Comprehensive Discovery : AWS vs Azure vs Google Cloud

Published:16 June 2021 - 7 min. read

In the battle of cloud computing, a trifecta of solutions stands out vs. the overall marketplace. Companies predominantly turn to AWS vs. Azure vs Google Cloud to handle their needs. Thankfully, there isn’t a bad choice in the bunch. Each has strengths and weaknesses, and a team’s perception of these will depend on their infrastructure requirements.

Not a reader? Watch this related video tutorial!
Not seeing the video? Make sure your ad blocker is disabled.

How do they compare? This overview breaks down each service and assesses what they bring to the technological table.

Aws vs Azure vs Google Cloud Market Share

By a fair margin, AWS reigns supreme as the most popular cloud service—capturing 33% of the global SaaS/PaaS market. Next up is Azure, which claims 18% of the market. Google Cloud earns the bronze medal here.

It’s important to note that AWS launched in 2006—two years ahead of both Azure and Google Cloud. The service has had more time to mature and permeate the cloud realm. Amazon has also captured many customers first in a space where the “familiarity factor” is strong. That’s been an important leg up. To switch, competitors must offer something truly compelling or undercut on price (perhaps delivering better value).

That said, AWS has moved beyond its explosive growth phase for the time being. A 2020 Synergy Research Group report showed that AWS grew in lockstep with the rest of the market. Interestingly, Azure has outpaced its rivals growth-wise by a decent margin—stealing away three percentage points last year, alone. Google Cloud’s story is mixed. While the service saw 47% more year-over-year revenue, it still maintains a marginally larger operating loss compared to the prior year. It’s believed that Google’s background in AI will be a boon for future growth as related technologies become more mainstream.

No service is lagging behind because it’s objectively “poorer” compared to the others. AWS, Azure, and Google Cloud cater to different infrastructure setups. Additionally, AWS and Azure are both important pieces of their respective company’s revenue puzzle. Since Google Cloud operates at a loss, Alphabet as a whole is offsetting those losses in other profitable segments of their business. That’s not to say Google Cloud and Google Cloud Platform (GCP) aren’t important—they’re just more in wait-and-see mode than their peers are. However, CFO Ruth Porat is optimistic about the future. Google Cloud’s growth has outpaced the market recently.

Aws vs Azure vs Google Cloud Pricing Models

Each service prices itself differently, though variations these days don’t tend to be too wide. Standard pricing aside, some vendors offer occasional discounts to attract new customers or retain existing users. This has been Google’s calling card. The platform primarily offers two types of discounts: one for committed use, and the other for sustained use.

Google Cloud Pricing

The former ties discounts to a contract term—meaning you agree to use Google Cloud resources (specifically Compute Engine) for 1 to 3 years. This works primarily for VMs. Teams can purchase commitments pertaining to certain projects, or share these discounts across multiple projects. Meanwhile, sustained use discounts are automatically applied once you run resources beyond a certain time threshold. If you operate Google Cloud vCPU and memory frequently throughout a billing cycle, your overall costs are reduced. The same applies to GPU devices.

Otherwise, Google Cloud is a pay-as-you-go service. You’re charged for the resources you use as opposed to reserving certain capacities—which may be overkill or inadequate. Customers may also use over 20 products for free before hitting a specific usage limit.

AWS Pricing

Like Google, AWS offers a form of sustained use discount called a volume discount. Essentially, you’re charged a standard billing rate for resource consumption until you reach a threshold point. We can equate this to memory and compute resources. However, AWS also offers a suite of data management services like S3 and Data Transfer. You might be charged a certain rate for your first 10TB of transferred data. Amazon will then discount your per-GB rate once your data totals exceed that cutoff.

AWS also has a free tier that operates very similarly to Google Clouds. Generally, you may pay-as-you-go or reserve tiered capacity. The latter option, Amazon claims, can save you money via committed use.

Microsoft Azure Pricing

Like its peers, Azure offers reservation discounts. These allow you to save money by committing to 1 or 3-year contracts, and claiming a certain level of capacity in unison. Customers can also unlock unique savings thanks to Microsoft’s software ecosystem.

While AWS and Google can run Windows Servers or SQL Servers, it can be expensive. Conversely, you might save substantially with Azure if you’re already under the Windows umbrella. Azure specifically calls out AWS here in its comparisons. You can expect to save 71% when running Windows VMs via Azure over AWS EC2. Similarly, you might save 85% over AWS RDS for SQL Database Instances, or save 45% when managing SQL Server VMs. Overall, Microsoft claims that AWS is five times more expensive when running these workloads. That’s something to keep in mind.

Azure chiefly touts its pay-as-you-go model on its pricing page. This is expected and aligns well with its competitors’ offerings. Note that Azure has 21 product categories, each with multiple products. Many of these products have specific pricing models, which—while not uncommon—may result in fragmentation with complex use cases.

AWS vs Azure vs Google Cloud Pricing

What are some high-level observations?

