Reliable Software Systems (CSE490R1)

Reliable Software Systems
CSE 490 R1 Winter 2019 Meets Thursdays 1:30 – 2:20 PM Room CSE 403 Discussion group/office hours after class in the 4th floor breakout	DESCRIPTION	Nowadays, software engineers build pieces of systems that rely on other systems, and other systems rely on them. In this interconnected world, every engineer needs to know how to identify and mitigate failures in their system. Furthermore, many companies even have their software engineers doing operational/oncall work for their own systems, so engineers should be even more motivated to build reliable systems. This seminar will help students become familiar with various industry practices for creating and running reliable software systems.
	ADMINISTRATIVE	Instructor: Alyssa Pittman Contact: smooo [at] cs.washington.edu (yes that's three o's, make sure you get them all) Expectations: This is a one credit, C/NC course. Please show up, ask questions, and discuss! Feedback: I welcome your input, either in person or online (supports anonymous feedback). Don't know how to say it? Check out the SBI model for a structure of how to give effective feedback.
	SCHEDULE	Week 1: Reliability? Systems? Slides: PDF Motivating outage: Maersk Shipping Optional readings: We’re all ops now Google on Distributed Systems Week 2: Expect Failure Slides: PDF Motivating outage: Slack Optional readings: Using retries Cascading failure Black swans Week 3: Monitoring Slides: PDF Motivating outage: Instapaper Demo: Honeycomb monitoring Optional readings: "Monitoring Distributed Systems" from the SRE Book Stack Overflow monitoring overview USE method for monitoring and debugging Week 4: Preproduction Slides: PDF Motivating example: Healthcare.gov launch Optional readings: Martin Fowler on CI Introduction to CI/CD Increment on staging environments Week 5: Production Slides: PDF Motivating example: Windows 10 October 2018 Update Optional readings: Testing in Production Canarying Releases Chaos Engineering: the history, principles, and practice Bonus interview on Chaos Engineering since we ran out of time in class: Testing in Production the Netflix Way Week 6: Data Integrity Slides: PDF Motivating example: Gitlab data loss Optional readings: SRE Book: Data Integrity The Hardest Part About Microservices: Your Data Week 7: Designing Scalable APIs Slides: PDF Motivating example: Twitter v1 API outages and retirement Optional readings: Four Mantras for Designing Scalable APIs Best Practices for API Error Handling Defining Stateful vs Stateless Web Services Week 8: Scalable Design Patterns Slides: PDF Motivating example: Foursquare Optional readings: Introduction to Architecting for Scale Scalability for Dummies: Clones Database Cache Asynchronism Week 9: Team Culture Slides: PDF Motivating example: Titan II Optional readings: Simplicity High Performing Teams Need Psychological Safety Post Mortems to the Rescue Week 10: Students' Choice: Netflix, Youtube, Twitter Architecture Slides: PDF Motivating example: early Netflix Optional readings: Netflix: What happens when you press play? (and more) YouTube Scalability (and more) Twitter, an Evolving Architecture (and more)