AWS & Azure & Google Cloud Pricing Models

Date: Oct 19, 2022

Related Tags：1. Get the most out of Alibaba Cloud with flexible pricing options
2. Why Alibaba Cloud

Abstract: Google Cloud, Microsoft Azure Cloud, and Amazon Cloud (AWS) offer hundreds of different products, each with their own service structures, technologies, and pricing models. This article attempts to make a comprehensive comparison of these three products.

Aws vs azure vs google cloud pricing overview

Amazon Cloud (AWS) is the world's leading cloud computing platform. It offers Infrastructure-as-a-Service (IaaS), Platform-as-a-Service (PaaS), and Software-as-a-Service (SaaS) offerings. Amazon Cloud (AWS) can provide users with on-demand computing power, storage, application services, and content delivery services aws vs azure vs google cloud pricing.

Microsoft Azure is Microsoft's public cloud computing platform. It provides cloud services such as computing, analytics, storage and networking. Users can develop and extend applications based on these services, or run existing applications on the public cloud. The Microsoft Azure cloud offers a rich platform-as-a-service (PaaS) offering and robust security features that can be integrated with Microsoft's own security products such as Azure Active Directory and Azure Defender aws vs azure vs google cloud pricing.

Google Cloud Platform is a public cloud computing service provided by Google. The platform includes various managed services for developing compute, storage, and applications that run on Google hardware. Google offers a simpler pricing model than other providers and lower prices for many service categories. In addition, it offers unique compute offerings, including industry-leading managed Kubernetes services and Tensorflow Processing Units (TPUs) for hosting AI workloads aws vs azure vs google cloud pricing.

Google Cloud, Microsoft Azure Cloud, and Amazon Cloud (AWS) offer hundreds of different offerings. Each product has its own service structure, technology and pricing model. Each cloud has thousands of possible deployment combinations. This can feel overwhelming, and cost management for these cloud vendors can be difficult aws vs azure vs google cloud pricing.

Fortunately, several cloud providers have pricing calculators and other aids that can help estimate and forecast costs. Timely use of these tools before and during migration is critical to containing costs.

Amazon Cloud (AWS) / Microsoft Azure Cloud/ Google Cloud Price Comparison

Price is one of the most important factors to consider when choosing a cloud platform provider. Since the three suppliers have different pricing models and discounts, it is difficult to compare them precisely. Here's a brief description of each vendor's pricing model:

Amazon Cloud (AWS): Amazon Cloud (AWS) has complex pricing for some modules, and they offer specialized tools like AWS Calculator, AWS Cost Explorer, and Trusted Advisor to help you estimate costs or get discounts. Meanwhile, Amazon Cloud offers deep discounts on its services through a variety of savings models, including Spot Instances, Reserved Instances, and savings plans.
Microsoft Azure Cloud: Microsoft Azure Cloud pricing is easier to understand. It shows the expense calculation section on the dashboard, so you can clearly see how much money you spend and where it is spent. The Microsoft Azure cloud provides a pricing calculator that makes it easy to estimate service costs, and a powerful enterprise budgeting system that lets you spread costs across departments or business units.
Google Cloud: In terms of the simplicity of the pricing criteria, Google Cloud stands out with its attractive and customer-friendly pricing strategy. It's trying to beat other cloud service providers through a price war.
All three providers have a free tier where you can try out their services before buying -- they even offer "free forever" tiers that offer limited service.

Amazon Cloud (AWS) Pricing Model

Let’s take a closer look at the Amazon Cloud (AWS) pricing model.

pay-as-you-go

The default pricing model for Amazon Cloud (AWS) is pay-as-you-go, based on actual usage per hour or per second. This is flexible, but also the most expensive option. Many customers start with this model and move on to other models as they learn more about their needs for cloud services.

Reserved Instances

Amazon Cloud (AWS) allows you to book instances for 1 or 3 years with discounts as low as 25% compared to pay-as-you-go. In the Reserved Instance model, unwanted Reserved Instances cannot be deleted (although Reserved Instances can be sold on a dedicated Amazon marketplace). To scale up, you need to buy more expensive pay-as-you-go resources.

While this reduces the flexibility of Amazon Cloud (AWS) services, you can still benefit from Amazon Cloud (AWS), which offers advanced automation options and a rich ecosystem of services. Most customers combine models to use Reserved Instances for long-running workloads and pay-as-you-go resources for more volatile workloads.

Auction instance

Spot Instances are available on Amazon EC2, Amazon Fargate, and some other computing services. It offers the best discounts, with discounts as low as 10% compared to pay-as-you-go. Spot Instances are reserved computing power bids on the Amazon Cloud (AWS) open market. The price changes every 5 minutes, and if your bid is higher than the current market price, you get a Spot Instance.

