Kubernetes is a popular DevOps tool thanks to its container-centric environment and portability across infrastructure providers. In 2018, Kubernetes had a big year, being the first project to graduate from the Cloud Native Computing Foundation (CNCF) and landing at #9 for commits and #2 for authors/issues on GitHub, coming in second only to Linux. “Three of the largest cloud providers offer their own managed Kubernetes services,” explains CNCF. “Furthermore, according to Redmonk, 71 percent of the Fortune 100 use containers and more than 50 percent of Fortune 100 companies use Kubernetes as their container orchestration platform.” CNCF also points out that Kubernetes is used in production at a massive scale by global companies like The New York Times, eBay, Uber, Goldman Sachs, Buffer, and others.
Given its widespread (and growing) use, it’s no surprise that there are many evolving Kubernetes trends and best practices at present. For DevOps professionals, staying on top of the latest trends and learning emerging best practices for working with Kubernetes requires a commitment to ongoing education. (Check out our list of the 50 best Kubernetes tutorials for some useful learning resources.) DevOps pros should also keep Kubernetes security in mind, and build security into the development cycle as early as possible. To measure your organization’s security maturity level and learn more about how Threat Stack can secure your containerized environments, take Threat Stack’s Cloud SecOps Maturity Assessment.
In 2019, one interesting trend will be Kubernetes getting closer to bridging the gap between development and operations. Developers who design and implement the applications that ultimately run on Kubernetes are getting more attention from Kubernetes maintainers. We are seeing increasing maturity in the tools that allow developers to take their application source code and seamlessly deploy it to a running cluster without having to write any setup code in between.
Another emerging trend grows out of the fact that early Kubernetes adopters were plagued by the inconsistent performance of stateful applications. In 2019, we will see greater support for stateful applications that run on standard block storage or networked distributed storage. This will allow operators to run various databases, queues, and other caching components of their microservices with resiliency built-in.
To find out what Kubernetes trends are piquing interest in the DevOps community at present, we reached out to a panel of DevOps pros and Kubernetes experts and asked them to answer this question:
“What are the most interesting trends to look for in Kubernetes right now?”
Meet Our Panel of Kubernetes Experts:
Read on to find out what Kubernetes trends you should be looking for in at present and beyond.
Ben Bromhead is the CTO at Instaclustr, which provides a managed service platform of open source technologies such as Apache Cassandra, Apache Spark, Elasticsearch, and Apache Kafka. Prior to co-founding Instaclustr, Ben worked as an independent consultant developing NoSQL solutions for enterprises.
“In 2019, developers and enterprises using Kubernetes are welcoming a growing trend toward easier database integrations…”
On its own, Kubernetes has a relatively nascent understanding of database functionality: Kubernetes is blind to important details of the database that is being written to, and its capabilities for storing data in-state are fairly limited. While running popular databases on Kubernetes is simple to begin with, these limitations make enterprise-grade implementations very challenging.
However, new workarounds are arriving that greatly facilitate Kubernetes database integrations. For example, enterprises looking to run the NoSQL Apache Cassandra database on Kubernetes now have the option of using an open source Cassandra operator, developed by Instaclustr and partner contributors, which functions as a Cassandra-as-a-Service on Kubernetes. The Cassandra operator, freely available on GitHub, takes deployment and operations duties completely off of developers’ plates, while providing a consistent environment and set of operations founded in best practices, and reproducible across production clusters and development, staging, and QA environments. Going forward, solutions such as this will enable developers to much more easily realize the full advantages of using Kubernetes and their chosen databases in combination.
Glen is the VP of Product at NeuVector, a container network security company. He has over 20 years of experience in enterprise security, marketing SaaS, and infrastructure software. He has held executive management positions at Trend Micro, Provilla, Reactivity, Resonate, Quantum, and Rignite.
