Upgrade to Pro — share decks privately, control downloads, hide ads and more …

Beyond DevOps - Operability for Serverless and IoT

Beyond DevOps - Operability for Serverless and IoT

Focus on operability to get the best out of Serverless and IoT

Working with Serverless and IoT platforms means that much of the infrastructure “heavy lifting” has been done by the cloud vendor, leaving engineering teams free to focus on operability as a key differentiator for products and services.

In this talk we’ll cover:
● The coming technology landscape for business systems
● Why operability is key for future software
● Practical operability techniques for Serverless and IoT

Background: DevOps practices have been hugely valuable over the past 10 years in making software systems more resilient, reliable and valuable for users. However, many of the challenges that DevOps set out to solve are being addressed by cloud vendors at a rapid pace: application platforms, automated deployments, dynamic infrastructure provisioning, and more. Organisations starting their digital transformations in 2018 and beyond should look to new approaches such as Serverless and the cloud vendor IoT offerings and avoid
older approaches such as IaaS and even containers.

---

A talk given at University College, London in 2018

Matthew Skelton

March 22, 2018
Tweet

More Decks by Matthew Skelton

Other Decks in Technology

Transcript

  1. What we’ll learn • Beyond DevOps • Operability as a

    key focus • What is Serverless? • What is an IoT platform? • Operability for Serverless & IoT 5
  2. DevOps 7 • Infrastructure Automation vs • “Highly effective, daily

    collaboration between software developers and IT operations people to produce relevant, working systems” https://skeltonthatcher.com/blog/a-useful-working-definition-of-devops/
  3. Before “Cloud” 8 • Costly physical infrastructure • Lengthy deployments

    (months) • Mostly manual configuration • Slow feedback speed • Optimised for cost
  4. “Infrastructure as Code” 18 • Programmable via APIs • Scripts

    in version control • Testable and Test-driven • Software Engineering practices applied to web infrastructure
  5. CAMS 19 • Culture • Automation • Measurement • Sharing

    John Willis - https://blog.chef.io/2010/07/16/what-devops-means-to-me/
  6. CAMS 20 • Culture • Automation • Measurement • Sharing

    John Willis - https://blog.chef.io/2010/07/16/what-devops-means-to-me/ 25% !!!
  7. Platform Teams • Not on-call for applications • Responsible only

    for underlying platform infra • Provide the platform “as a Service” to Product teams 22
  8. Product-aligned Teams • On-call for application software • User Experience

    (UX) • Product viability • Software dev & Testing • Operational concerns 24
  9. Productivity Teams • Advise & enable Product teams • Experienced

    engineers (software, testing, operations) • Coordinate cross-team work • “Heavy lifting” 26
  10. 27

  11. Infrastructure concerns • SaaS: Software as a Service • FaaS:

    Function as a Service • PaaS: Platform as a Service • CaaS: Containers as a Service • IaaS: Infrastructure as a Service • On-premise: traditional/manual 29
  12. Infrastructure concerns • SaaS: Software as a Service • FaaS:

    Function as a Service • PaaS: Platform as a Service • CaaS: Containers as a Service • IaaS: Infrastructure as a Service • On-premise: traditional/manual 33
  13. Operability for Serverless/IoT 50 “Even if the cloud provider is

    doing everything, ... is my latency where my customers need it to be? The provider’s going to do the best they can to give me a great service, but if my customers don’t agree, then I have a problem.” -- Kelsey Hightower, @kelseyhightower https://read.acloud.guru/you-need-sre-skills-to-thrive-in-a-serverless-world-kelsey-hightower-340a002b3730
  14. Operability for Serverless/IoT 51 • UX: latency, “service” • Prevent:

    pro-active, improve • Security: perimeter explosion • Audit: data, traceability, archive • Compliance: PII, GDPR, SOX • Cost control: pay per execution
  15. User Experience 52 • Latency • “Service” “How much does

    this organisation care about my end-to-end experience?”
  16. Preventative approaches 53 • Billions of metrics • Automatic correlation

    • Automated anomaly detection • Act as ‘sensing’ for Product “It just works”
  17. 54

  18. Security 55 • Perimeter explosion (FaaS, IoT) • Auto-renewal for

    Certs • Egress detection for data • Dependency scanning “I feel safe with these people”
  19. Audit & Compliance 57 • What data do you hold?

    • Changes: where, when, who? • Archive: secure, relevant “Show me your approach to data removal for consumers”
  20. Cost control 58 • Detect run-away executions • Retire little-used

    functions • Prevent DoS → $$$ • Rapid product decisions “What is the cost of this feature?”
  21. Modern logging & tracing 3/ • Dev teams and Ops

    teams collaborate on logging details: ◦ Log messages ◦ EventID ◦ Correlation • Invest in logging (time & tools) 62
  22. Run Book dialogue sheets 2/ • Checklists for typical operational

    considerations • Team-friendly exploration around a large table • See runbooktemplate.info 64
  23. Operability for Serverless/IoT • Is infrastructure automation the best approach

    for your org? • Invest in areas that differentiate your organisation • Make systems operate well 66
  24. Further reading Team Guide to Software Operability Matthew Skelton &

    Rob Thatcher skeltonthatcher.com/publications Download a free sample chapter 68