Podcast
Episode 2
Microservices and Orchestration with Kubernetes and Docker
This month we’re talking with Shane Castle and Gui Martins about Kubernetes: how it’s enabling the effective orchestration and management of cloud-first microservices and architectures like service mesh, and what that means for your business.
Listen to the Episode
Listen to the Episode
- 0:30 – Jeff Zado introduces the show and the topic of Serverless (PaaS) vs Microservices
- 2:30 – Dave Judd introduces himself
- 2:55 – Dave discusses what is meant by “serverless”
- 3:30 – Dave discusses serverless vs. microservices / Kubernetes and when you might choose one architecture over another
- 6:30 – Dave discusses what’s motivating a transition to serverless, including multiple language support and consumption based pricing
- 7:00 – Dave discusses how the three major cloud vendors (Microsoft Azure, Google Cloud, AWS) are following similar patterns
- 8:00 – Dave talks about auto-scaling on serverless platforms and global high availability that comes with PaaS
- 10:00 – Dave compares Serverless Functions to Microservices
- 13:20 – Dave talks about how serverless allows you to model your systems to how things actually behave in the real-world, with event-driven / reactive-based architecture
- 14:00 – Dave discusses serverless in Azure, specifically Functions, Event Grid (and the Cloud Event Schema), and Cosmos DB
- 16:30 – Dave talks about managing state in a serverless world
- 18:00 – Dave, Jeff and Nick talk about fanning out thousands of functions to do massive processing in parallel, scaling just the pieces that are needed at a given time as opposed to scaling an entire application with a serverless orchestrator, granular tweaking of performance and cost
- 20:45 – Jeff and Dave talk about performance improvements in serverless runtimes and relative speeds in spinning up functions
- 22:00 – Dave talks about vendor lock-in and whether its potentially illusory in some cases
- 23:00 – Nick and Dave talk about how serverless also enables cross-language collaboration on developer teams and the benefits for engineering managers
- 24:30 – Dave compares serverless offerings on Azure to AWS and Google Cloud
- 25:30 – Dave talks about how serverless can be used not only for new greenfield applications but also to augment older legacy or brownfield monolithic application
- 27:30 – Dave discusses some examples of implementing serverless in the real world, specifically in the fintech sector
- 30:32 – Jeff talks about how the example of the implementation for the fintech company enabled innovation with even better profitability and more calculated costs
- 32:50 – Nick asks Dave to provide a reality check and Dave discusses some issues to consider when moving to a serverless architecture – because everything’s a function, there’s more to manage and consider re deployment of functions that work as a group, versioning and security
- 35:28 – Dave also talks about observability, logging, metrics and analytics and how important that is in a serverless architecture to ensure application resiliency and debug issues that might arise