Stateful Sessions for Intelligent Apps
summary
Demonstrates a practical way to scale real‑time audio transcription by keeping per‑call audio state in the container and using OpenShift sticky sessions so cookies bind a phone line to a specific pod. Uses a Kafka kappa architecture to stream decoded text to a sentiment notebook, runs six call simulators in the demo, and shows live results; it works end‑to‑end but Jupyter notebook processing becomes the bottleneck rather than the session routing.
AI-written summary · about
chapters (12)
- 0:00Introduction
- 0:21What is Stateful
- 1:17Puzzle Builder
- 3:15RealTime Call Center
- 5:47Open Data Hub
- 7:52Phone Lines
- 8:30Autoscaling
- 9:56Sticky Sessions
- 10:59Saving Cookies
- 11:27Demo
- 11:57Openshift Demo
- 15:12Recap
related talks
01h 23mIntelligent applications in containers
30mContainer workflows for data science and ML
26mRethinking Cloud-Native Patterns on Kubernetes
54mExploring Stateful Microservices built with Open Liberty in Kubernetes
41mMicroservices and Event Sourcing with Kafka and Kubernetes
53mEffective Kubernetes for Cloud Native Java Developers
27mPragmatic Pod Patterns: Leveraging sidecar containers in Kubernetes
47mGoing from containers, to pods, to Kubernetes – help for developer environments!