“I think the most interesting Kubernetes trend this year will be…”
Early deployment experiences with service meshes on top of Kubernetes. Service meshes like Istio and Linkerd are swelling in popularity, as enterprises are excited about what they provide (such as routing, discovery, and encryption — though many have not analyzed the performance and management costs of a service mesh). Additionally, service meshes don’t support all protocols, so hybrid environments will need to be managed. This will be a particularly interesting trend to watch play out in the Kubernetes ecosystem this year.
Andrei Vasilescu is a renowned digital marketing expert and CEO of the money-saving platform, DontPayFull. He has been providing cutting-edge digital marketing services to various international companies and different online coupons of various brands for years.
“At present, we are watching a few interesting and new trends in Kubernetes…”
Two of these new trends are:
- Affinity to Hybrid and Multi-Cloud: A general shift towards hybrid-cloud and multi-cloud strategies is being noticed in Kubernetes at present. A vast part of public cloud users are adopting a multi-provider strategy, and this has increased from 13 percent in 2016 to 27 percent at the beginning of 2019. In addition, hybrid-cloud usage reached up to 32 percent from 24 percent in the same time frame. Seeing this increasing trend, we can conclude that Kubernetes will remain the most-used platform for container deployment in hybrid-cloud and multi-cloud infrastructure environments. Kubernetes has developed their enterprise version for multi-cloud and multi-cluster deployment.
- Fast Increasing Interest in Windows: Kubernetes 1.9, the beta version for Windows, was released at the start of 2018. Since the launch of this Windows version, the interest for Windows has been growing very quickly, and it has not shown any hint of slowing down. A recent Kubeadm survey indicated that one out of six users is interested only in the Windows version. In addition, as a part of Tigera Essentials offerings, Calico is now available for Windows for Kubernetes users. Kubernetes will be commonly available for everyone in 2019.
Markku Rossi is the CTO at SSH.com.
“Platforms are getting more and more complex…”
Containers and container orchestration are adding their own layer to be managed, in addition to traditional server (virtual/cloud) access management and security. There are three dimensions to this. One is system complexity — it is harder to know what’s happening and who has access to which data with what rights. The complexity adds a lot of configuration points that are susceptible to configuration mistakes. Another is security as cryptography — securing communication between containers and container clusters and multi-cloud. And lastly is keeping the system up-to-date with security updates — Kubernetes, underlying OS, and software components inside the containers.
Matthew Barlocker is a startup fanatic. Creating businesses, teams, products, and process is where he thrives. His background is in software engineering, so he prefers deep tech companies, like Blue Matador, his alert automation startup.
“One of the biggest trends to look for in Kubernetes this year is visibility…”
Now that Kubernetes is the de facto standard for container orchestration, we should see some filling out of the ecosystem by way of visibility. The best thing about Kubernetes is also incredibly dangerous — it hides errors, restarts, config issues, and more by moving containers automatically, creating a huge avalanche of issues when it finally breaks. The next trend is visibility into the running system.
Will McGrath is a marketing manager for Red Hat Storage.
“Some of the biggest trends in Kubernetes at present are…”
Expansion of the Kubernetes Operators framework to not only package and deploy Kubernetes applications but also actively manage them.
Continued maturity of the KubeVirt technology to help migrate certain types of virtual machine-based workloads that cannot be easily containerized by allowing them to run side-by-side with your containerized applications.
Continued maturation of the rook.io CNCF project. Rook is an open source Cloud Native Computing Foundation project which positions itself as a cloud-native storage orchestrator — providing the platform, framework, and support for a diverse set of storage solutions to natively integrate with cloud-native environments.
Alex is the Principal Consultant at Scalar Software.
“One trend that I find interesting is turning Kubernetes into a platform…”
I’ve worked with teams in the past to turn Kubernetes into a platform, but at the time, a lot of the tooling had to be built in-house because it didn’t exist otherwise. With all the new tooling that’s available, turning Kubernetes into a platform is becoming easier and more accessible for everyone. Every major cloud provider has started to offer managed Kubernetes services, which makes it more user friendly.
