Planning a Cloud Native Tooling Strategy in Uncertain Times
When you start to adopt Kubernetes (or indeed any other technology), you’re making decisions from the very first moment.
Which cloud vendor do you choose to host your workloads? Which Kubernetes distribution? Which service mesh? Which logging tool?
Choice is a wonderful thing — it gives you the flexibility to build the right tech stack to suit any use case and tackle any problem. Your stack will look like nobody’s else’s, and that’s just the way it should be.
But every choice inherently comes with trade-offs, particularly in a fast-paced world like cloud native, where there’s always a hot new thing to try.
The decisions your organization made in the past can constrain the choices it needs to make now and in the future; it’s not always easy to jump ship when you have running workloads to move, contracts to end, processes to rework, dependencies to untangle.
So you need to make your choices carefully — strategically — in order to maximize return and minimize risk. And, as you may have noticed, there’s a lot of risk out there right now.
A New Era of Cloud Native Risk?
You’re probably very familiar with the Cloud Native Computing Foundation landscape map and its staggering 1,100 logos. It’s not the only such landscape out there. If you’re doing Kubernetes at the edge, you may have come across a similar map from Topio Networks, with a whole new set of logos, and you can find others covering AI, 5G, blockchain — you name it.
A burgeoning community of thousands of vendors and projects is a sign of healthy innovation in early-stage ecosystems.
But it won’t last forever.
Every market goes through a “natural selection” process, with a few products in each category dominating, and others either dying off or moving into specific niches. Kubernetes itself is one winner, having displaced Swarm, Mesos and Nomad as the mainstream choice in the race for container orchestration.
This process is complicated by periods of market turbulence, such as we’re seeing today. As investors get cold feet, some startups run out of cash and stall. Large corporations tighten their belts and sharpen their axes for rounds of layoffs.
The knock-on effect for the community is that project maintainers lose their jobs and free time, projects lose their vendor sponsorship backing, underperforming “hobby” or experimental products get shut down, and free editions get turned off, sometimes unexpectedly.
This is going on right across the tech landscape, but there’s another dynamic affecting Kubernetes in particular: we’re also seeing a phase of large-scale mergers and acquisitions, such as SUSE’s purchase of Rancher Labs and Broadcom’s recent $61 billion buy of VMware.
“Some acquisitions can be very good,” said Holger Mueller, vice president and principal analyst at Constellation Research. “They keep a struggling vendor alive commercially, bring investment into a product and reduce integration challenges and pain.”
But the churn in the market can present challenges, too, said Nitha Puthran, senior vice president for cloud and infrastructure at Persistent Systems, a technology services company. “We are continuing to see private equity takeouts,” said Puthran. “Some of these are very helpful and add funds, focus and growth; others make cuts that might impact the technology that is important to you.”
Indeed, any change of ownership and leadership always brings the very real risk of changes in a company’s strategy, pricing and service levels.
“We have seen some products change their pricing drastically,” said Prashant Iyengar, chief technology officer and co-founder at Miko, a consumer electronics company. “Others have reduced their contribution to the software’s community edition or stopped it altogether.”
“In some cases, we’ve seen a vendor has been acquired, and their tool stops being supported in just a few months,” said Iyengar.
So in a risky market like today — or even in a safe one — what can you do to make better vendor and product choices?
Choose Vendors and Products with Momentum
This time of economic uncertainty reinforces the importance of strategies that minimize risk, said Iyengar. It’s important, first of all, to look for stable vendors that will likely still be around with funding, a user base and strong talent.
“Has the product or open source project been a top performer for at least a couple of years?” Iyengar asked. “Has it gained momentum and positive buzz across tech blogs and other media or strong GitHub ratings, reviews and comments?”
“In some cases, we’ve seen a vendor has been acquired, and their tool stops being supported in just a few months.”
—Prashant Iyengar, CTO and co-founder, Miko
Ask your peers about a vendor or product you’re evaluating, suggested Carmen Taglienti, distinguished engineer at Insight, an IT consulting firm. “Talk to your colleagues or the community,” he said. “Make sure the vendor has a good reputation, especially for the use cases you are considering.”
Past performance is often a good indicator of whether the company can execute on its roadmap, said Puthran. “Where is the company’s funding at?” she asked. “What is the track record for security vulnerabilities and the remediation process and time?”
Balance Stability with Innovation
If you’re looking for safety, it can be tempting to turn to larger, longer established vendors. But just “buying big” is not automatically a safe choice: companies like Google routinely sunset products and features or even shut down whole business units without offering affected customers an adequate offramp.
Besides, safety shouldn’t be your only consideration: you want agility, too. There’s a risk in buying a product from a larger vendor, particularly if you’re a smaller customer. Large vendors may exert their market power to ratchet pricing up. They are rarely as responsive to customer requests as hungry startups are; and nor do they deliver so fast — you’ll often be waiting a long time for roadmap features.
So choosing the right vendor is about looking past the market cap at the quality of the product and the executive team, understanding their competitive landscape, and their financial health.
It’s important to assess a vendor’s roadmap — and whether they’re delivering on it — before swapping out a part of the tech stack with their product, said Puthran: “If you don’t see continuous product and feature enhancements from the vendor and a strong customer-first attitude, you should look elsewhere.”
