Real-Time Data Anonymization the Serverless Way
Talkat KubeCon + CloudNativeCon North America 2021 · on Thursday, 2021-10-14· at 11:55· Concourse Hall 151 + Online·
summary
Architecture that anonymizes images on upload by extending Ceph with a Ceph SQS-compatible API so KEDA can pull notifications directly from object storage and avoid exposing HTTP endpoints. Rook provisions the Ceph S3 buckets and bucket-notifications; Redis stores notifications with visibility timeouts and retention; a KEDA external scaler spins worker functions which download images, run Haar cascades for faces and a TensorFlow model for license plates, blur regions, and overwrite objects. Runs on MicroShift and can operate on a single-node (2 CPU, 2 GB) edge cluster.
AI-written summary · about
chapters (14)
- 0:00Introduction
- 0:21Agenda
- 1:13Data Privacy
- 3:37Solution Architecture
- 6:03Building Blocks
- 7:33Rook
- 12:45Why do we need something else
- 14:09Summary
- 15:31Message Queue
- 17:25External Message Queue
- 18:12CNF Service Frameworks
- 20:50CATA Service Framework
- 22:18Object Detection
- 24:08Microshift
related talks
35mKEDA - Real Time And Serverless Scaling In Kubernetes
36mKEDA: Event Driven and Serverless Containers in Kubernetes
50mDeveloping Highly Scalable Image Storage Solution with AWS Serverless
46mHighly Scalable Image Storage with AWS Serverless
29mServerless on Kubernetes: What We Learned From Customers
37mKubernetes Event-driven Autoscaling with KEDA
37mKEDA: Unlocking Advanced Event-Driven Scaling for Kubernetes
37mReinventing Event Photography: Transform Community Moments