  • AWS and Google Cloud are priced similarly for systems leveraging general-purpose or on-instance versions of cloud that are memory-optimized
  • Google Cloud is clearly the most costly on the compute-optimized side, whereas AWS and Azure draw nearly level with one another
  • Memory-optimized and accelerated resources seem exorbitantly expensive with Google, since it favors 40vCPUs and 12vCPUs over smaller 4vCPUs
  • For memory-optimized instances, Google Cloud has the most economical one-year commitment price
  • Google Cloud is much cheaper overall for compute-optimized cloud instances

All three providers utilize hourly and per-minute billing for their services. However, each has introduced per-second billing at a certain usage-time threshold. This structure is more granular than those which preceded it.

Quick Notes on Serverless

The serverless deployments with each service also vary in price. AWS and Azure are nearly neck-and-neck in terms of pricing, while performance-based charges make Google Cloud noticeably more expensive. However, this cost probably won’t break the bank. It might also make sense if those CPU MHz differences are meaningful to your workloads. Each provider offers serverless free tiers as well.

Best-Fit Deployment and Infrastructure Types

While it’s true that companies are moving away from traditional data centers, some elements of these on-premises deployments are appealing to technical teams. Making a complete cloud transition isn’t always simple. Consequently, many companies have opted for a hybrid-cloud environment—one which blends on-premises control and customization with virtual scalability. This approach allows teams to make critical apps and resources available anywhere, at any time. This is how many teams get their feet wet.

Immediately, this setup meshes well with Azure. Microsoft is an enterprise stalwart and has always respected the needs of data-center clients. If you’re adopting a hybrid approach, Azure’s tools and resources are extremely useful. Azure’s control plane works seamlessly with one’s on-site hardware and remote resources—like VMs and other instances. Identity management and network integration with existing on-premises infrastructure also simplify this transition. Azure Stack and Azure ExpressRoute are two tools enabling this. Finally, Azure has made notable strides in the edge computing space.

On the opposite end of the spectrum, AWS centers on the public cloud. While you can reserve capacity based on your needs, Amazon has customers who share the same resources. Networking devices, storage, and hardware are divvied up between tenants. Public clouds tend to be cheaper than their alternatives. Microsoft Azure also offers public resources, though this isn’t the platform’s sole focus. For networking purposes, Amazon does offer its own Virtual Private Cloud (VPC).

Meanwhile, Google Cloud’s Virtual Private Cloud (VPC) is one of the platform’s bright spots. This allows teams to scale diverse workloads, whether they’re regional or global. These VPCs are shareable and expandable with ease. Google Cloud also brings solid hybrid and multi-cloud application setups to the table. On the multi-cloud front, customers can simultaneously use two public cloud providers to tackle specialized workloads.

That said, how does availability compare by location? AWS operates 77 availability zones across 24 regions—found in a whopping 240 countries and territories. Azure is available in over 60 regions across 140 countries. Lastly, Google Cloud operates 73 availability zones in 23 regions, across 200+ countries. AWS seems to claim the crown here.

One caveat is in the edge computing realm, where Google holds a clear advantage. While AWS has 44 edge networking locations, Google Cloud boasts an impressive 144 locations. Azure doesn’t explicitly state how many total edge zones it maintains.

Why You’d Choose Amazon Web Services (AWS)

Overall, you simply cannot beat the vast number of tools offered under the AWS umbrella. If you’re working with numerous applications, workloads, and architectures (server vs. serverless), there’s almost certainly something for you. The AWS platform offers over 42 services and tools. These are categorized under the following:

  • Compute services
  • Storage services
  • AI and ML
  • Database services
  • Backup services
  • Serverless computing

While we can categorize Google Cloud’s and Azure’s tooling similarly, their overall breadth is lacking compared to AWS. While it might seem that AWS is throwing a lot at the wall to see what sticks, these services aren’t simple throwaways. The company has the resources to offer and develop them over time.

Additionally, AWS enjoys the greatest global reach at comparable prices to its competitors.

Why You’d Choose Microsoft Azure

If you’re dabbling in the cloud for the first time—or leading a modernization project across your physical infrastructure—Azure will be the logical choice. The solution is arguably the leader when it comes to hybridization, and especially since so many enterprises rely on Windows. Azure also supports open source add-ons for extensibility. Familiarity with Microsoft services and processes will help out immensely here.

Why You’d Choose Google Cloud

If your business is AI or ML-driven, you should look no further. While other providers have added support for libraries like TensorFlow, Google Cloud has integrated with this resource for quite some time. Google is placing plenty of eggs in this basket. That commitment signals that the company will support developers and other workloads in the data science space for years to come.

Otherwise, Google Cloud is built to support cloud businesses first—focusing less on data centers and “traditional” approaches to technology. Its tools are DevOps friendly (Cloud Functions, App Engine, and Kubernetes) and mesh well with open-source contributions. Finally, Google offers customers a complete container-based model for cloud computing.

Make a Holistic Vendor Choice

As you can see, the AWS vs. Azure vs. Google Cloud debate is far from black and white. You’ll need to consider the overall needs of your team and environment, then zero in on potential options from there. No matter what you choose, you’ll benefit from dedicated support and a bevy of technological expertise on the vendor side.

Hate ads? Want to support the writer? Get many of our tutorials packaged as an ATA Guidebook.

Explore ATA Guidebooks

Looks like you're offline!