The problem is that when capacity becomes unavailable or the current Spot price exceeds your maximum bid, your Spot Instance is terminated with only 2 minutes' notice. Amazon Cloud has also introduced a new feature to send advanced warnings when Spot Instances will be terminated, but this does not guarantee that you will be notified and deal with instances in a timely manner. It also provides an advanced mechanism, Spot Fleet, that manages the scalability of Spot Instance groups and Pay-As-You-Go regular instances.

Microsoft Azure Cloud Pricing Model

In addition to the pay-as-you-go model, Microsoft Azure Cloud offers two main cost optimization strategies: Microsoft Azure Cloud Virtual Machines and Spot VMs.

pay-as-you-go

Microsoft Azure cloud services are billed by the second based on actual usage, with no long-term commitments or upfront fees. You have the flexibility to increase or decrease resources as needed. Microsoft Azure cloud virtual machines are automatically resized with autoscaling.

This pricing model is primarily aimed at users who like flexibility and want to convert capex to opex, and applications with variable or short-lived workloads.

Reserved Virtual Machine Instances (RVMI)

Reserved VM Instances in the Microsoft Azure cloud are pre-purchased VMs that need to be used for 1 or 3 years in a specific region. Reserved VM instances can be up to 72% cheaper than pay-as-you-go.

The Microsoft Azure cloud may replace one reserved virtual machine instance with another during the term. Users can also cancel Reserved Instances before the term ends, but early termination will incur a fee.

This pricing model is suitable for applications with constant load, customers with fixed budgets, or large applications that require constant use of a certain number of virtual machines (such as a central management component).

Bid VM

Compared to pay-as-you-go, Microsoft Azure Cloud allows you to purchase their spare computing power at a discount of up to 10%. However, Spot Instances can be interrupted by sudden notifications, so are only suitable for workloads that can tolerate interruptions. The Microsoft Azure cloud only provides a 30-second notice that a virtual machine is about to be disrupted.

The Microsoft Azure cloud offers Virtual Machine Scale Sets (VMSS). This is an autoscaling mechanism that allows you to automatically add Spot Instances based on a predefined strategy. Unlike Amazon Cloud's Spot Fleet, Virtual Machine Scale Sets cannot mix Spot VMs and Pay-As-You-Go VMs.

Spot VMs are primarily suitable for distributed fault-tolerant applications, stateless applications, and non-emergency or highly parallelized workloads.

Google Cloud Pricing Model

Google Cloud Platform offers the following pricing models.

pay-as-you-go

Google Cloud can pay as you go. This is great for individuals who don't plan to use cloud services continuously, as it allows you the flexibility to add or remove services as needed. This level of flexibility comes at a price, as the pay-as-you-go model has the highest cost per hour on Google Cloud Platform.

long-term commitment plan

If you plan to go to the cloud for the long term, you can use a long-term commitment plan, which will result in significant cost savings. Google Cloud offers a long-term pricing model with a choice of 1 or 3 years in advance. Google calls its program "Committed Use" and offers discounts as low as 30% off pay-as-you-go.

Preemptive instance

Preemptible VM instances (aka Spot Instances) are significantly cheaper (60-91% cheaper) than standard VM prices. However, if Compute Engine needs to reclaim compute capacity to other virtual machines, it stops (preempts) those instances with a 30-second notice. Preemptible instances use Compute Engine's spare capacity, so availability depends on Compute Engine usage.

If your application is fault tolerant and can tolerate instances being preempted, preemptible instances can significantly reduce compute engine costs. For example, batch jobs can run on preemptible instances, and if some of them stop, the work will be slowed down but not terminated. Preemptible instances can complete batch jobs without placing the workload of the batch job on an existing instance or paying full price for an additional regular instance.

However, Google Cloud preemptible instances offer significantly less management capabilities than Amazon Cloud (AWS) and Microsoft Azure Cloud, making them more difficult to scale and combine with pay-as-you-go instances .
in conclusion

In this article, I've listed the key differences in pricing models between Amazon Cloud (AWS), Microsoft Azure Cloud, and Google Cloud. I mainly focus on three main payment methods:

Pay-As-You-Go: Allows you to pay hourly or minutely to use cloud resources.
Reserved Instances/VMs: Allows you to pre-purchase compute resources for 1 or 3 years, saving up to 72% (depending on vendor).
Spot/Preemptible Instances/VMs: Lets you buy spare capacity for as little as 10% off the cloud provider's Spot Market, but with the risk that the instance may be terminated abruptly at some point.