Cutting-edge IT infrastructures aren’t just for nimble startups anymore.
At this year’s Kubernetes Conference, Cloud Native Computing Foundation community members including Red Hat, NetApp and Rancher debuted solutions that will help organizations that aren’t cloud native make a new and difficult type of digital transformation: from legacy architectures to containers and microservices.
Precise numbers for container and Kubernetes adoption among enterprise companies are tough to nail down, Rancher Labs vice president of global channels and alliances Jim Sarale told me, but he knows that general Kubernetes use is poised for explosive growth.
“I wouldn’t say it’s at its infancy, but it’s definitely on the hockey stick — we see it exponentially growing,” he said. “And with the remote workforce and the kind of new world order that we’re all living in, I think we’ll see that adoption continue to spike and get to that 60 or 70 percent we’re all hopeful for in the next couple of years.”
For reference: CNCF released survey results indicating that container use leapt 300 percent since 2016, from 23 percent of respondents to 92 percent. Thirty percent of respondents in 2020 came from companies with more than 5,000 employees.
“It’s a difficult task for enterprises to flip the switch.”
The benefits of technologies like containers and microservices for small, cloud-native startups are well documented, and the advantages for large enterprises are similar: faster deployments, more performant software and better UX.
But, unlike tiny startups, enterprises have some serious barriers to adoption. Rejigging IT infrastructure requires time, money and organizational buy-in. Companies are finding more and more need for edge computing and data hosting, which complicate that infrastructure even further.
“It’s a difficult task for enterprises to flip the switch,” Sarale said.
To navigate the double-challenge of edge migration and container transformation, enterprises must quickly get savvy to a couple of important concepts, Red Hat senior director of product strategy Brian Gracely told Built In.
One of those concepts is new-ish: hybrid cloud set-ups and their advantages. The other is very familiar: The benefits of sticking with open-source solutions.
Public Cloud, Private Cloud or the Edge — Why Not All?
The economics and user experience of centralized cloud computing fall apart when companies want to attach sensors to thousands of delivery trucks, add voice recognition to self-checkout machines at remote grocery stories or scale a 5G network.
“Having to bring everything back via the network to the cloud and then send it back again — sort of multiple round trips all the time — would prohibit some of these new types of applications from ever happening,” Gracely said.
The solution to that problem lies in edge computing, or placing servers and data centers geographically close to end users so that information doesn’t have to pass back and forth in the inefficient loop Gracely described. But the edge comes with its own set of problems.
Imagine a vertically integrated delivery company, Gracely suggested. As a package makes its way from a warehouse to a customer, it moves through national, regional and local tracking systems. Last-mile delivery systems must communicate with customer-facing web apps. The company may roll out different applications for corporate, warehouse and delivery employees — and the list goes on.
Perhaps that adds up to five different technologies — A, B, C, D and E. As the system becomes more complex, everything running on the edge becomes harder to see and understand. After all, Gracely said, there’s not just one path from the warehouse to the customer — there could easily be 5,000. That means keeping up with components A, B, C, D and E times 5,000, which quickly becomes an impossible task.
“Like if technologies B and C had updates that came in every six months, but technologies A and E had had software upgrades or patches that came in every three months, I’m constantly doing nothing more than maintaining all this craziness in different versions and different upgrade paths,” he said.
“The more consistent you can be ... the more you’re going to focus on delivering a great experience, instead of how complicated it is just to keep the whole thing running.”
Red Hat — which rolled out updates last week to its Enterprise Linux and OpenShift Kubernetes platforms — has been preaching the importance of hybrid cloud infrastructure for this very reason. With hybrid cloud, enterprises have a centralized plane to manage edge deployments, as well as software hosted in public clouds and on premises.
“The more consistent you can be from the cloud side of your system to the edge side, the more you’re going to focus on delivering a great experience, instead of how complicated it is just to keep the whole thing running,” Gracely added.
NetApp, an end-to-end data management provider that announced an integration with open-source enterprise Kubernetes platform Rancher last week, provides one such plane, a hyper-converged infrastructure (HCI) with a single UI. By partnering with Rancher, NetApp can take some pressure off enterprise IT departments with hybrid cloud set-ups, NepApp senior director of product management McClain Buggle said.
