Antithesis of dev ops
Download as PPTX, PDF1 like425 views
The document discusses the impact of COVID-19 on DevOps practices, highlighting how remote work and social distancing have disrupted the core tenet of close collaboration. It emphasizes the importance of creating reusable pipelines and automating processes to maintain efficiency in software development. Additionally, it draws parallels between modern workflow challenges and historical solutions like assembly lines, underscoring the need for systematic communication in DevOps workflows.
Antithesis of dev ops
- 1. Covid-19 The antithesis of DevOps https://xkcd.com/2282
- 2. Avi Cavale VP of Engineering at JFrog @avinci
- 3. COrona VIrus Disease 2019 An infectious disease caused by a newly discovered strain of coronavirus
- 4. A much harder question, What is DevOps? A set of practices that brings Dev & Ops together with the intent of shortening SDLC , improving software quality, and providing CD focusing on a “Culture of Collaboration”
- 5. Focusing on culture and breaking down silos is a good thing dev ops • build things • inspired by creativity • make things reliable • inspired by stability
- 6. Benefit seen is rarely > 15% improvement Dev Test Ops 2 Pizza Teams Functional Silos larger larger Agility costIntegration cost
- 7. Workflows evolve in the way organizations are structured Conway’s Law
- 8. Do these workflows communicate stochastically or systematically?
- 9. The DevOps emphasis on culture & bringing devs and ops closer together has forced all communication to be stochastic SCRUM Meetings Chat Rooms Hallway Conversatio ns Do you have a minute? Email Messages
- 10. Now, Covid-19 has forced a huge change • Socially distancing is here to stay • WFH has become the new normal • Work schedules have become totally random Core tenet of DevOps - “collaborate closely” - has been turned upside down…
- 11. Interestingly, more than a century ago, a visionary figured out how to increase throughput… he created “Assembly Lines”
- 12. parts are added to create semi- finished components upstream components can become parts for additional semi-finished components …and so on until the final “shippable” assembly is reached
- 13. Ops sets-up test infrastructure with VPC, Kube cluster, Load- balancers etc. Test infra info MS1 Imag e MS2 Imag e MS1 Dev team takes base image and create MS1 image with CI tests MS2 Dev team takes base image and create MS2 image with CI tests Test team takes MS1and MS2 deploys to test and run sys and perf tests and promotes MS1 and MS2 if all OK Test ed MS1 Imag e Test ed MS2 Imag e Ops sets-up prod infrastructure with VPC, Kube cluster, Load-balancers etc. Prod infra info Deployment team deploys MS1 and MS2 to prod Prod rel’s e info DevSecOps creates security hardened base docker images for microservices Base Imag e Pipeline stat e A connected enterprise, aka “Software Supply Chain”
- 14. is the workflow that is needed to produce a component from other input components Pipeline stat e is the output of a pipeline (aka “resource”), which could either be a finished component, or a semi-finished component used as an input to other pipelines Resources are the bedrock of creating systematic communication in DevOps workflows
- 15. How to get the monkey off your back • Create reusable pipelines for different parts of your SDLC and modularize it • Choose a pipeline tool that has inbuilt functionality for state or resources • Whatever you get pinged for, create a resource for it • Automate the heck out of your process
- 16. Thank you for listening Avi Cavale avic@jfrog.com
Editor's Notes
- #13: The beauty of this is that downstream components don’t care when, where and who created the dependency parts