“With Containers as a Service, regardless of what you plan to bake, you no longer have to build the kitchen first”
What’s in store for DevOps this year? Will cloud adoption continue to grow? Will containers-as-a-service gain momentum? What challenges should Kubernetes and Docker address in 2018? We talked with Peter Williams, CTO at CJC Ltd. about all this and more.
Containers, microservices, cloud, serverless: A tale of DevOps
JAXenter: What are your DevOps predictions for 2018? What should we pay attention to?
Peter Williams: As the name suggests, DevOps looks at things from a developer’s perspective, while reducing the friction of getting systems into production. Approaches like microservices and control planes should enable enterprise integration and governance across systems. However, challenges remain around which approaches will define the standards, giving plenty of focus for DevOps in 2018.
JAXenter: What makes a good DevOps practitioner?
Peter Williams: As strange as it may seem, for me a good DevOps practitioner is both hardworking and lazy. By that I mean that a good practitioner is always training, always striving to increase the breadth and depth of their skills and knowledge – and thereby lessen their potential workload. They will be open to gaining knowledge in whatever technologies or disciplines will improve them as engineers/developers.
They can then apply that knowledge to the creation of automated solutions and tooling to avoid repetitive manual endeavors that get in the way of the good stuff!
JAXenter: Will DevOps stay as it is now or is there a chance that we’ll be calling it DevSecOps from now on?
If you aren’t already thinking in terms of security, or you don’t have a plan for security, it will come back and bite you.
Peter Williams: Realistically, what we call it is less important than how we’re approaching it. At CJC, we’re supporting the financial services industry, so security is a primary concern for us. We have to think and act in terms of security by default.
Within our developer teams, we’re looking at greater utilization of Kubernetes and service mesh type deployments, where we can use technologies like Istio to offload some of the heavy lifting of securing applications/services and the need to hard code these functions and features. Obviously, you may not necessarily want to abdicate responsibility for those functions, but we’re exploring options.
Now that approach may not suit everyone, but in my opinion, if you aren’t already thinking in terms of security, or you don’t have a plan for security, it will come back and bite you.
JAXenter: Do you think more organizations will move to cloud in 2018?
Peter Williams: Absolutely yes. Cloud adoption will continue to grow for the same reasons it always has. More so now, as I think it’s better understood and there are more public success stories. Another factor that I think will actually accelerate adoption is AI and machine learning. These services are easier to achieve in the cloud. People want to leverage these technologies and utilizing the cloud is a great way to get at them quickly.
JAXenter: Will containers-as-a-service become a thing in 2018? What platform should we keep an eye on?
Peter Williams: As great as containers are, they need to be deployed and managed and that takes a skilled hand. What interests me is the democratization and commoditization of technology. If you imagine where we were when cloud first came to prominence with the relatively simple concept of the IaaS, suddenly we didn’t have to worry about hardware anymore. Then, as things progressed, we got PaaS and SaaS, which further stripped away what we needed to worry about in order to access or deliver the services we wanted.
Technology has continued to progress up this curve, where we as consumers and creators of technology services need to worry about less and less about the underlying delivery piece and instead focus on our applications and services. People often refer to deploying code in containers as “baking” vs traditional deployment methodologies, which they call “frying.” I guess the point I’m making here is that with CaaS, regardless of what you plan to bake, you no longer have to build the kitchen first. So, keep an eye on the platforms that require you to build less kitchen.
Is Java ideal for microservices developments? Should companies continue to invest resources in this direction?
Peter Williams: As we know, Java is super mature, well understood and there are a lot of skills in the market supporting a rich ecosystem. I think if your dev team is already well-versed in Java and you need to make strides quickly, then it will absolutely do a job. However, I’m a believer in the right tools for the job and microservices really lend themselves to this mode of thinking. If the only reason you’re using Java is “it’s what we’ve always used” then I think you need to take a moment to consider your options.
JAXenter: Containers (and orchestration tools) are all the rage right now. Will general interest in containers grow this year? If yes, how are we going to handle the older applications?
In 2018, I think the roles could reverse with Docker benefiting from the growth and adoption of Kubernetes.
Peter Williams: In my view, containers and orchestration tools will absolutely continue to remain very interesting and go hand in hand. Following on from my earlier point on CaaS, if we can get to a stage where all the complexity is handled within the system, then you can put the power to deploy services on-demand directly into the hands of the business and therefore quickly get value from your technology.
Regarding legacy technology, this is a subject very close to my heart as our dev teams are currently trying to solve this exact problem. In our industry, like many others, we’re at a real turning point. We have all this new exciting technology that we want to embrace, but we have to be mindful of legacy systems that are architected in the old ways.
As it stands we’re writing a lot of wrappers, which allow us to containerize these legacy applications and deploy them as efficiently and flexibly as you would any other container-based deployment. This is not an easy task, but fortunately, we’ve got several hundred engineering years of experience around these legacy systems and we’re relying heavily on that to build our experience/expertise into our solutions.
JAXenter: What challenges should Kubernetes address in 2018?
Peter Williams: I think 2018 could be a “difficult second album” for Kubernetes. How do you follow the year  they had?! That’s not to say I think they will struggle, quite the opposite, I just think they have some hard work ahead to deliver on their initial promise and to support a community that’s growing fast.
JAXenter: What challenges should Docker address in 2018?
Peter Williams: Docker are in a strong position and Kubernetes has benefited in no small part because of it. In 2018, I think the roles could reverse with Docker benefiting from the growth and adoption of Kubernetes. However, this is not an opportunity for Docker to sit back and relax as there is still room for improvement. There is always room for innovation and you are never too far away from disruption.
In fact, now could be a critical moment for Docker. While commercially they are looking to move up the stack, they could actually find themselves blocked because of Kubernetes. Meanwhile, microkernels are growing in prominence thanks to IOT and could look to attack Dockers’ market from below.
JAXenter: How will serverless change in 2018? Will it have an impact on DevOps?
Peter Williams: It’s still relatively early days, so I expect a lot of change. The key change will happen around tooling, which should improve in terms of quality and variety. That should have an impact as it will make serverless easier to utilize and debug, which will lead to more choice – always a positive in my view.
JAXenter: Will serverless be seen as a competitor to container-based cloud infrastructure or will they somehow go hand in hand?
Peter Williams: Going back to my thoughts on languages for microservice development, this is all about the right tool for the right job. I prefer to think of them as options instead of competitors because applying the competitor label suggests it’s one or the other, or a right way and a wrong way. This boils down to how I can best achieve my desired outcome with the available technologies and cost models. Is your goal speed, efficiency, simplicity, lower run costs, getting the most out of limited dev resource?
Take a look at our interview series with nine DevOps influencers:
- Part 1: Collaboration or survival of the fittest: Who runs the DevOps world?
- Part 2: DevOps without automation is like peanut butter without jelly
- Part 3: DevOps dream team or superstar DevOps engineer? Tips for aspiring DevOps engineers
- Part 4: Key DevOps metrics that matter: How well does your team sleep?