Favorite talks from Google Cloud Next ’17

This year Google Cloud Next 2017 was a constant source of amusement for me. I just wanted to reflect on several talks I watched that you might find interesting if you are into DevOps and SRE topics.

DevOps/Kubernetes

• Networking in Google Container Engine and Kubernetes: Mind blowing talk by Kubernetes engineers Tim Hockin and Michael Rubin explaining lifetime of a TCP packet in a Kubernetes cluster. This talk goes in detail about the Kubernetes and GKE networking such as routing, load balancing and iptables. Definitely watch if you are interested in container networking.
• Globally scalable microservices with Google Container Engine: Sandeep Dinesh presents going from a single Kubernetes cluster to a multi-region federated cluster using kubefed, then shows how to globally scale a service using Google Cloud Load Balancer and single-IP anycast on Kubernetes. Check his blog posts for a more in-depth tutorial on this.
• Spinnaker: continuous delivery from first principles to production: This talk starts with great storytelling about a Waze outage, continues with deployment strategies for real-world microservices and shows a great demo of how they are using Spinnaker, an open source continous delivery tool by Netflix.
• Container management and deployment: from development to production: Great talk by Kelsey Hightower on end to end development and deployment experience on Google Container Engine and Cloud Container Builder.

Reliability

• Ten common causes of downtime and how to avoid them: Great talk on how you should be thinking of designing your infrastructure for high availability and scale.
• Metrics that matter: A short talk by the founder Google’s SRE team Ben Treynor about measuring the right metrics for user-facing services.
• Designing reliable systems with cloud infrastructure: If you are new to the SRE mindset and have not read the free SRE book check this talk out. Paul Newson talks about how to architect your infrastructure on Google Cloud to achieve higher availability. Great walkthrough of capacity planning for an example distributed service and illustrations of traffic flow during various failure modes.

More

You can find videos for all 216 talks from the conference on YouTube.