Down-to-earth offshoots of artificial intelligence are increasingly accessible for digital transformation work. These projects tapped into machine learning – with existing talent.
Multi-cloud strategy: 5 mistakes to avoid
Working with multiple cloud providers? Avoid these missteps as you craft a multi-cloud strategy
We’ve rounded up some of the common pitfalls, to help you navigate around them – and avoid unnecessary complexity, resource drains, and all manner of other headaches.
[ Are you ready to discuss cloud security concerns? Get our concise guide and learn from the experts: Hybrid cloud security: 5 questions skeptics will ask. ]
When you’re working with multiple cloud providers, it’s typically not a short-term play. You’re crafting a long-term strategy with high-stakes goals, such as the ability to be extra agile when needed. We tapped several multi-cloud experts for their advice for fellow IT leaders on shaping a multi-cloud strategy. Here are some major mistakes they advise you avoid.
1. You go multi-cloud without building enough cloud competency
As with so many IT initiatives, unrealistic multi-cloud expectations tend to be a harbinger of future pain. Organizations that are really making their first significant move to cloud, especially from an infrastructure standpoint, are typically best served mastering a single cloud environment before using multiple cloud providers.
“Unless you are a glutton for punishment, make sure you have really gone cloud before you try to go multi-cloud,” advises Anders Wallgren, CTO and co-founder at Electric Cloud.
Some organizations might overestimate their internal expertise, he says. Your team may already be using multiple SaaS apps – but that doesn’t mean you can seamlessly slide into managing multiple IaaS environments across multiple platforms, for example.
By “going cloud,” Wallgren doesn’t just mean becoming familiar with a particular vendor or platform, but ensuring that your people, processes, and tools are optimized for building, deploying, and operating services across multiple cloud environments.
“You should be able to reliably and repeatably build, test, deploy, qualify, [and] release code into your existing cloud provider,” Wallgren says. And you should so before trying to do the same across multiple platforms: “If not, you’ll want to fix that first.”
If you’ve got people, process, or technical issues with running workloads in a single cloud environment, moving to a multi-cloud approach is not by itself a solution, he says.
“Your release automation and software pipeline automation should be working well enough that you don’t have ‘release anxiety’ or ‘deploy anxiety,’” Wallgren advises.
2. Failing to design complexity into your strategy
Wait, did you just say we should design complexity into our strategy?
The “multi” in multi-cloud means more. More of the right things – flexibility, scalability, resource optimization – is great, and why the multi-cloud approach can have powerful allure for strategic IT leaders. But more also means more to manage, so you must plan for that.
“Getting the strategy right in terms of design and operations is critical,” says Amin Lalji, senior manager at EY Canada and a certified cloud architect and cloud security instructor at Learning Tree International. “Presently, the promise and vision of multi-cloud compute anywhere is still in its infancy and, operationally, there are a few items that need to be considered in order to manage operational risk.”
Don’t assume, for instance, that you can run most of your workloads seamlessly with any cloud provider. “The expectation is that workloads can be deployed and operated at any vendor an organization chooses,” Amin says. “In reality, there are architectural complexities that need to be carefully designed in order for current day multi-cloud implementations to be successful.”
Amin shares a few examples of those potential complexities that you’ll want to bake into your plan:
- Load balancing
- Health monitoring
- Code deployment
- Staff skill levels
3. Assuming cost savings are ensured
In our recent post on how to get started with multi-cloud strategy, we recommend being specific in your use cases. Otherwise, an appealing outcome – avoiding vendor lock-in, for example – might be an unfounded assumption, rather than a well-defined goal.
Cloud spend is another great example. Indeed, multi-cloud scenarios can enable the flexibility and choice you need to manage your costs. But don’t assume your costs will come down.
“[IT leaders] should be aware of the complexity of managing multiple clouds, the continuous upkeep of those environments, ongoing management and maintenance, and cost, cost, cost,” says Eric Sanders, cloud services consultant at Shore Group. The emphasis on that list item? It’s because Sanders sees it as the biggest potential issue if you’re not planning properly for it.
That gets back to strategy and long-term execution. As Lalji told us recently, there can be great cost advantages in a multi-cloud scenario, in part because of the cutthroat competition among providers. But it’s not a given.
He offered this example of tailoring your strategy toward optimizing cloud spend if that’s one of your multi-cloud goals: “Cost optimization should not be approached as a ‘shift workloads back and forth between vendors’ strategy, as much as starting new ones with one provider and retiring workloads on the other.”