An interesting tool that was recently released is Knative. Knative is a tool that simplifies the way teams deploy applications to Kubernetes. This improves the user experience for developers who are using the platform.
Another interesting tool is Rancher Kubernetes Engine (RKE), which simplifies how you deploy a Kubernetes cluster, which reduces a lot of complexity for operators. RKE makes it easy to launch clusters anywhere.
Turning Kubernetes into a platform allows teams to focus on development while still receiving all the benefits that Kubernetes offers. With the Serverless add-on that Google released for GKE, it will be interesting to see how it plays out in production. Serverless Kubernetes could change the way teams are using Kubernetes today.
Shawn Moore is the CTO and co-founder of Solodev, a web development and Web Content Management System (CMS) provider in Central Florida.
“This year will see a container market in flux…”
There is no clear winner, and the requirements continue to challenge IT operations.
Both Docker and Kubernetes have become essential for any digital transformation journey. The evolution to the Cloud and the emergence of DevOps is all about change, and the cumulative acquisition of skills — first with Docker, then Docker Compose, Swarm, and now Kubernetes — have been driving this transformation. As we approach Virtualization 3.0, it’s clear that any real digital transformation strategy can’t be achieved without containers. Quite frankly, if it involves the servers, the Cloud, and the internet, it must have containers.
Rancher is propelling Kubernetes to a new level of multi-cloud computing performance. The new Rancher 2.2 was just released a short time ago, providing a more unified platform for managing Kubernetes clusters in production and deploying clusters anywhere, on any provider. New features include Global DNS for provisioning and maintaining DNS records, RKE (Rancher Kubernetes Engine) with backup and disaster recovery snapshots that leverage S3 storage, and mutli-cluster application deployment.
As more and more services move to the cloud and containerization, organizations will be able to focus more on deploying their applications and less on servers. For example, at Solodev, we see containerized applications like our Web Content Management System (CMS) and DevOps merging as one, allowing you to scale up and down at will, rapidly move from testing to production, add new services, and drive innovation faster than ever.
Kubernetes is complex — that isn’t about to change. Organizations are still struggling to understand the need and the role of containers in their DevOps strategy and build teams that can develop and support these technologies.
The bigger players will fully dominate the space, specifically Amazon Web Services with their AWS Elastic Container Service (ECS) for Docker, AWS Fargate, and now AWS Elastic Kubernetes Service (EKS). This is driving greater validation and trust around these orchestrations, but more importantly, making Docker and Kubernetes easier and faster to adopt and manage. AWS will also drive the greatest adoption in large part because of its consumption-based pricing.
This year, more businesses will be bringing container skills in house to build advanced architectures and develop AWS CloudFormation templates to modernize their IT operations. Unfortunately, the talent pool for Docker and Kubernetes is still relatively small and nascent, creating a shortage of qualified people with competitive salary demands. The field is also changing rapidly, putting a strain on current skill sets.
Security will continue to be a key issue because of shared operating systems. Significant vulnerabilities impacted Kubernetes in 2018, causing IT operations to reconsider their deployments. This is another reason why AWS will continue to attract more users with their secure and scalable offerings. Trust is a major factor; while you might sacrifice some capabilities by aligning with a partner at scale, AWS is delivering ease of use, reliability, and less risk around containers.
There’s been a lag in support for Docker and Kubernetes, but expect that to improve as we progress through 2019 — led by AWS. Having the market presence and global reach, infrastructure partners with the scale of AWS will provide the necessary resources for containers and orchestration to truly reach the next level of market performance.
Sean Porter is the creator of the Sensu project and the co-founder and CTO of Sensu, Inc. Sean is a seasoned systems operator and software developer with a decade of experience in automating infrastructure. As CTO of Sensu, Inc., he oversees the development of Sensu.
