Maksym Yemelyanov - stock.adobe.
Container adoption has gone from near-zero to 60 in a staggeringly short period of time.
In 2017, technology consultants viewed containers as a promising technology and something customers could start experimenting within test/dev environments. Projects were limited and containerization was in no way a channel partner business mainstay, however. As 2019 comes to a close, containers have hit their stride, if not the mainstream. Containers have gone from an interesting possibility to widely accepted technology.
Ask Miles Ward, former cloud solutions lead at Google and now CTO at SADA Systems Inc., a Los Angeles-based IT and business consulting firm. "An easy point of reference is the main conference for containerization technology, KubeCon," he said. "In 2017, that was maybe just shy of 1,000 people and had fairly limited corporate sponsorship. Now, the event is a sprawling, multibuilding thing in San Diego."
The November conference, hosted by the Linux Foundation, attracted 12,000 attendees.
The spiking attention isn't limited to technology tire kicking, however. Consulting executives report an enormous uptick in container-related projects.
"Less than 10% of customer projects do not include containerization as part of the work that we are taking on," Ward said. "If you are going to start a software project or augment and existing [project], you are likely to include containers somewhere in that cycle, if not for the entirety of the system."
Container adoption grows
Other industry executives also cite the growth of container technology.
"It has accelerated massively," said Jonathan LaCour, CTO at Mission Inc., a managed service provider based in Los Angeles. "Pretty much every customer we have either has started the process, is evaluating or is well down the path of leveraging containers."
"The number of projects involving containers is growing and will be a bigger piece of our services effort in 2020 and beyond," added Juan Orlandini, chief architect at Insight Enterprises, an IT services provider based in Tempe, Ariz. "Quantifying [growth] is hard to predict, but we definitely do see acceleration," he said.
A key factor behind the rapid uptake: Enterprises see an opportunity for order of magnitude improvements in software development cycle time. That boost translates into competitive advantage as businesses of all kinds deliver more of their value in the form of software.
"The ability to get new patches, new features into the hands of consumers faster is the defining difference between high-performing companies and low-performing companies," Ward said. "The delta between the laggards and the leaders is not measured in percentage points."
Ward pointed to the 2018 book Accelerate: The Science of Lean Software and DevOps: Building and Scaling High Performing Technology Organizations, which states the highest-performing businesses experience 440-times faster lead time "from commit to deploy" compared with software development's backmarkers.
Orlandini said the benefits of containers may not be immediately obvious to traditional IT infrastructure teams.
"Containers and container orchestration systems function as an enabling technology for development teams to develop faster and bring value to an organization quicker," he explained. "In short, when used correctly, they accelerate the time to value."
In addition, Orlandini said containerized systems, when properly deployed, "make a hybrid cloud approach simpler than more traditional methods," such as virtualized systems that rely on a specific vendor's hypervisor. He emphasized the term "simpler" vs. "simple," however, suggesting there's considerable room for improvement.
"Much work is being done in the larger ecosystem that containers and Kubernetes live in to bring this to fruition," he noted. Kubernetes is an open source container management technology.
Old and new
Consultants said new applications and workloads are where they're finding much of their current business in containers and Kubernetes. Containerization breaks down applications into readily updatable microservices. That's much easier to do with a new software project compared with a legacy system. The cost of redeveloping an old app along microservices lines may overshadow the benefits.
Rick Erickson, co-founder and chief cloud strategist at Agosto Inc., a cloud services and development firm based in Minneapolis, said most of the container and Kubernetes activity he's seen of late fits into the new-workloads category. On the other hand, customers tend to avoid investments in older workloads associated with legacy SAP, Oracle and JD Edwards environments.
"They don't want to do an update in any way if they can help it," he said of customers managing such workloads. "Companies with millions of dollars invested in ERP or MRP [manufacturing resource planning] -- those tend to be more of a challenge."
Erickson cited the example of one enterprise with an enormous SAP implementation spanning 10,000 virtual machines.
"The practicality of moving that, with the investment load, would be really hard," he said. "It would require a heroic effort."
Mission's LaCour also cited greenfield development as a typical setting for containers and serverless technology. In the case of new applications, conversations with customers have gone beyond questions of technology acceptance.
"It's not, 'Should we adopt containers?'" he said. "We are going to, 'How do we do it best?'"
Legacy applications -- large SAP installations and off-the-shelf e-commerce -- systems, for instance -- represent the opposite end of the container adoption spectrum. "It happens, but it's rare," LaCour said.
An exception may be applications on the brink of transformation. He said companies may have created applications just before container orchestration emerged. Those apps, while not built entirely on microservices, may exhibit other aspects of modernization, making them potential candidates for containerization. For example, such applications may adhere to AWS' "well-architected" framework and follow DevOps practices, LaCour noted.
Ward said he's seen a pattern toward making Kubernetes easier to deploy for specific types of development. He pointed to Kubeflow, which aims to make Kubernetes a platform for machine learning, and Agones, which sets up Kubernetes as a game development environment.
Such tools seek to simplify Kubernetes, which, Ward said, has plenty of "nerdy bits" for adopters to navigate. Prior to Agones, for example, an organization would need to figure out how to build a Kubernetes game platform on their own, gluing together the necessary third-party products, he noted.
"The ecosystem is building preconfigured platforms using Kubernetes as a substrate," Ward said.
While those specific Kubernetes platforms rollout, Google is working to reduce the complexity of Kubernetes, in general. Ward said this effort ties into Google Kubernetes Engine, a managed environment for containerized applications, as well as Google's Anthos multi-cloud application management offering.
Erickson said he expects the ecosystems sprouting around Kubernetes to become a key source of activity in 2020. Those ecosystem developments will have less to do with the actual technology of Kubernetes and more to do with what the technology can enable, he noted.
"The technology is already there," Erickson said. "What you can do with it is what's interesting."