Designed for the Cloud

Live blog post from re:Invent 2018, Las Vegas

One consistent theme in Werner Vogels’s keynote speech at re:Invent yesterday was that products designed for the cloud are very different from those designed for on-premises hardware. In the first waves of migration to the cloud and cloud-first delivery of new systems, it was natural for all of us to think of moving the software we use every day to the cloud and simply running it on cloud compute and storage instances. For example, AWS quickly made RDS available as a managed way to run old-style databases, since this was what our customers wanted.

But in the keynote, Werner walked us through the logic AWS engineers used in rethinking what a database specifically designed for the cloud needed to look like. The cloud presents all sorts of opportunities to increase resilience, security, and performance at scale, but to gain these benefits, applications—databases, for example—must be designed in a way that takes advantage of those cloud capabilities. For example, he talked about how designers of AWS services were able to develop “cell-based” architectures to limit the blast radius of failures and make their services more resilient, and how database-aware storage could vastly increase performance over traditional relational database models. He showed how AWS used these principles in designing and implementing Amazon Aurora, a relational database compatible with MySQL, Postgres, and MariaDB, but highly performant at scale and resilient because it takes advantage of these cloud-native capabilities.

This improved resilience, security and performance at scale have become increasingly critical for enterprises who use the products. As our datasets become larger and larger, as more and more users take advantage of digital services—more and more frequently—the need for scalability continues to increase. As consumer expectations increase, and as real-time applications such as many IoT applications proliferate, performance (speed) must also increase, and database access is very often the limiting factor. And as bad actors develop more and more sophisticated ways of compromising our security, our software platforms must become more and more rugged. Design for the cloud helps us in all of these areas.

Companies as varied as Dow Jones, NTT Docomo, Capital One, Netflix, and Verizon all use Amazon Aurora to achieve high performance at scale in the cloud. But database software is just one example of how cloud native design opens up new possibilities for software platforms. All of the software platforms that we build our enterprise applications on can be improved by redesigning them for the cloud, where they can take advantage of elasticity, containerization, and the variety of compute instance types—as a few examples of new capabilities—to achieve much higher levels of performance, scalability, resilience, and security. The floodgates have been opened…now let us see how creative software vendors and providers like AWS can be with their rethinking of product and service design.

–Mark

@schwartz_cio
A Seat at the Table: IT Leadership in the Age of Agility
The Art of Business Value
War and Peace and IT: Business Leadership, Technology, and Success in the Digital Age (now available for pre-order!)