Blog

Super organized, Kubernetes focused—Thank You, CNCF!

It might seem counterintuitive, but when you’re starting a company it’s good to have competition. If there’s no one in your niche… well, it might be because no one needs the particular service or product you offer.

In line with that reasoning, seeing our competitors at KubeCon was one of the highlights of the conference. There’s a growing feeling among companies of all sizes who attended the conference that building and managing Kubernetes infrastructure in-house is not a winning business strategy—it’s too risky, labor intensive, expensive and slow. I talked to people from Kaiser Permanente, The Home Depot and Bloomberg over the course of the conference — each at various maturity in their Kubernetes journey. Yet every single one agrees that attracting top talents is a top concern. Other companies debated whether an in-house container management process or orchestration would provide a business advantage. I urged them to think about what the opportunity costs are if they are wrong. One thing was clear: These companies would not have sent representatives to KubeCon if they weren’t already working with containers.

Agile Stacks’ conference team included both of our co-founders, John Mathon and Igor Mameshin, Anton Kranga, our Director of Engineering, Rick Richardson, Principal Engineer, Gene Saxe, Account Executive and myself. We all felt at home because of the interest we got from people who visited our booth. We enjoyed a myriad of conversations with IT leads, developers, ops engineers, a competitor or two, complimentary vendors, press and analysts.

Here were some of my takeaways.

The CNCF is Super Organized

Ok, perhaps this should not be surprising, but I’ve attended my fair share of conferences and they do not always run smoothly. I was very impressed by how well organized registration was and everything thereafter.

If you have never been to a KubeCon, consider it—not only is it good from an educational / business perspective, but it’s incredibly well-run allowing you to focus on learning and making connections, not figuring out why your registration was lost.

Everyone Needs Kubernetes

It simply is not feasible to manage a complicated Kubernetes infrastructure without additional tooling—that was clear. But each company has unique requirements. And analyzing Agile Stacks' product fit requires looking at the target application’s lifecycle. Projects that have already been deployed and essentially in maintenance mode — especially true at large corporations — are not great candidates for a new infrastructure management platform. Larger companies tended to be interested in learning how Agile Stacks could support net, new projects where architecture decisions can start anew and be done iteratively at low cost.

On the other hand, many start-ups and smaller companies were interested in implementing Agile Stacks immediately, for all of their applications and microservices. This likely reflects the fact that smaller companies have more limited resources to manage infrastructure in-house, are more agile in general and are more concerned about fast scaling as they grow.

Too this point, our machine learning stack powered by KubeFlow received a lot of attention at the booth. Showing our visitors the list of ML services got the conversation started and frequently led to discussions about specific use cases and practical next steps. Certainly top of mind were the challenges of, first, getting Kubernetes setup and, second, how to get all ML tools and components to work seamlessly in a pipeline. We believe the next generation of data scientists will never have to wonder whether infrastructure and DevOps will be able to keep up; the assumption will be that integration and scaling is automatic. Hence, companies both big and small get a level playing field when it comes to powering their AI infrastructure now that Agile Stacks takes complexity out of the equation. The main focus will shift from infrastructure woes to ground-breaking data and machine learning models.

There’s Room for Collaboration

One of the most exciting things about KubeCon was seeing not just our competitors, but also other companies that produce complementary products and services. I was particularly interested in talking with representatives from Containous, Nutanix, Microsoft Azure, Portworx and Armory.io about possible future collaboration between our companies.

Kubernetes is Here to Stay

While specific data about Kubernetes adoption rates both varies by source and is inherently out-of-date already, it is clear that Kubernetes is far-and-away the most popular container orchestration tool with somewhere around 60% of companies using it. Docker, Rancher, and Mesosphere used to have competitive offers — now they have both pivoted to offer complementary products and promote the fact that they offer integration with Kubernetes. All of the major cloud providers offer services specifically tailored to Kubernetes environments. Most AI applications are developed using containers to make it easier to update and scale each step in an ML workflow, and Kubernetes is an ideal platform for that scenario.

KubeCon Redux

Having only been our second officially sponsored event since coming out of stealth earlier last year, KubeCon made such an impression with our founders that we will once again be at KubeCon in 2019. The venue for 2019 will be San Diego, the location for one of our engineering hubs. As Igor Mameshin, CTO, put it, “Our team had such a great time at KubeCon and we appreciate all those who stopped by our booth to have a conversation, ask a question, or to simply say hi. We also enjoyed meeting many of our partners. We even hired several engineers whom we met at KubeCon. It’s safe to say that our first KubeCon + CloudNativeCon was a success, and we hope to see everyone again in San Diego!

 

Founders Igor Mameshin, CTO, and John Mathon, CEO, taking a break for a photo opp wearing their new Agile Stacks gear.

KubeCon 2018

Topics: Partners, Marketing

Subscribe to Email Updates

Recent Posts

Posts by Topic

see all