In a recent episode of the Continuous Discussions (#c9d9) podcast, we hosted Gene Kim and several speakers from the upcoming DevOps Enterprise Summit Las Vegas (DOES18) conference. We’re proud Founding Partners of the event and always enjoy having IT thought leaders join us ahead of the show to talk about DevOps, the conference program and more. In this episode, we discussed an emerging conference theme of next-gen operations and infrastructure.
The panel of DOES18 speakers included: Cornelia Davis, senior director of technology at Pivotal; Damon Edwards, co-founder and Chief Product Officer of Rundeck; Mik Kersten , CEO of Tasktop; Tom Limoncelli , SRE at StackOverflow.com; Gene Kim , founder of IT Revolution and co-author of “The DevOps Handbook,” “The Phoenix Project” and “Accelerate”; and CloudBees’s Sam Fell and Anders Wallgren.
Continue reading for some of the top takeaways from the discussion!
Programming and Insights
As we continue to evolve, the best is yet to come for operations teams, per Kim : “If we follow the course of what the development languages and development frameworks have done for development productivity, and extend that to operations, I think the best time for operations and infrastructure and platforms are to come.”
Davis encouraged operations teams to treat their platform like a product instead of infrastructure: “When I speak with the IT operations folks, they’re still largely short order cooks. They’re very reactive, doing things that people need them to do. It’s about being able to apply some of the same principles of continuous delivery to the way that you do ops. At the core of that is this philosophical notion, which is to treat platform just like you treat any other product.”
The way enterprise operations teams do work is unique, per Edwards : “For enterprise operations, you have to fundamentally rethink the conventional wisdom of how you think about, approach and do operations work. DevOps Enterprise Summit started in the early days of the DevOps explosion. This is distinctly an enterprise discussion because it’s all about that history of success that we have to account for.”
Kersten on the different ways development and operations teams measure outcomes: “I so often see the development side of the organization and the infrastructure operations side collecting completely different metrics and telemetry. The development side uses the more familiar telemetry that comes from DevOps. And the operations side is still in a burnup, burndown world. There are no common metrics or languages for what really should be an internal value stream.”
Great operations management advice from Limoncelli : “Get one team to do it well, and let people copy that success. And do it in a way that creates cultural minimums or cultural ways of doing things, so that people are competing to make things better instead of being told you have to make this change.”
Moving away from legacy operations is a dauting challenge especially for the enterprise, explained Wallgren : “The things that we refer to as legacy are the things that keep the lights on and bring the money in. It’s a big leap to move away from that and it’s going to take a lot of work and quite a lot of time. It’s like changing the engine and tires on a car while you’re driving down the highway at speed.”
Functional Programming, SREs and More Musings
Davis on functional programming, something she will be talking in-depth about at DOES18: “Treating function as a first-class citizen is absolutely essential because you start to create these functions very dynamically, and the functions now encapsulate the state. So instead of having state and then having side effects against a state, what we have are these functions encapsulating a sequence of changes that are going to generate a state.”
Kim described an “aha” moment he had during Davis’ talk at DOES18 London about sidecars in Kubernetes: “The ability to liberate developers and address cross-cutting concerns like logging was one of the primary impetuses why people flocked to Spring from orthodox frameworks that were so unsatisfying. And that was, for me, that was one of the biggest aha moments of the conference.”
Edwards discussed the implications of toil work and how to spend more time doing things of value: “Toil work is work that we should have automated, it’s repeatable, and not adding enduring value to what we’re doing. Versus work that requires human ingenuity, the engineering work that adds enduring value to our organization. Set a limit as an organization to not accept more than 50 percent of time being consumed by toil and figure out how to help those organizations where they have too much toil and not enough engineering work.”
Most organizations don’t have $2.5 billion in their back pocket to build a product-oriented value stream, but Kersten said we should take a page from their book: “This is a testament to how invested the tech giants are in their own platform as products. If we don’t at least adopt their thinking around having this connected end to end platform that supports all your products and value streams, you’re just not on the right path.”
Limoncelli on the ability of SRE’s to reach across silos: “In the SRE model, you can reach across silos. I like to think of it as you have the CICD pipeline and the SRE’s stand at the end of the pipeline asking, ‘What are the operational problems that we’re having? What’s the constraint? What’s the bottleneck?’ They fix that by asking how they can collaborate with the other team within our organization. And either sit with them for a week and fix it or collaborate with them to fix the problem.”
Wallgren elaborated on treating pipelines as code: “Treating your pipeline as code is something we’ve been evangelizing. And really, that should be the first thing that you really worry about in some sense, because you’re not going to ship anything unless you have a pipeline.”
The DOES18 Las Vegas agenda is now live ! Check out the sessions from Davis, Edwards, Kersten, Limoncelli, Wallgren and dozens more. Tickets to DOES18 are running low, so be sure to register before it’s too late.
Watch the full episode:
Originally published on DevOps.com .
Stay up to date
We'll never share your email address and you can opt out at any time, we promise.