Accelerating Cloud Adoption

Chapter 8. Conclusion: Moving Forward, Embracing Change

When I first started my cloud journey in 2008, the public cloud was used primarily for infrastructure services for compute, network, and storage. By 2010, the CSPs were offering fully managed services for databases that auto-provisioned, managed, and scaled all of the necessary infrastructure and database software for you. Then came containers, functions as a service (FaaS), data streaming and ingestion services, and more abstractions of specific technology components. Next, the providers started abstracting entire technology stacks for specific use cases like Internet of Things (IoT), blockchain, and gaming. As I write this in 2020, the CSPs are abstracting entire business functions. Google’s Healthcare API, for example, is designed to ingest healthcare data and images in industry-specific standard formats, complete with de-identification logic for masking personal data.

The difference is striking. Think about the recommendations Amazon offers customers: based on what you’ve bought and what it knows about you, the site recommends new purchases, often with incredible accuracy. Back in the 1990s, I had a team of developers who worked on building code that would produce similar recommendations. We wrote hundreds of thousands of lines of code, which ran on massive infrastructure that needed to be maintained. Today, Amazon offers its purchase-behavior recommendation logic as an API. I could replace all of that code, infrastructure, ...