Summary
In this chapter, we examined the differences between Data Warehouses and Data Lakes and the advantages of both approaches. We have the structured model of the Data Warehouse with its security, tuning possibilities, and accessibility for Self-Service BI on the one hand, and the capabilities of Data Lake systems to process vast amounts of data in high performance to support machine learning on the other.
Both concepts, when implemented in isolation, can help solve certain problems. However, even with the growing data, the disparate source systems, their various formats, and the required speed of delivery, as well as the requirements for security and usability, neither can succeed on their own: there is life in the old Data Warehouse yet!
The combination of the two concepts, together with the extended offerings of the Hyperscaler cloud vendors such as virtualization, container offerings, and serverless functions can open new opportunities in terms of the flexibility, agility, and speed of implementation. We are getting the best of both worlds.
In the next chapter, we will discuss a generic architecture sketch. You will learn about the different building blocks of a Modern Data Warehouse approach and how to ask the right questions during your requirements engineering process. In the second part of the next chapter, we will examine Azure Data Services and PaaS components. We'll explore alternative components for different sizes and map the Azure Services to the Modern Data Warehouse architecture.