“Kubernetes requires a change in traditional approaches to how we think about systems and visibility…”
Originally launched by Google in 2014, Kubernetes has seen impressive growth driven by a rich ecosystem — and fast-tracked its way to become the go-to for container orchestration and management. But with that growth comes challenges. Kubernetes adoption means there is now a higher volume of (generally) smaller moving parts to monitor. As a new set of problems unfolds and businesses look for effective ways to deploy and manage apps this year, Kubernetes requires that we continue to evolve our approach to monitoring.
As more businesses adopt Kubernetes as part of their multi-cloud strategy, I expect the realities of day two operational challenges, such as avoiding expensive downtime and maintaining visibility, to be an ongoing challenge this year and beyond. They’re also faced with the complex challenge of integrating otherwise loosely connected systems — for example, attempting to collect data from modern systems like Kubernetes using legacy tools (e.g., Nagios) or getting data from legacy systems (e.g., SNMP traps, or metrics collected in outdated formats) into modern tools like InfluxDB. I’ve also noticed a growing trend of Kubernetes naysayers who are commenting on how complex it is, and it’s a fair argument. There are a ton of moving parts in Kubernetes and integrating them is itself a full-time job — hence the need for the right tools (and people!) to connect these systems. Even responsible cloud-native thought leaders like Kelsey Hightower will readily admit that there’s no magic bullet (not even K8s!) that you can use to solve every problem. You still have to use the right tool for the right job — a.k.a., the right solution to connect your multi-generational (i.e., Kubernetes and bare metal) infrastructure.
Tom Petrocelli is a research fellow at Amalgam Insights.
“Kubernetes will continue to be adopted by organizations for microservices architectures and fuel the adoption of microservices architecture…”
The most interesting trend will not be in Kubernetes itself. Instead, it will be the ecosystem around Kubernetes. Expect the expansion of networking, storage, monitoring, and deployment software, designed for containers and Kubernetes. These will make systems based on Kubernetes enterprise-capable.
Ian McClarty has over 20 years of executive management experience in the cybersecurity and data center industry. Currently, he is the CEO and President of PhoenixNAP Global IT Services.
“I believe the following three trends surrounding Kubernetes and will continue to gain momentum in 2019…”
The growth of hybrid and multi-cloud with Kubernetes being the common platform.
- There is an ongoing initiative from the community to push Federation V2 API to the GA stage. Multi-tenancy, during last year’s Kubecon the multitenancy-group was formed to work on a draft document. I’ve seen the first version, and it’s going to make things much easier. Also, I see many little improvements entering K8’s control plane, for instance, Continerd with its low-level plumbing API which is gRPC gaining fast market share in the container runtime world. Then things like IPVS and CoreDNS replacing iptables in kube-proxy and kube-dns.
- The integration of Kubernetes with service mesh, such as Istio, will become more prevalent. Service mesh provides more robust service-to-service communications and traffic management (e.g., load balancing) than what is provided natively by Kubernetes, which can ease the challenges of scaling microservices running on Kubernetes.
- Centralized management of multiple Kubernetes clusters, such as development, test, and production clusters, or multi-cloud setup for dynamic workload handling, will become more common. Multi-cluster management tools, such as Rancher, are becoming more powerful, which can simplify the complexity of managing Kubernetes clusters and help with enterprise adoptions of Kubernetes.
Mat Igleheart currently serves as a Cloud Architect with Lakewood, Ohio-based Onix. An AWS-Certified Solutions Architect, AWS-Certified Cloud Practitioner, and Google Cloud Sales-Certified technical salesperson with in-depth knowledge of cloud-based solutions and a focus on customer satisfaction/service, Mat has worked with Telecom, Dell, and Avaya.
“Firstly, Kubernetes is a tool to manage what the cloud world refers to as a ‘container’…”
Containers simplify the build/test/deploy pipelines in DevOps. With Docker containers, developers own what’s within the container (application and service, and dependencies to frameworks and components) and how the containers and services behave together as an application composed by a collection of services.
Because of this, the strong shift toward cloud containers and using Kubernetes should continue in 2019 and beyond.
