- Dexter Pontañeles
It’s been a week since we got back from aws:reInvent 2022. Even though I spent most of that week in workshops, what endured and lingered in my mind is a keynote by Dr. Werner Vogel. Dr. Vogel is the CTO and VP of Amazon in charge of driving innovation within the company. I managed to watch his keynote only at a live stream session despite me being at the event because I had to prioritize a workshop at a different location. Had I known that this would turn out to be my favorite session, I should have planned my agenda better so I could go to the Venetian, where the keynote was held.
Livestream (overflow) session of the keynote at Mandalay Bay.
Passion for systems architecture
Dr. Vogel’s keynote was filled with passion for systems architecture. He opened the keynote with a matrix themed introduction video showing the downsides of a hypothetical synchronous world (red pill). He then entered the stage with a lambda T-shirt, which could very well be a half-life game logo or both.
The real world
While surely the introduction was done to inject humor to the talk, it drove home the point that we should design our systems based on the real world, which is entirely asynchronous. Building asynchronous, loosely coupled event-driven distributed systems allows us to build systems that scale. This also allows us to make resilient and evolvable systems as in real-world systems. He narrated that a lot of these concepts and patterns are the keys to the success of S3. AWS S3 was developed with these distributed concepts in mind: Asynchrony, autonomy, local decision, decisions on local data, simplicity, controlled asymmetry, symmetry, and simplicity. He further reinforced that event-driven microservice cloud architecture as the way in order to solve most of our common systems requirements today.
The author, Dexter Pontañeles at the aws:reInvent 2022 Expo.
Dr. Vogel noted that event-driven serverless applications are complex and difficult to build. With this challenge in mind, he announced new tools and features that help ease this: AWS Application Composer, EventBridge Pipes, Amazon CodeCatalyst, and AWS Step Functions Distributed Map.
**AWS Application Composer**
In the keynote, AWS launched the new AWS Application Composer which is a graphical tool for creating serverless applications. It allows the user to create the layout of the application architecture in a simple drag-and-drop interface where he can connect the resources and design their functions. Besides making it easier to author serverless applications, it should also enable easier sharing and collaboration of the application. The application can automatically sync the configuration generated with the developer’s local filesystem so they can work with their IDE of choice if they want to.
It is currently in preview mode and available only in the following regions:
- Europe (Frankfurt & Ireland), US West (Oregon), Asia Pacific (Tokyo), US East (N. Virginia), US East (Ohio)
In addition to the Application composer, Dr. Vogel also announced the following new features:
Allows to easily stitch AWS services together. It is specifically designed for integrating messages from different AWS services and a way to manipulate events that flow through the pipe. It reduces glue code usually needed to enable this. It also has built-in event filtering.
A unified software development service that makes it faster to build and deploy in AWS using blueprints. You should be able to create private blueprints so you can reuse blueprints within your company. The blueprints contain easy CI/CD pipeline setups. It also makes it easy to switch between multiple codebases and supports multiple IDEs and integration with different code repository systems.
Available today in US West(Oregon)
AWS Step Functions Distributed Map
Allows you to easily process large amount of data using simple lambda functions by using Map state in Distributed mode. This mode allows input from large-scale AWS S3 data sources.
Finally, the keynote demonstrated how AWS powers next generation applications by featuring customer applications. Applications that closely model reality using enormous data sets based on multi dimensional data, and inputs from multiple sensors; and state of the art systems that perform complex simulations. He ended the keynote with a call to build bigger, better, bolder systems much faster by finding inspiration from the universe itself.