Distributed applications and Kubernetes: Frameworks or service meshes or both
summary
Compares Java frameworks (Spring Boot/Spring Cloud with Eureka and Spring Cloud Gateway, Micronaut, Quarkus using Resilience4J/SmallRye) that embed service discovery and circuit breakers versus Kubernetes features (Services, ConfigMaps, scaling) and extensions: per-pod sidecar proxy service meshes that enable percentage-based canary routing and observability at the cost of added hops and measurable latency, and node-level eBPF on nodes (Cilium) that attaches to worker nodes to capture/control traffic with lower overhead while still requiring a control plane.
AI-written summary · about
chapters (8)
- 0:00Introduction
- 3:09Why build distributed apps
- 10:33Distributed app frameworks
- 17:34Kubernetes for systems
- 22:35Service meshes and sidecars
- 29:59eBPF as Kubernetes extension
- 35:17Choosing the right approach
- 36:22Q&A
related talks
34mDistributed applications and Kubernetes
48mWeaving Through the Mesh: Making Sense of Istio and Overlapping Technologies
39mDo you need a Service Mesh?
46mExperiences from using discovery services in a microservice landscape
27mMicroservices: Awareness and Discovery Pattern on K8s
56mService Meshes - Powering the next wave of microservice architectures
05mService Mesh Architectures Explained - Sidecar and Beyond
40mReimagining Application Networking and Security
