In an epic three-hour keynote address at AWS re:Invent that featured a live UFC match on stage, famed AWS CTO Werner Vogels presented a dizzying array of insights, predictions, suggestions, and directives. Oh, and he tossed out a few cool product announcements, including the new Amazon Cloud9 IDE, Amazon Alexa for Business, and the AWS Serverless Application Repository.
But Werner is an always fascinating visionary, so we wanted to share 20 of the most interesting things he told AWS re:Invent attendees during his talk on “21st Century Architectures, re:Imagined.” He offered his unique take on everything from the coming primacy of voice interfaces to the security responsibilities of developers to the cost of high availability. (Some points have been slightly edited for clarity.)
- Today’s top technology drivers are data, IoT, powerful P3 Instances, and deep learning frameworks.
- We allow our customers to set our roadmap with quickly iterated development tools that allow you to develop the way you want to develop … for the systems you want to be running in 2020. We continue to build services for the future, not the way we developed in the past. You have to understand that the system you’re building now is not the system you will be running in 6 months or 2 years.
- Our goal has always been to build a collection of very fast and nimble tools for you to use, but with 3,951 new features and services introduced since the first re:Invent in 2012, we know the pace can be confusing.
- The six principles of AWS Well Architected are:
- Stop guessing capacity needs
- Test systems at production scale
- Automate to make architectural experimentation easier
- Allow for evolutionary architectures
- Drive your architecture using data
- Improve with the help of Game Days
- With P3 Instances, we can now build neural networks that we can execute in real time. That’s a big shift in how we access digital systems and has major implications for machine learning.
- Digital access has been driven by the capabilities of the machine, not how we want to work with them. But with new computing power, the future will be human-centric instead of machine-centric, with natural-feeling interfaces where the whole environment will be active via voice, vision, touch—I’m not really sure we can do smell.
- Voice is the next major disruptor in computing. Voice is the first disruptor that will happen by the deep learning systems that we are giving you.
- Voice creates digital systems for everyone, not just digital natives who know how to use keyboards.… If you’re a young dad and your kid is ill, you don’t want to go online and fill out a form, you want to yell at something.
- Alexa devices are not actually that smart. All the smarts live in the cloud, in the Alexa Skills Kit and the Alexa Voice Service.
- Voice will also change the way backend systems work. Voice becoming the primary interface will change the way your systems work.
- Conference room systems at work annoy the hell out of all of us—or at least me. Connecting to the system is why meetings always start 10 minutes late. Alexa for Business will first make sure Alexa works well in conference rooms. You can just say, “Alexa, start the meeting”—no need to enter room codes anymore.
- The six ways development has changed:
- More security aware
- More collaborative
- More languages
- More services
- More mobile
- QA and ops are integrated
- Protecting your customers should be your #1 priority. Without that, you don’t have a business. It should come before any features.
- You need to protect data in transit and at rest. I believe we haven’t taken encryption seriously enough. Encryption is the only tool we have to protect your data. “Dance like no one is watching. Encrypt like everyone is.”
- Security is everyone’s job now, not just the security team’s. With continuous integration and continuous deployment, all developers have to be security engineers … we move too fast for there to be time for reviews by the security team beforehand. That needs automation, and it needs to be integrated into your process. Each and every piece should get security integrated into it … before and after being deployed.
- We all want to build high-availability systems that are up 100% of the time. But the next failure will happen.
- Business rules drive availability choices. We all want 100% availability, but that actually costs money. Lots of services may be completely fine with 99% reliability; it may go down occasionally, but it’s cheap. So before paying for 100%, think about your business reasons for wanting it.
- Test, test, test. Break everything to see how your systems respond. Game Days and chaos engineering can be invaluable.
- Galls Law holds that “A complex system that works is invariably found to have evolved from a simple system that worked.“ But complex services managed for you allow you do to less and less so you can focus on the business logic that you are writing.
- In the future, all the code you ever write will be business logic.
Bonus tidbit of wisdom: “You may be thinking you need more chaos in your life. And you absolutely do. Chaos doesn’t cause problems, it reveals them. Ask not ‘What happens if this fails?’ but ‘What happens when this fails?’”—guest speaker Nora Jones, senior chaos engineer at Netflix and co-author of Chaos Engineering, Building Confidence in System Behavior Through Experiments.
Watch Werner Vogels’ full AWS re:Invent keynote presentation in the video below:
Check out more coverage of this year’s AWS re:Invent here on Storify.