It’s hard to remember that Amazon was once just a scrappy e-commerce startup and they did things that many startups do in the heat of growth, build up technical debt.
In 2000, very much out of necessity, Amazon made their strategic technology decision to start building reusable modules for its internal development groups. This allowed these groups to create new features faster because they weren’t always reinventing the same thing over and over again.
As time went on, the collection of internal services grew and people inside the company started to realize that maybe there’s a business opportunity there.
First launched in 2004 and then relaunched in 2006 with three public pay-as-you-go services, Amazon Web Services (AWS) set sail into the uncharted waters of what we now call cloud computing.
Around the same time that AWS announced their first cloud services, Microsoft already had similar efforts in play, doing much the same thing as Amazon: stitching together disparate services and creating reusable components for internal staff.
Microsoft officially announced that it had something cooking for the cloud in 2008, which they called Project Red Dog because, well, “Pink Poodle” just didn’t have a really nice ring to it.
Officially released as Windows Azure in 2010, Microsoft’s first iteration of cloud service . . . wasn’t really that good. It couldn’t even run a standard version of Windows server. A bit of dog-fooding and an offsite retreat in 2011 was enough to convince the players in Azure’s leadership that they needed to do better if they hoped to be competitive with the likes of AWS and Google.
Microsoft Azure relaunched in 2013 much better positioned and a much more compelling option for organizations, especially those who were longtime Microsoft shops and were looking for a smooth path to leverage the growing cloud wave that was coming. In 2014, Windows Azure was renamed Microsoft Azure, which symbolically and literally marked a shift in strategy for Microsoft making the cloud a priority.
Google’s cloud origins have been baked in from the very start, but officially entered the public cloud in April 2008, through the preview release of Google App Engine, aimed at letting developers quickly deploy and scale web applications.
Now, in contrast to AWS and Microsoft, who emphasized heavily the infrastructure as a service path in their early days, Google chose a platform as a service, PaaS. And over the next several years, Google used its world-class development organization, acquisitions and deep pockets to grow Google Cloud Platform into something that can compete with its fellow public cloud providers.
Both Google Cloud platform and Google Workplace, formerly G Suite, are business segments within Google (and Google itself as a business segment within its parent holding company, Alphabet).
At present, Azure falls into the Microsoft Intelligent Cloud business segment, which covers all the server and cloud products, as well as the support and consulting parts of the business and GitHub, which Microsoft purchased in 2018.
AWS is a division within Amazon, but operates as its own entity within the greater company. The commerce side of Amazon is a customer of AWS, meaning that AWS has to compete for business just as if they were an external company.
Now, one thing to note when comparing cloud providers with regard to revenue is that their reporting groupings are not the same. Amazon reports AWS revenue separately, while Google includes both GCP and their Workplace product as part of their cloud revenue. Microsoft bundles Azure, on-prem server software, Visual Studio, and GitHub into their reporting segment. Sometimes Microsoft also refers to their commercial cloud category of products, which includes Azure, Office 365, Dynamics 365, and part of LinkedIn. As of 2021, Microsoft had yet to publish revenue numbers purely for Azure by itself.
Let’s look at some customers for each platform.
Again, you have to use caution here when looking at providers’ logo slides. Even if some small, obscure, tiny division of a huge company uses the platform to store a 3K file on object storage, that provider is more than happy to claim the logo all for itself.
Additionally, many of these large companies have lots of divisions, some of which they may have acquired or still in the process of integrating. It’s not uncommon for huge companies to be using four or five different cloud providers throughout their vast organizations.
How about some comparative strengths of each platform?
Well, with strengths come weaknesses. So what are some challenges facing our three big cloud providers?
So how do I possibly choose which cloud provider is best for me? Well, there’s really no best universal answer because all situations are different.
I suggest conducting some pilot projects with these providers. And I’d also throw in IBM, Oracle Cloud, SAP or Alibaba, if you have an existing relationship with those companies already. Maybe try to do the same project on multiple providers and evaluate things like ease of use, robustness of offerings, quality of documentation, and so forth.
If you’re concerned about becoming too locked into one provider, you can choose some architectures that are more open and universal, like using layers of abstraction, building around REST APIs and using containers.
But be cautious here because you can get too generic. If you’re just using your cloud provider to run the same old VMs and containers that you had on-prem, you’re likely missing out on lots of value that cloud provider can offer in the way of managed services and variable costs.
Now, I know this may be controversial, but I would steer clear of trying to spread your workload out over multiple cloud providers. I know, I know . . . some people think that using multiple cloud providers shields them from the dreaded vendor lock-in. I think that spreading your workloads across multiple clouds just dilutes your resources and makes things needlessly complicated. Rather, I believe it’s better to focus your upskilling and efforts to maximize the features and value of one provider. Go deep on that provider and get really good, whomever it happens to be.
Well, my friends, this has been another edition of Cloud Provider Comparisons. Thanks for reading, stay safe, take care of one another, and keep being awesome, cloud gurus.