Cloud wars: Who should you cast your lot with?
Google rolled out its new multi-cloud platform, Anthos, earlier this month. In the cloud wars, Google finally seems to be on par with Amazon and Microsoft.
But the battle is far from over. With Google gaining traction, the marketplace is bound to shift again, and again.
How will Google's strengthened position in the cloud wars change the pace of multi-cloud innovation? Most importantly: What does it mean for IT organizations trying to make a decision? Here's what you need to know.
The problem, framed
Amazon, Microsoft and Google are the three leading contenders in the public cloud space. The popular questions—about which one is the best, or which one is right for most enterprises—may not be the right questions to ask.
All three do a great job with the cloud computing basics, namely, the infrastructure that includes storage platforms, compute platforms, and databases. While other features and functions, such as the wide array of machine learning and big data cloud services now available, are nice to have, the reality is that most enterprises don't need those services right now, and perhaps won't ever need them.
The issue is that Amazon, Microsoft, and Google have been focused on a feature/function war of sorts. They release any hyped service that provides a leg up on the other two competitors and that promotes the use of the core cloud services (storage, compute, and data).
Just go to any of the user conferences put on by Amazon, Microsoft, or Google. They are all well-attended these days, but 90% of the time the focus isn't on what their customers are using, but on what's new.
These new services may not even be available at the time of announcement. The larger strategy is to promote the adoption of their public cloud over the others, with arrows pointing toward their latest release as proof that they have a reputation of being thought leaders and innovators.
In other words, they'll talk high in the stack, but you'll leverage the lower-stack services.
Cut through the hype
So how do you wade through the publicity to pick the right cloud for you? Okay, you know I have to say this: It depends on your core requirements now, as well as your enterprise's overall vision for the use of cloud computing.
It also depends on your enterprise's perceptions about which is the "right solution," and that may be based upon current beliefs and usage more than on the actual properties of a public cloud provider.
So how should you be thinking about this issue? It’s critical to understand a core set of requirements.
Know your storage needs
This is not just a GB or TB number; it's really about what you use now as storage, what will likely move to the cloud, and storage analogs of existing on-premises systems. If you deal mostly with storage of relational databases, then you need suitable versions of those databases.
For instance, MySQL on premises, or the AWS cloud-native version Arura, which is basically MySQL on the AWS cloud? There are many other examples like this on all the public cloud platforms.
Getting back to raw storage, look at how things are stored now before the migration, such as block, object, or file-based storage. Typically, you'll have versions of all three. Understanding this, you can pick the least problematic paths, and those become your requirements for cloud-based storage.
There is good news here: Amazon, Microsoft, and Google offer great cloud storage products. However, the costs differ, depending on what you need. It's time to put on that green visor, pull out the slide rule, and crunch some numbers.
You won’t see an obvious winner at first, but if you look at your own requirements, the right solution will to rise to the top. At the very least, you can use our requirements to provide a ranking that will help you make decisions and consider other requirements.
Performance and costs: Not what you expected?
Lift-and-shift is becoming the cheapest way to move to the cloud. This involves picking a clone of the platform you’re running on premises—say, Red Hat Linux with attached object storage—and simply moving the code and the data.
The simplicity of lift-and-shift, as well as the speed and the cost, has made this the go-to method to migrate workloads to the cloud. Since you're moving from platform A to platform A, things should function the same.
At least in theory.
While many enterprises view lift-and-shift as a slam dunk, the reality is a bit different. Some enterprises have lifted and shifted their way to Amazon, Microsoft, or Google only to find that their workloads aren't optimized for the host cloud platform. The cost ens up being higher than expected, and performance lackluster.
Here's the reality: Workloads moving to public cloud platforms use those platforms best if they are modified to take advantage of native services. These services include security, governance, cost management, monitoring, etc.—things that are of value only if the applications are aware of their existence.
This means a partial refactoring for cloud-native features, which results in a higher cost for migration. Moreover, "native" means you’ve married the specific cloud platform.
Amazon, Microsoft, and Google are all pretty much the same when it comes to the platforms they offers, and the ability to lift and shift. Some come with greater ease for specific types of workloads. For example, Microsoft public cloud is an easier port from existing .NET on-premises infrastructures.
So you need to figure out what you have, what you need, and how best to get there.
It's a multi-cloud world
Hybrid cloud was supposed to be the platform that offered the best of both public and private clouds. But private clouds could not keep up with the R&D spend of Amazon, Microsoft, and Google, and so have fallen out of favor with most enterprises that deployed them.
The silver lining is that the options improved with the rise of multi-cloud, which simply means leveraging more than a single public cloud at the same time.
While most people in the know in the cloud industry have watched this trend rise for some time, until the recent Google Anthos move, Amazon, Microsoft, and Google have all been in denial about the use of public clouds beyond their own. Enterprises now view multi-cloud as the likely solution for them, with the only question remaining being which cloud providers to place in their portfolio.
It's time for the leading cloud providers to accept that multi-cloud is a reality, and that public clouds should focus on working and playing well together. This is a critical path for success.
There are several add-on players that sell cloud management platforms (CMPs), cloud services brokers (CSBs), systems built on Kubernetes, and other layers of technology that can manage the underlying heterogeneous public cloud services. They typically do this either through abstraction, orchestration, or both. But until now, the key public cloud providers have not seen a benefit to lead this space.
How Anthos fits in
Now Google's Anthos (formerly the Cloud Services Platform) lets you build and manage applications across a multi-cloud infrastructure. A Kubernetes-based system that also uses other open-source technologies, Anthos is as much an architectural approach as it is a cloud service.
Its mechanisms allow you to modernize both new and existing applications using containers, microservices architecture, and a service mesh. You can run within a single cloud provider, or across different cloud brands. This new approach lets you focus on the function of the application, rather than where it's going to run. You can move quickly without compromising security or governance or increasing complexity.
Google's Anthos is a game-changer. It's a sure bet that Microsoft and AWS will provide some sort of multi-cloud product of their own, but right now Google stands alone. Chances are that Microsoft and AWS will scoop up one or more of the independent CMPs and CSBs to compete against Anthos. But Microsoft and AWS could also build their own products from the ground up.
Where to cast your lot today
One-choice answer is no longer a hard-and-fast requirement, and that realization should stop some of the internal bickering in IT around which cloud to pick. In most instances, if you truly understand your requirements, and you’re truly committed to selecting best of breed, then multi-cloud is your answer.
In short, multi-cloud gives you the ability to pick the right products for today's needs and to adjust as your needs change in the future.
If you think this will lead to complexity, risk, and higher cost, you're right. That said, you can mitigate that cost and risk by using tools such as Anthos, or the landslide of tools that will likely follow it.
Moreover, as the public cloud providers fall in line behind Amazon, Microsoft, and Google to build technology that plays well with other cloud platforms, and the company moves its focus to portability and operations, the game will permanently change.