Perhaps IT professionals are transitioning applications from on-premises to public clouds for the first time, switching from traditional IT to DevOps or scaling out edge capabilities. Whatever the challenge, toolsets meant specifically for hybrid infrastructures will help teams create secure, governable, trackable systems — instead of chaos.
“The world is hybrid, maybe not in the way that we probably talked about it five years ago, where we thought it was all about bursting [in which companies push computing to public clouds only when IT demand spikes], but more about bringing the tooling that the teams need,” Buggle said.
Leveraging the Power of Open Source
When it comes to their hybrid cloud set-ups, enterprises have a few options: proprietary systems, open-source systems or a mix of the two.
Red Hat provides support services for the entirely open-source Enterprise Linux and OpenShift. In Gracely’s opinion, open-source solutions are the only way to go for two reasons.
The first is cost. Implementing and maintaining open-source software generally costs less than paying a vendor.
The second is use cases. For many enterprises, use cases at the edge are still being determined. For example, the contactless shopping and remote work brought on by the coronavirus pandemic drove companies to bring applications and their data closer to end users — that shift would have been impossible to predict even a year ago.
Unexpected use cases can throw a wrench in proprietary software. Production cycles are longer, and redeployments more challenging. Open source, on the other hand, prioritizes open, flexible ecosystems in which components “play nicely together” and developers rarely run into immovable technical roadblocks, Gracely said.
“If I choose something that’s proprietary, I’m pretty much going to lock myself into those four or five or six initial use cases,” he added. “I’m almost guaranteed that those things I didn’t know were going to happen — but eventually almost always happen — just aren’t going to work.”
“We can’t ask these organizations to be nimble in the way they deliver and let people consume things if we’re constantly restricting them and pushing them in a proprietary direction.”
NetApp, meanwhile, is an original equipment manufacturer, so its partnership with Rancher will serve as a “bridge” that connects enterprises with the open-source, cloud-native Kubernetes, Buggle said. For big companies attempting to go cloud native, an understanding of the open-source ecosystem and its pre-vetted solutions is a massive advantage.
“There’s all sorts of innovative technology and tools available to accelerate application development and digital transformation,” he said. “It’s just a reality that [enterprises] are not going to be in a position to test and validate and lock down and secure every single tool out there.”
However, open-source solutions aren’t for everyone, according to Buggle. Depending on company culture and industry, proprietary software may add security benefits that justify any drawbacks.
“I think it just depends on what culture you’re embracing within the company,” he said.
At the end of the day, though, the overlap between open source and cloud native is too big to ignore, Sarale said. Siloes between different departments, developers, connected devices applications and data centers are the No. 1 obstacle to the next big digital transformation from centralized systems to hybrid, containerized ones. In Sarale’s view, open source has always been the key to making disparate systems connect.
“We can’t ask these organizations to be nimble in the way they deliver and let people consume things if we’re constantly restricting them and pushing them in a proprietary direction,” he said.
‘One Size Doesn’t Fit All’
No matter how ready enterprises are for the shift to the edge and the adoption of containerized software, the trend’s direction is clear.
Forrester predicted that containerized, distributed software will “herald a new era” in the next five years, in which large enterprises can compete better with cloud-native startups. Sixty-three percent of respondents to Red Hat’s State of Enterprise Open Source survey indicated that they’ve deployed a hybrid cloud infrastructure, while more than half of those that haven’t plan to in the next two years. And Gartner estimated that, by the end of this year, 20 percent of business-critical apps will run on HCIs.
Meanwhile, IDC’s Worldwide Edge Spending Guide projected that software for the edge will make up 21 percent of spending in the $250 billion edge-computing market by 2024.
With sprawling organizations and complicated IT, enterprises have plenty of hurdles to leap in their journey to cloud native. Luckily, there’s more than one path to get there.
“We really need to be able to provide that bridge for large enterprises,” Sarale said. “One size doesn’t fit all.”