Cloud Native / Containers / Kubernetes

Container-Native Application Development in 2018, an Oracle Executive’s Take

26 Dec 2017 9:44am, by

Bob Quillin
As vice president of software development for the Oracle Container Group, Bob Quillin is responsible for the products and team from the StackEngine acquisition by Oracle in December 2015. Based in Austin, the group develops container-based services designed to help developers and DevOps teams build, orchestrate, and scale enterprise-grade container apps. In 2016, they launched Oracle Container Cloud Service (OCCS), a container management service (CaaS) for building and running Docker applications on Oracle Cloud built from the StackEngine technology.

It’s been a wild and eventful 2017 in the container-native world. Kubernetes has effectively won the container management and orchestration wars, open source software has gained wide acceptance across organizations, and cloud providers (at least most of them) have answered the call for more open and cloud neutral technologies. What lies in store for 2018 builds on all this industry standardization and progress.

Enterprise developer teams will gain valuable flexibility and choice after several years of single cloud vendor lock-in struggles, as these open source technologies become pervasive across cloud providers. What comes next is what you might expect as container technologies mature, enterprises become more selective based on these new choices, and microservice and container technologies move up the stack.

After reflecting on the excitement and progress of 2017, I’ve taken a look at the unbridled opportunities that await all of us in container-native application development (or is that Kubernative? — read on), and offer a few predictions for 2018!

1. Managed Services Replace Open Source DIY

This past year has seen a huge increase in cloud-native open source commitment and adoption led by the Cloud Native Computing Foundation. In 2018, enterprises will continue to accelerate open source commitments but will face a new challenge — how to digest open source software faster and more comfortably (something we all might face during the holiday season!). There’s a huge risk of continuous indigestion/continuous digestion (the bad kind of CI/CD) as enterprises assimilate these technologies.

So, how will the enterprise digest open source faster? By turning to managed services based on open source software — and stopping the Do It Yourself projects! The year 2017 will be the last year for the “I Stood Up My Own K8s” trophies. There’s no reason in 2018 to spend valuable developer and DevOps time running and maintaining your own open source platforms.

2. Integration Beats Disintegration

2017 has continued to see a tidal wave of discrete and disjoint cloud services rolled out for development teams — to the detriment of an integrated, coherent developer experience.  “Disintegration” is the “process of losing cohesion or strength” or “the process of coming to pieces” — which is the problem developer teams face when assembling these cloud services into a cohesive whole.  As an analogy, many cloud providers have been just a huge grocery store full of ingredients — but they left it up to you to shop, prep, cook, and clean up afterwards — when all you really want to do is eat. Yes, it’s fun to cook delicious and memorable meals yourself every once in a while — just not all your meals, every day, every week, 365 days. In 2018, enterprises will turn to cloud providers for integrated experiences — a full stack meal that includes wine pairing and cleans up after itself, too!

3. Quality Over Quantity — a Year of the “ilities”

If all cloud providers offer the same basic set of open source based services (e.g., Docker, K8s) — all priced for just the IaaS resources you use, then how do you choose? For many enterprises in 2018, the swing factors will come down to the “ilities” like scalability, security, availability, reliability, and usability. Often that can be described as “enterprise-grade” or we’ve heard it described as “open source for grownups.”

It’s akin to graduating from college and starting to appreciate the good stuff — be it the food you eat, the wine or beverage you drink, the car you drive, or the clothes you wear. Open source can be free and fun, but when you need to run your enterprise apps on it, you’ll want to consider the good stuff — and that’s where the “ilities” come in. In 2018, quality comes forward.

4. Container Management Moves Up Stack

If KubeCon/CloudNativeCon 2017 is any indication, 2018 will be the year we move on from pods, kubectl’s, and masters and on to open service brokers, service meshes, and circuit breakers. Up the stack, we go with projects like Istio leading the way. Enterprises will leverage patterns from early, scale-out microservice practitioners and apply a standard set of security, observability, and reliability patterns that have proven indispensable for successful, planet-scale microservice deployments.

Just when you were starting to get comfortable with Kubernetes concepts, the state of the art leaps forward based on the simplifying assumption that we are starting with a Kubernetes foundation. That’s the power of a platform standard that can now be Certified Kubernetes across providers and vendors.

5. Java Gets Even More Wired

Java continues to be one of the most popular development and application platforms in the world — so let’s add a couple more shots of caffeine for faster release cycles and cloud innovations.  With applications transitioning to container-based, modular microservices and functions, Java in 2018 will address developers’ hunger for more frequent releases, flexible licensing needed to support the changing landscape driven by the cloud, and innovation from diagnostics to cloud.

6. Serverless Unlocks Cloud Lock-in

Serverless has been one of the big remaining closed and proprietary cloud-native technology areas to date. This has forced enterprises to choose between cloud lock-in or adopting early service tools like Lambda. That’s all about to change in 2018 as a set of open serverless projects, such as our own Fn Project and CNCF efforts move forward. “Open on Open” is the only way to move open serverless forward in 2018 — building a serverless solution on an integrated stack on top of a Kubernetes foundation.

And in 2018, perhaps we can work on weaning ourselves off of the term “serverless” in our tech nomenclature and focus instead on “functions.” But I wouldn’t count on it.

7. Kubernetes-Native Becomes a Thing

If 2016 was all about cloud-native and then 2017 shifted up-stack to container-native, what about 2018? Looks like we move on to Kubernetes-native. Kubernetes will become the lens through which people interact with their products. Look for the rapid emergence and wide availability of K8S operators for key products. If Kubernetes is the standard management and orchestration fabric, then more enterprises will demand standardized management through that layer and vendors will respond by adding Kubernetes operators to make their interactions “Kubernetes-native” — or more simply “Kubernative?”

8. Developer Productivity Explodes as Toolchains Embrace Containers, Kubernetes

The transition to modern, Docker-centric CI/CD continues as development team realizes that in this new container native world, having your toolchain be Docker-aware and Kubernetes-native gives them productivity advantages to increase release cadences and release cycles. CI/CD becomes less a standalone tool and more of a full container lifecycle platform that seamlessly integrates the entire build, deploy, and run DevOps cycle — but does it naturally with built-in container registry integrations and one-click native Kubernetes deployment automation.

Kubernetes growing pains are now being addressed: manageability, governance, compliance. There’s always a second-order set of “hard” problems that are left to operations teams and enterprise management teams after a strong software adoption wave like we’ve seen with Docker and Kubernetes over the last few years.

The good news is that there’s already a full body of work and industry activity moving forward on an integrated operations tool stack for monitoring, observability, and logging with projects such as PrometheusOpen TracingJaeger, and Fluentd leading the way in the CNCF. Furthermore, work on Kubernetes Federation is starting to address the more complex multicluster, multicloud management operational challenges like managing and auto-scaling global applications or deploying spot clusters on-demand.

Even more so, the thorny problem of applying application-aware decision logic to deployment choices factoring in cost, regional affinity, performance, quality of service, and compliance constraints.

The Cloud Native Computing Foundation is a sponsor of The New Stack.

Feature image via Pixabay.


A digest of the week’s most important stories & analyses.

View / Add Comments

Please stay on topic and be respectful of others. Review our Terms of Use.