Bias Toward Open Standards and Integration
When choosing a vendor, look for one that builds on open standards with minimal lock-in — so it is easier to move on if the worst happens.
This is not just about whether the vendor or their project is strongly opinionated, with limited integrations, but how committed the vendor is to working with you and supporting your freedom to choose. If you encounter an interoperability issue, will the vendor point the finger elsewhere or step up and work with you on a solution? How strong is their documentation?
“We’ve found that many tools do not scale well globally, and assessing this helps us determine whether we adopt different tools for different regions,” said Iyengar.
An organization should ask and verify if a product will integrate with the rest of its stack, he added: “If the tool’s integrations do not perform well, or if it fails to integrate with a critical new part of our stack, that would cause us to seek alternatives.”
This is far from a theoretical consideration — it matters today. According to a new report by Spectro Cloud, two-thirds of production Kubernetes users already suffer interoperability issues between software elements in their stacks, and a further 27% put significant effort into avoiding interoperability issues.
Be Real about TCO and Hidden Costs
When making an investment decision, naturally most decision-makers focus first on whether the technology product meets the specified feature requirements, then the onus is on procurement to secure the keenest price. This is an opportunity to refocus the discussion on the whole value picture.
“Rationally, it is about TCO, or total cost of ownership,” said Mueller. “Does the vendor create value, keep creating value and is expected to create value for an enterprise?”
A good strategy for avoiding sudden increases in product cost is to lock in pricing for the vendors to whom you commit. “For core tools, we prefer to lock in pricing via long-term contracts for at least three years,” Iyengar said. “This minimizes changes related to the economy.”
But TCO extends way beyond pricing. “Value can take many forms, but the most important ones drive competitive advantage, increase efficiency, provide security, and reduce support and maintenance costs,” said Taglienti.
Two-thirds of production Kubernetes users already suffer interoperability issues between software elements in their stacks, and a further 27% put significant effort into avoiding interoperability issues, according to a new report by Spectro Cloud.
In particular, you should evaluate the impact of good old-fashioned technical support over the expected lifetime of the product. This is vital in the cloud native ecosystem, because in many cases, the options available are pure open source software, built by a small team of contributors.
This is great from an innovation perspective, but when you’re deploying code from a small project into production at scale, who can you call when things go wrong? An apparently free product may actually end up costing an organization a lot more than a commercial product that includes support, onboarding and hosting, particularly in the long haul.
Support is of course something you can DIY — but there’s clear demand from users for a robust option backed by service level agreements. According to Spectro Cloud’s report, 45% of organizations picked 24/7 technical support for the whole Kubernetes stack as a feature they’d look for from a Kubernetes management vendor.
Hedge Your Bets — and Keep Your Eyes Open
One way to avoid placing the wrong bet on a technology is to hedge and adopt multiple products. According to Iyengar, his company has started incorporating at least two vendors for some core tools and technologies.
“In some cases, we develop a version of the tool internally in an A/B testing or staging environment,” he said. “This way, the tool can easily be replaced in the upcoming revision cycle.”
“Rationally, it is about TCO, or total cost of ownership. Does the vendor create value, keep creating value and is expected to create value for an enterprise?”
— Holger Mueller, vice president and principal analyst, Constellation Research
This is not a perfect solution, for obvious reasons. Maintaining multiple technology investments to serve the same purpose not only means you’re paying twice for integration and setup costs, but potentially incurring additional costs in training and documentation, and accumulating complexity.
Another option is to maintain a state of readiness for switching. While switching always has a cost, complacency and sticking with a familiar incumbent have a cost, too.
“Each year, we perform a thorough evaluation to ensure our current tools are still the best choice for our needs,” said Iyengar. “This evaluation often reveals other tools with more benefits and capabilities.”
Persistent Systems’ Puthran echoed this notion. “If you are not able to achieve significant benefits that increase over time with your current choices, you should look at alternatives.”
Move Methodically
When you’ve weighed the ROI and calculated the TCO, and decided that it’s time to make a change in your stack — now’s the time for a cool head.
“Run a proof of concept to make sure the assumptions you are making about the value proposition are correct and any risks are known,” said Taglienti. “In most cases, start small with an MVP, or minimum viable product, to learn the nuances of the vendor’s platform or products and understand the learning curve within the organization.”
In other words: look before you leap.
Whatever You Do, Choose Wisely
Figuring out a tooling strategy for distributed, cloud networks isn’t optional. More than 95% of new digital workloads may be deployed on cloud native platforms by 2025, up from 30% in 2021, according to Gartner.
The analyst firm also predicts that over 85% of organizations will embrace a cloud-first approach by 2025 and won’t be able to fully execute their digital strategies without using cloud native architectures and technologies.
This means it will only become more important for an organization to have a good strategy in place for selecting tools and the vendors that sell and maintain them.
There are many ways to build (or rebuild) the stack that works best for your organization’s needs. Weighing all these factors can help you find the right solutions and sustain them through uncertain times.
But remember, everything you do is a choice, even if you do nothing. So take the opportunity and choose wisely.
