Explore the concept of perfect scalability in software systems during this 51-minute conference talk from Scala Days New York 2016. Delve into the architectural constraints and design principles that enable near-linear scaling of systems. Learn why scalability is often more about what to avoid rather than what to implement. Examine real-world examples from IoT to high finance, and discover how to leverage the Lightbend ecosystem, including Scala and Akka, to build highly scalable applications. Investigate the role of microservices, continuous delivery, and monitoring in creating systems that can scale across data centers and continents. Gain insights into overcoming common scalability pitfalls, understanding Amdahl's and Gunther's Laws, and implementing strategies to avoid contention, shared state, and system failures. Explore the balance between scalability and other system requirements, and learn how to design architectures that can handle extreme loads without compromising performance.
Overview
Syllabus
Intro
What?
A Perfectly Scalable System
Why does this system scale?
How do we fix it?
Closer to Perfection
Amdahl's Law
Gunther's Law
Avoiding the Wall
The Enemies of Scalability
Avoiding Contention
Avoiding Sharing
Avoiding Shared State
Limit Communication
Avoiding Ordering
Avoiding System Failure
Avoid Linear Time
Designing for Perfect Scalability
Scalable Architecture
Scaling Down
Ultra Large Scale Systems
Spike Load
Command-Sourcing
Degrading Gracefully
Microservices
Simple is Good
Persistence is (Not) Futile
Keep your hands out of "OPD"
Distributed Transactions
Allow me to repeat myself..
DDD and Scalability Onion Architecture
Balancing Scalability
Trust, but Verify
The Log is not Enough
DevOps Matters
Conclusion
Questions?
Taught by
Scala Days Conferences
