This is a survey paper that explores six Cloud-based deployment archetypes for Cloud applications and the tradeoffs between them to achieve high availability, low end-user latency, and acceptable costs.

00:00:13 Abstract

00:01:29 1. Introduction

00:04:28 1.1 Principles of Availability

00:09:32 1.2 Types of Applications

00:11:03 1.3 Data Durability, Availability and Backup

00:14:02 1.4 Six Deployment Archetypes for Cloud Applications

00:18:02 2. Zonal

00:18:36 2.1 Single Zone

00:20:11 2.2 Primary Zone with Failover Zone

00:26:47 3. Regional

00:28:13 3.1 Single Region

00:32:56 3.2 Primary Region with Failover Region

00:37:21 4. Multi-Regional

00:40:38 4.2 DNS Load Balancing

00:45:30 4.3 DNS Load Balancing with Isolated Stacks

00:52:45 4.4 DNS LB with Custom Multi-Regional Load Balancing

00:54:22 5. Global

00:56:15 5.1 Global Anycast

00:58:36 5.2 Global Anycast LB with Isolated Regional Stacks

01:00:54 5.3 Global Services Stack

01:08:13 6. Hybrid

01:11:36 7. Multi-Cloud

01:19:21 8. Comparison

01:22:58 8.1 Failover-to-Standby vs. Load Balancing

01:25:54 8.2 Regional vs. Global Application Stacks

01:28:14 8.3 Instantaneous Recovery

01:29:30 8.4 Cost of Availability

01:32:34 9. Related Work

01:38:08 10. Summary

01:39:22 Acknowledgements