Lin Sun – The Crazy Service Mesh Ecosystem (Introductory)
Ever wondering how the crazy service mesh ecosystem explodes in the past few years? Why suddenly everyone is jumping onto the service mesh bandwagon? What are the differences between these key players such as Netflix OSS, Istio, Linked and Ramp etc. This talk will provide an overview of why service mesh, the crazy service mesh ecosystem, where we are, and live demos and comparison of what is out there.
This talk will focus on the following:
- What are the key players in this space?
- What is Istio and the takeaway of the latest Istio 1.x release?
- What is Linkerd 2.0 and key takeaway of Linkerd 2.x?
- Wait…What about AppMesh?
Sam Handler – DevOps As Engineer Empowerment (Introductory)
We’ll start with a familiar situation: the “DevOps team” has finally finished migrating a “legacy” deployment pipeline to a buzzworthy new containerized application delivery platform. Backs are slapped, champagne poured. Then, a few weeks later, an engineer deploys while the DevOps team is at lunch. The site crashes — the app can’t boot without a new environment variable. Alerts fire. The engineer remembers something about using a CLI for environment variables. But wait, what’s a kubeconfig? Fifteen minutes pass. The DevOps lead runs into the room panting, a slice of pizza still in hand. The engineer spots him and tells him what the application needs. The lead unlocks his computer and runs a command to get the site back up. Phew.
This outage is just one example of how DevOps initiatives can actually end up disempowering engineers, while DevOps ought to be doing the opposite. Fortunately, if you center your DevOps practice around engineer empowerment, you can avoid sticky situations and calmly enjoy your pizza. You’ll reap lots of rewards besides uptime! Here’s what to do:
- Educate engineers about the infrastructure their application runs on and the mechanics of a deployment.
- Give engineers appropriate privileges. Architect your systems such that they have access to what they need without endangering the health of the platform.
- Make engineers responsible for the health of the platform, and give them the tools necessary to meet this responsibility.
- Provide all engineers with an understandable view into the health and performance of the platform.
- Supply tools and training for engineers to measure performance characteristics from their code.
- Equip the team with production-like playground environments.