IBM Taps Kubernetes for the Bluemix Container Service
IBM continues to empower developers with new services and capabilities for building cognitive applications in the cloud, delivering a new cloud service that taps the power of the Kubernetes container orchestration system and Docker containers.
The new IBM Bluemix Container Service, launched at the company’s InterConnect 2017 event, is a container service on IBM’s Bluemix cloud platform that helps developers simplify the creation and management of apps using Kubernetes and the Docker engine.
Jason McGee, an IBM Fellow and vice president and chief technology officer of the IBM Cloud Platform, said the IBM Bluemix Container Service automatically provisions, updates and monitors the containers that developers use to create apps, helping them to more easily build cognitive applications that tap into IBM’s real-time security intelligence.
The Power of Containers
“Building microservices with containers helps developers to do what they do best – invent – by creating and stitching together different services that work in tandem,” McGee said in a statement. “Our service does this in a simple, straightforward way by using Kubernetes to automate critical parts of an app’s environment, while giving developers the ability to build in cognitive intelligence, blockchain and Internet of Things services.”
Containers lie at the forefront of the developer movement towards microservices, as they enable developers to rapidly iterate, deliver and operate on the components of an app, without being bound to one underlying infrastructure or cloud provider, IBM said.
The popularity of containers has placed them among the hottest segments of the developer landscape right now, with 451 Research predicting the market opportunity for containers to exceed $2.6 billion by 2020.
The IBM Bluemix Container Service includes fully-managed Kubernetes clusters for deploying, scaling and managing containers when building apps, as well as native Kubernetes APIs, built on Bluemix. It also features automatic load balancing and provides users with the choice to store and access data across multiple platforms and repositories, and flexible isolation and capacity controls to design a container cluster to fit specific needs and budget constraints of an application, McGee said.
The Bluemix Container Service also provides a metrics on each container cluster’s performance, and offers access to IBM cloud services such as Watson APIs, blockchain, cloud data services, Internet of Things tools and more.
Optimized for the Enterprise
Overall, the technology is “pretty horizontal” and IBM has seen adoption in many industries such as transportation, financial services, media, insurance and others, McGee said.
“We haven’t done a lot to optimize for certain industries,” he noted. “We have done a lot to optimize around what we think are the kind of enterprise requirements our customers might have. We’ve done a lot on isolation and having options out of public cloud. We’ve done a lot of work around security and vulnerability analysis. We look at their container images. We integrate with X-Force to get the latest intelligence about what going on with security in the industry and apply that to the workloads. We even do things like analyze their application development function to see if they have a vulnerability that could surface when deployed.”
Indeed, security features are built in. For instance, IBM’s Vulnerability Advisor, Bluemix’s tool for scanning and flagging security issues within containers, is a key component of IBM Bluemix Container Service. And the service will continuously monitor each container through an integration with the IBM X-Force Exchange, the company’s threat intelligence sharing platform.
Community and Contributions
McGee said IBM has been contributing to the development of Kubernetes since the project’s launch in 2014. Big Blue also has contributed to the Docker project.
By pairing a Docker-powered engine with the simple management capabilities of Kubernetes, Bluemix developers can access a highly usable interface and dashboard to easily write code within a container and quickly deploy it to multiple apps, IBM said in a statement.
“The Kubernetes thing is really important,” said Angel Diaz, IBM’s vice president of cloud technology and architecture. “The single tenant for cluster stuff that we do in our Kubernetes can help ease the road for people who use Docker. This is no coincidence. I have been bringing these worlds together physically over the last two-plus years. We opened Docker with the Open Container Initiative. I stood up onstage as OSCON with Google and opened up Kubernetes, and I just brought Docker into the CNCF [Cloud Native Computing Foundation] party where they just announced containerd. To me, that is an aligning of the stars that is fundamental.”
McGee noted that IBM has had a container service in Bluemix for two years as people are looking at how to build next-generation apps and are looking at how containers play in building those new apps
“Containers are seen as a really cool technology,” he said. “And I think what’s happened over the last year or so is people starting the transition from using containers for development to using containers as part of their DevOps process to do packaging. But with the exception of early adopters, most people have not changed their operational model yet. They take that container, they shove it in a VM they run it in production as if it were a VM. They use their old processes. But what’s starting to happen now as they’re consuming more and more containers and trying to get more agility and trying to figure out how to adopt operationally with a container model. That’s why you see orchestration systems like Kubernetes gaining tremendously in popularity and maturity and contributions.”
Moreover, as IBM looks at the enterprise IT space from a cloud perspective, “like many things in cloud we think there’s real value in ‘as-a-service’ delivery of these capabilities, of removing from customers that burden of setting up and managing container systems so they can just focus on building applications,” McGee told The New Stack. “So, what we’re trying to do with the container service is give you all the power of these advanced container orchestration systems, but we’re the ones who set them up and manage them and upgrade them and scale them and integrate them and provide you with all of that as a service so all those distractions are removed.”
Original Bluemix Container Service
IBM’s original container service on Bluemix was different in a couple of respects. One is it was built it two years ago before container orchestration was really a thing. It provided a model where you simply could run a container or a cluster of the same container — IBM took care of running everything and ran it in a multitenant way.
“So, you had a user experience like you get with VMs today,” McGee said. “If you go to a cloud today and you want to run a virtual machine, you don’t start by installing a hypervisor and then running VMs on it. You just run a VM. That’s how we started.”
And the reality is that is likely where IBM will land, he added. But that’s not where the market is right now and that’s not where the technology is right now. Indeed, good ideas need good timing, too, he noted.
So, over the last six months, IBM took a step back and said they needed to get onto a community-driven orchestration solution. And as Big Blue had done a lot of analysis and were active in both the Docker and Kubernetes communities, they settled on Kubernetes.
“We feel that the containers market has not picked a winner yet, so I won’t say Kubernetes is the winner,” McGee said. “I think we feel like they have the most momentum. It’s a very good architecture where we’re contributors. I think it has the best representation right now of all of the concepts that you need to really build a meaningful production application with containers. So, we think it’s the best tech.”
IBM also is betting its own services on Kubernetes. The company is building much of the Bluemix stack going forward on Kubernetes, including Watson and other things, McGee said.
“Now, if you want to hit the market broadly, the other big player is Docker,” he noted. “We use Docker engine and Kubernetes uses the Docker engine. Well, Docker Swarm orchestration is the other big one. And clearly, we’ll have a story there. We don’t have that now, but we think the market requires that you have a story for both. We’ll see where the market goes.
IBM kicked off the beta for its Bluemix Container Service in Dallas and will be expanding it to other markets as the company prepares to make it generally available soon.
“It’s fantastic that IBM is progressing open technology by drawing on a major community project — Kubernetes — as the foundation of its new container service,” said Alexis Richardson, technical oversight committee chairman of the Cloud Native Computing Foundation and Weaveworks CEO, in a statement. “Using the flexibility and control of both Kubernetes and Bluemix, we expect to see a greater number of developers accessing and taking advantage of cognitive services and Watson APIs to build apps.”
The Cloud Native Computing Foundation is a sponsor of The New Stack.