Multicloud Is An Exercise In Strategic Thinking
While some enterprises examine multicloud strategies prior to deployment, others adopt it without giving the concept a second thought.
This article was written by Tolga Tarhan - Senior Vice President & GM, AWS Services, and originally published on Forbes.
There are plenty of schools of thought on multicloud adoption strategies. While some enterprises examine the topic prior to deployment, so many others adopt multicloud as a standard approach without giving the concept a second thought. Either way, one thing is clear and consistent: Not everyone in the enterprise understands how to optimize their multicloud strategy.
While there is some benefit to pooling everything into one preferred cloud (the most compelling to me being the economies of scale and simplified security), there can be equally compelling reasons to transition to multicloud. They just need to be the right ones.
Let’s begin with why organizations consider multicloud in the first place. As the senior vice president and general manager of AWS services at a cloud-native services company, I see two main reasons companies choose to embrace a multicloud strategy.
The weak reason — and not an uncommon one — is thinking that a multicloud strategy is a way to pressure service providers into discounting their fees. Organizations think they stand to gain by pitting two cloud providers against each other or using the flexibility of a multicloud application to leave one provider for another. The reality is that the discount structures used by cloud providers are pretty well set, and few organizations have the influence needed to motivate providers to negotiate on their price points.
In most scenarios, adopting a multicloud strategy for the purpose of price arbitrage is not a solid choice because it can end up costing more, and the drawbacks can negate the benefits.
For example, deploying a single application to multiple cloud platforms can not only cost additional time and money in development, but it also forces the application to use only those services that are available from all cloud providers, effectively limiting the application to the lowest common denominator.
I believe the best reason for going multicloud is when you're seeking out a specific capability, service or benefit that is significantly superior in one cloud environment compared to others. Enterprises should consider selecting a default cloud for a majority of their workloads and utilizing other cloud providers on a workload-by-workload basis.
Contrary to what many might think, the complexity in going multicloud is not always technology related. The connectivity requirements are straightforward, and the services, for the most part, are not so different from each other, which makes it relatively easy for engineers to become familiar with them and adopt them in a relatively short period of time.
There is, however, more to the process than signing up online and hoping for the best. The real complexities lie in the areas of corporate governance, security, operations and privacy. With multicloud, security teams have to build additional security postures and policies. In addition, operations team members may need to be available 24/7 to manage and monitor problems in another cloud. That could be beyond the scope of an organization’s financial and human resources.
Once your organization decides that the move to multicloud makes strategic sense, there are two areas that you should consider upfront: What is the best way to go about it, and how do you know when the time is right?
The following are some basic factors to consider:
• Start by determining the primary cloud supplier from which to grow out your specialized needs. Then identify workloads that may be candidates for a different cloud, and evaluate whether that investment in a second provider is worth the added costs and complexity. Be sure to validate if the supplier can deliver before committing.
• Assess the current state of cloud usage within the organization. Many companies are surprised to discover how many independent teams randomly register for cloud services for independent projects without any kind of cohesive strategy. It’s important to contain this multicloud “sprawl,” determine what exists and decide what those clouds' future should look like, as well as establish appropriate policies, moving forward.
• Determine if you have a team in place that can support a multicloud environment. This is especially crucial in terms of your ability to manage the added security and privacy functions.
As for timing, review your budget to see if adding a cloud service is viable or whether it makes more sense to wait. The reality is that adding a cloud service at a later date will be the same amount of work as it would be today, so there is no need to rush the decision.
It doesn’t make sense to preemptively spend the time and energy on a strategy if you aren’t ready to use the full depth and breadth of multicloud. The only thing rushing into a decision will generally accomplish is a more expensive environment that you cannot fully utilize. I've found it's better to spend the money on more immediate strategic needs and build out your multicloud strategy when it makes the most sense.
Ultimately, multicloud is a strategic business decision that requires considerable scrutiny. It’s not that it’s hard to make that decision; the challenge lies in making the right choices for the right reasons.