In a nutshell, DevOps people (web developers) like containers because it makes their coding easier and simpler and more repeatable and efficient. When they push new code out in containers and use management tools like Kubernetes, they don’t see the old mistakes of coding occurring, like new updates breaking existing applications, bringing down entire environments, and having major outages. Security will always be important. Meanwhile, the continued adoption of multi-provider strategies — hybrid and multi-cloud — will also continue, with Kubernetes being the common platform for container deployment across a variety of environments.
Zachary Smith is the CEO of Packet.
“2018 was a year of huge adoption for Kubernetes, and nearly every cloud provider rolled out managed Kubernetes offerings (AWS EKS, Digital Ocean, etc.)…”
As we move through 2019, I think the three most interesting trends in Kubernetes will be:
- Kubernetes is Going On-Premises: As the enterprise market for container orchestration accelerates, and service mesh helps to make multi-cloud deployments easier to manage and secure, we’re going to see the major cloud and hybrid cloud vendors taking their Kubernetes experiences on-premises. From Google (GKE on Prem) to Red Hat (Openshift Dedicated), this will be a major theme.
- Kubernetes is Getting Smaller: Users are looking at Kubernetes as a control plane that goes with them anywhere — and not just in the public cloud or on-premises, but at the edge and with IoT devices. Rancher’s k3s, which is deployable in a low memory environment, is the first of what we expect will be many iterations of the Kubernetes experience for constrained environments.
- Kubernetes is Tackling Telco: The enterprise market is shifting towards cloud-native practices, but equally important are the massive infrastructure investments being made by existing and new telecom providers rolling out 5G and CBRS wireless. With hundreds or thousands of locations needed to drive low-latency and cost-effective operations, the appeal of highly deployable and portable Kubernetes-based infrastructure is hard to resist. Major strides from the CNCF Cloud-Native Network Functions (CNF) working group and others mean that telcos will soon have a robust Kubernetes-based alternative to Openstack.
Itay Ariel is the Senior Software Developer of cnvrg.io, a data science platform helping teams manage build and automate machine learning pipelines. Itay has deep knowledge of Kubernetes capabilities and specializes in Kubernetes. He leads the integration of Kubernetes into the cnvrg.io stack.
“Kubernetes has already been widely adopted by data scientists for simple usages like running Jupyter Notebooks…”
In 2019, we will continue to see this trend of Kubernetes for machine learning, but will also take a role in completing the AI loop. The AI loop consists of data cleaning using Spark Map Reduce, model training using Tensorflow or other ML libraries, and deploying models to production. In addition, Kubernetes will be used to facilitate continual learning, which allows published models to continually train and adapt to live data and improve accuracy. Kubernetes will be used to scale, schedule, and run complicated scenarios written in simplified YAML declarations. Kubernetes is a fundamental component on the way towards continual and auto-adaptive AI.
Chandler Song is the Co-Founder and CEO of Ankr and an engineer-turned-entrepreneur. He worked at Amazon Ads Team, Amazon Lab126 in IoT and FireOS, as well as SAP Enterprise Supply Management. Chandler invested early in Bitcoin and was involved with the blockchain student organization during college at UC Berkeley.
“Kubernetes is clearly in charge now, and Docker Swarm will drop in importance…”
Kubernetes will be built in as an underlying service, so users will no longer (suffer to) operate it. CSP Marketplaces will be the top choice for distribution and adoption of K8s and container-based apps for most of the globe.
Security will improve as image and build metadata and rootless usernetes are gaining traction.
Alex Sokolov is a senior software engineering manager at iTechArt Group. He has 10+ years of experience in software development, building, and mentoring engineering teams.
2019 will see increased usage of Kubernetes as it improves on its efficiency in managing applications automatically. It is also likely that multi-level encryption and authentication mechanisms will be implemented in Kubernetes to help boost its security. Lastly, we’re likely to see added contributions to Kubernetes by the major cloud providers — AWS, Azure, and GCP to allow for deeper native integrations.