What AWS re:Invent Taught Me About Love / Hate Relationships

I had the opportunity to attend AWS’s 2023 re:Invent conference for the first time, alongside my teammates here at Rafay. Our CPO, Mohan Atreya, posted a great wrap-up blog a few days ago that I highly recommend reading, as it summarizes our experiences well and includes all the announcements and product updates we released.

Attending these kinds of events is always an incredibly clarifying experience for me, because it helps me to get outside our marketing bubble a bit (we product marketers are often too good at creating our own Reality Distortion Fields) and check my perceptions against the actual experiences of platform teams, our target market. One of the many reasons I love staffing booths at events like these is because it allows me to collect a wide variety of perspectives very quickly. And this year at the booth, we had an excellent tool at our disposal to collect those perspectives.

I’m speaking, of course, of our swag.

Marketers know that a good swag item at a conference doesn’t just attract attention to your company, it also helps your people drive the right conversations with the people it attracts. It’s equal parts shibboleth, discovery tool, and calling card. And I’m proud to say our marketing team knocked it out of the park with this t-shirt design. I mean, look at it:

“What, that’s it? A catchphrase in a hacker font?” you’re thinking. Or, “0h, h0w 0r161n4l.” But, I’m telling you, having these shirts out front immediately told us who our people were … and our people were loving them. We opened the show with 500 of these shirts in our lowly 10’ by 20’ booth, and they were gone within hours. We ordered 10,000 more overnight (not actually that many, but it felt that way as we were folding and rolling them) and they were completely gone by the end of the show.

Why? Because the “don’t hate k8s” phrase resonated with people, and they would immediately tell us about their relationship (love, hate, or otherwise) with Kubernetes when they inquired about it. That, in turn, helped us immediately understand their experience, expectations, and needs.

Generally speaking, the people we spoke to fell into one of three camps:

K8s is great! (But it could be better.)

A great many organizations have standardized their container operations around Kubernetes. A great many of THOSE orgs are happy with their choice of EKS as their preferred managed service to deliver it, and we spoke to quite a few of them last week!

While the benefits of running Kubernetes in a public cloud platform seem obvious, we heard that many EKS users are still working hard to deliver the dream of truly seamless delivery of Kubernetes services to the people who depend on them. The time needed to build access control, add-on management, policy enforcement, compliance checks, and multi-tenancy with team isolation can be considerable … and that’s before you even get to creating the standard templates that developers need in order to provision the infrastructure they need at the push of a button.

Rafay has been the leader in empowering platform teams to deliver cluster-as-a-service (CaaS) and namespace-as-a-service (NaaS) for years with Kubernetes Manager. NaaS in particular, when properly configured, can be a game changer for enterprise organizations, because it allows them to keep cloud costs low without sacrificing the security and workload isolation provided by dedicated clusters. But both CaaS and NaaS benefit from Rafay’s built in blueprint templates, drift detection, policy management, RBAC management, and more.

I hate K8s, this shirt is dumb (Also, can I have one?)

We also met many platform teams and infrastructure engineers who are either living in a post-K8s world, or trying their best to avoid it entirely and skip to the next thing. And a few of those folks definitely had some feels about it.

Listen, I get it. Kubernetes was supposed to solve many of the challenges of existing cloud infrastructure, such as orchestrating complicated deployments, automatic scaling and load balancing, application portability, and a declarative configuration approach. But modern applications need more than just a container environment to function – they need connections to databases, object stores, and other services too. And platform teams want to give developers golden paths to complete cloud environments and application landing zones quickly, for dev, test, or prod, without losing control over how they’re configured or how much they cost.

Earlier this year, Rafay introduced Environment Manager specifically to help platform teams deliver cloud environments and landing zones as a service for developers. In addition to providing access controls, integrating with DevOps toolchains, and tying into Infrastructure as Code repositories like Terraform as its source of truth, Environment Manager can automate the complete deployment pipeline of an environment, and maintain persistent visibility over the environment once it’s deployed.

What is K8s, anyway? (I’ll take a shirt too … for a friend … )

I loved meeting these folks, because at the end of the day, helping them is what we are all here for. Also, how in the heck does ‘ubernete’ shorten to the number 8? It makes no sense. Who decided this? …What was I talking about

When it comes down to it, whether you’re creating cloud applications or using them to get your work done, you shouldn’t have to care that there is something called ‘Kubernetes’, ‘ECS’, or any other technology making it all work. Most roles living outside IT shouldn’t have to worry about this stuff – even DevOps engineers would often like to scale back how much they have to worry about it. The most recent roles to join this group are data scientists and data engineers, both of whom were in attendance at re:Invent 2023 in significant numbers.

Companies in every industry have an urgent need to jump into the opportunities presented by AI and ML applications, but not every industry has kept up with the latest developments and technologies that make those applications work. But no matter their technological maturity, Rafay’s templates for AI development (including development of generative AI applications) are designed to give these teams a rapid start so they can provision an AI workbench, connect to services like GenAI large language models (in AWS Bedrock or elsewhere), build their code, and GO.

Which group do you fall into?

Drop us a line and let us know how we can help you! And if you’re interested in one of those fly t-shirts (or weren’t able to get one in your size when you visited us at re:Invent), sign up for a demo here and we’d be happy to send you one.