The future is Serveless | Filip Knapik | #4 Serverless UG Warsaw
•
2 likes•225 views
Filip Kanpik - Product Manager @ Google Cloud Platform raised a topic related to the challenges and opportunities of the servaste architecture based on Google Cloud. The presentation will also cover open source solutions and launch of Serverless services on Kubernetes clusters using Knative; also outside of Google Cloud. We will talk about issues related to the efficiency of Serverless solutions in the public cloud and the direction of development of the entire area.
Recommended
Recommended
More Related Content
What's hot
What's hot (20)
Similar to The future is Serveless | Filip Knapik | #4 Serverless UG Warsaw
Similar to The future is Serveless | Filip Knapik | #4 Serverless UG Warsaw (20)
More from Serverless User Group Poland
More from Serverless User Group Poland (8)
Recently uploaded
Recently uploaded (20)
The future is Serveless | Filip Knapik | #4 Serverless UG Warsaw
- 1. The future is serverless Filip Knapik Product Manager, Serverless
- 2. Principles of serverless Invisible, auto-scaling infrastructure No paying for unused infrastructure$
- 3. Serverless features Operational model Fully managed security Pay only for usage No ops, Autoscaling Event-drivenFocus on code Stateless Programming model {}
- 4. Serverless isn’t just compute Compute AI Data Processing MessagingDatabase Data warehouse Storage
- 5. Serverless on GCP Cloud Dataflow Cloud Machine Learning Cloud Storage Cloud PubSub Cloud Functions App Engine BigQuery Cloud Datastore
- 6. What’s so hard about traditional app development? User Select IaaS OS Setup Containerization Storage Container Orchestration Networking Monitoring Logging Testing Dev
- 7. Highly customizable Highly managed HTTP or Async Functions High scalability for apps Zero server management Multi Container- based system Run on VMs Used for Lift & Shift Compute Engine Kubernetes Engine App Engine Cloud Functions Serverless Re-platform once; run anywhere Connect and extend cloud services Compute continuum Serverless Containers GKE Serverless Add-on Serverless abstraction for GKE Can run with GKE on-prem Fully managed Build, deploy and serve containers
- 8. Functions
- 9. FaaS characteristics Simple Auto-scales with usage Run code in response to events Pay only while your code runs
- 10. FaaS Use Cases APIs Events Data and light ETL Webhooks Device/iOT analytics e e e e Web/mobile client API calls (FE queries, client UX) Data change events Log event actions 1P/3P data ingestion and lightweight ETL Client behaviour event triggers, analytics
- 11. Cloud event triggers Cloud Functions Cloud Vision API Cloud Functions Cloud Storage Cloud Storage
- 12. Pros and cons of FaaS Focus on code, not infrastructure Flexible cost Very scalable Enforces microservice architecture Event-based Not all workloads can use FaaS abstraction Not always portable
- 13. Containers
- 14. Highly customizable Highly managed HTTP or Async Functions High scalability for apps Zero server management Multi Container- based system Run on VMs Used for Lift & Shift Compute Engine Kubernetes Engine App Engine Cloud Functions Serverless Re-platform once; run anywhere Connect and extend cloud services Serverless & containers Serverless Containers GKE Serverless Add-on Serverless abstraction for GKE Can run with GKE on-prem Fully managed Build, deploy and serve containers
- 15. Serverless containers Event/Response Fully managed Stateless Pay for use Sign up: g.co/serverlesscontainers Serverless Containers (Alpha)
- 16. Serverless add-on 1-Step deploy Run serverless workloads in your GKE Cluster Auto-Scale Sign up: g.co/serverlessaddon
- 17. Serverless future FaaS + Containers + Apps Event-based architecture No-ops as a preferred model Everyone needs scalability Portability is a must