How to Connect the Non-Connectible: Tips, Tricks and Tears
summary
Shows concrete connectivity patterns and failure modes for apps spread across GCP, AWS and on‑premise: recommends using a transit‑style hub (AWS Transit Gateway / GCP VPC‑Transit) instead of many VPC peerings, and demonstrates differences between VPN tunnels and Direct‑Connect equivalents (order of tunnel setup, route limits, cross‑project CIDR blocking). Calls out DNS and internal load‑balancer pitfalls (GCP global internal LB beta, DNS forwarding issues), and discusses runtime/auth details including a claimed large performance gap between containerd and dockerd and complications with IAM/workload‑identity mappings plus partial workarounds.
AI-written summary · about
chapters (12)
- 0:00Cloud technologies
- 3:27Multicloud growth
- 6:20Containers for portability
- 8:50Network challenges
- 13:22Transit networks
- 16:42DNS and routing
- 19:49Private resource access
- 23:02Internal routing and load balancing
- 26:14Authorization and security
- 30:19Cluster management
- 33:29Service connection automation
- 35:17Conclusion & Q&A
related talks
25mNetDevOps: Applying lessons of DevOps to networking
01h 03mConnect remotely like a pro! Hack VPNs, avoid firewalls, connect even when your machine is down
38mThe Reality of Porting Kubernetes Apps Across Cloud Providers
45mMultiCloud
32mDive into technical decisions from successful migrations
20mHelp Protect Your Data Centers with Safety Constraints
05mGCP vs AWS - Our head in the clouds
41mKubernetes the Very Hard Way