20170721 future of reactive architectures
TRANSCRIPT
Future of Reactive Architectures
Scala Up North, July 21, 2017
Director of Engineering Unified Commerce Platform
Jamie Allen
Microservices Engineers
Senior Microservices Engineers
Principal Microservices Engineers
Technical Managers
It’s easy to lose sight of what companies are really trying to accomplish
Vendor life
Nothing matters more to stakeholders
Delivering business value quickly
Cannot be done unless you can rapidly develop teams
Caveat
FP? Reactive? Yes, use both, where they add value
Only one team had FP help
Our teams each learned differently
Painful making calls between services
Hand-stitched interactions
We’ve avoided that so far
Use Akka Cluster?
FP knowledge doesn’t scale
Critical fact
Critical fact
Neither does Akka knowledge
I need to deliver business value!
It doesn’t matter
Make it easier to quickly deliver business value in a Reactive architecture
So what do we do?
Let the toolchain handle the majority of resilience issues
Use Reactive principles where they’re most needed
Simplify service interactions
gRPC
Orchestration is as much about the ecosystem as the capabilities
Kubernetes
Service mesh with additional telemetry and security support
Istio
Absolutely not
Is Starbucks abandoning Scala and Akka?
What about data? Turns out, we were wrong here too
CAP Theorem isn’t what we thought
Next generation storage
CAP Theorem isn’t what we thought
Next generation storage
Google: Calvin Fauna
Do we need to be Reactive everywhere?
Choose how to be reactive in delivering business value quickly
Thank you!Questions?