The world of technology loves hot, new, sexy trends like AI, ML, Big Data, AR, VR and a handful of other acronyms.
Cloud computing is no different.
And in the cloud world, two trends have stuck around for years: multicloud and hybrid cloud.
In this article, I’ll explain what these two terms mean and why most startups shouldn’t care about them.
I’ll also talk about another hot trend/media fabrication – cloud repatriation.
Let’s get started.
What Is Multicloud?
Put simply, multicloud means using multiple public cloud providers at the same time.
In theory, a multicloud workload should be portable and functioning within two or more cloud ecosystems.
For example, you can use different IaaS vendors or get your IaaS, PaaS and SaaS solutions from AWS, Azure and GCP separately.
There are a few reasons why organizations decide to go the multicloud way like redundancy or backup concerns, for example. We’ll discuss these potential benefits in a bit.
For now, let’s move on to hybrid cloud
What Is Hybrid Cloud?
Hybrid cloud is pretty similar to multicloud and people often confuse the two terms.
Here’s the fundamental difference:
Hybrid cloud usually refers to using two or more different infrastructure types, most often a public cloud and a data center.
Multicloud, on the other hand, means using two (or more) public clouds.
In short, they’re pretty similar concepts.
That’s why (at least in theory) they should bring similar benefits to organizations.
The Benefits Of Multicloud And Hybrid Cloud
As I said earlier, there are many potential reasons why businesses decide to adopt these practices.
First, avoiding vendor lock-in seems to be a common argument for using multicloud architectures.
In general, vendor lock-in is often listed as one of the main concerns when discussing cloud computing’s pros and cons. By going multicloud, businesses can attempt to mitigate some of the risks of being locked into one provider’s infrastructure.
Redundancy is another argument people use to justify multi/hybrid cloud.
For example, if for some reason, AWS stops working entirely tomorrow, part of your infrastructure will still function inside your own datacenter or in Azure/GCP.
Finally, some workloads might not make sense in the public cloud because of regulatory or economic reasons. In these cases, organizations are forced to adopt a hybrid approach out of necessity.
Again, there are also other reasons which I won’t cover here.
Right now, it’s important to focus on a topic that people don’t discuss often – why most startups really shouldn’t care about either of these trends.
Why (Most) Startups Should Avoid Multicloud And Hybrid Strategies
As a cloud consultant, most of my customers are startups and small to mid-sized businesses.
They typically want to build and scale a product, app, service, website and etc.
As you can guess, their teams are usually small.
They don’t have 10 cloud engineers in-house, nor do they have the time or manpower to coordinate between two different cloud providers and a data center.
This might be anecdotal, but I suspect the same is true for most startups.
If you’re in a similar situation, please, don’t bother with multicloud or hybrid cloud.
Here’s why:
For a startup, the cloud provides three key benefits – speed, simplicity and agility. All three go out of the window if you try a multi/hybrid cloud approach without enough manpower.
First, it takes tons of time (speed) and effort (simplicity) to truly understand one cloud provider’s nuances, let alone two or three. This is time that I’m sure you’d rather spend on developing your product.
Second, each cloud provider gives a discount when you commit to using their services for a set amount of time. That discount gets way smaller if you divide your cloud expenses between multiple vendors.
Finally, you’re also not avoiding lock-in despite what anyone tells you. And the whole lock-in problem is way overplayed anyway.
We’re all locked in by the technology choices we make. Adding one more cloud provider doesn’t alleviate the lock-in problem. It simply adds another tech stack that’s hard to navigate (agility).
In short, don’t go multicloud or hybrid cloud just because they’re „good practices“.
Make your choices based on what makes sense for YOUR business. And in most cases, I’m willing to bet that going all-in on one provider is the best choice.
Now, let’s finish with the newest cloud trend I’ve read about – cloud repatriation
What Is Cloud Repatriation?
Cloud repatriation is supposed to be a growing new trend that sees companies moving back to their data centers.
It’s the opposite of the cloud migration trend from a few years ago.
I’m talking about cloud repatriation here because I’ve recently been seeing more and more articles and webinars on the topic.
Some data center vendors, consultancies and media companies are trying to push the narrative of „the cloud is too expensive, so everyone’s going back to the good old data center“.
So, let me save you some trouble:
Cloud repatriation doesn’t seem to be a successful trend, at least not in any meaningful shape or form.
People didn’t just wake up and realize that public cloud services can be expensive. Everyone who’s worked with AWS, Azure, or GCP for a few weeks knows how fast the bills rack up.
That’s why it’s considered a best practice (and common sense) to do some basic math before going to the cloud.
Of course, some workloads are better off in a traditional data center.
But that doesn’t mean that „Some IT leaders are learning that running certain applications in a public cloud can cost more than doing so on-premises…“ (direct quote from an article on a media website).
Yes, obviously, running some applications is more expensive in the public cloud, but there’s no industry-wide research to provide reliable data about the cloud repatriation.
Of course, it may turn out that 10 years from now, everyone’s forgotten all about the cloud (although I doubt it). We’ll just have to wait and see.
Finally, if you need AWS or DevOps advice , don’t hesitate to drop us a line.
