Admins cite “infrastructure” as the greatest challenge to container adoption, according to a recent Diamanti survey. What does that mean, exactly, and how do organizations overcome this challenge? We recently spoke with Jeff Chou, Diamanti’s CEO, to find out. Here’s what he had to say.
A migration to containerized environments doesn’t happen on its own, of course. You need a plan. And that plan has three main components, according to Chou.
Choose Your Infrastructure
The first step in container adoption, he says, is simply planning your infrastructure. “There are so many tools and open source software projects to understand” in the world of containers, Chou says, that figuring out which ones to use to build your containerized infrastructure can be daunting.
That’s why “you have to study the ecosystem to determine which elements are right for your organization’s purposes,” Chou says. “You have to pick out the resources comprising the layers of your stack. You have to guarantee that they all work together.”
Plan for Production-Environment Challenges
Building a containerized infrastructure is only the start of the battle, he says. Putting containers into practice successfully also requires managing a production environment successfully.
Managing containers in a production environment requires you to think about much more than just your infrastructure. It also means “guaranteeing service levels, providing all the infrastructure services around storage services, data management, resiliency, failover, service level agreements and all the things enterprises need before pushing anything into production,” Chou notes. “Yes, and you need 24-by-7 full-stack support.”
He adds, “After climbing the hill to build the container environment, enterprises quickly undergo an existential reckoning—now they have to actually care and feed this thing they built! What happens when they find a bug? How do they debug it? What happens if that bug is in the open source?”
Making the Jump to Multi-Cloud
The challenges of containers don’t end once an organization has a production environment running smoothly, Chou says. The growing popularity of multi-cloud architectures means that companies must contend with the additional challenge of managing containers across multiple clouds—while also, in many cases, retaining on-premises containerized infrastructure.
“Invariably, applications need to not only scale, they also need to be moved across cloud environments,” he says. “The cost of a cloud service or regulatory compliance will compel an organization to move an application out of the public cloud back to their own data center.”
The answer to this conundrum, according to Chou, is to automate, automate, automate. While multi-cloud architectures are inherently challenging, automated policies that govern how containerized applications run and scale make it possible to handle the complexity of a running containers in a multi-cloud architecture.