
AWS vs Azure vs Google: Choose the Best Fit for Your Organization
The Infrastructure-as-a-Service (IaaS) marketplace has reached its tipping point in the wake of emerging realities such as remote working. Competition within the IaaS marketplace is now fierce, and it’s a three-way race: AWS vs Azure vs Google.
If you want to shift to the public cloud, it may be challenging to choose a provider that best suits your business requirements. Parallels® cuts through the complexities surrounding the IaaS marketplace and the debate about AWS vs. Azure vs. Google Cloud to help you make an informed choice.
First off, check out our comparison table for a quick overview.
In this article we compare the following features:
AWS vs Azure vs Google: Overview
Users may create apps using pay-as-you-go computer resources and services from Amazon Web Services (AWS) in a couple of minutes. For instance, you could rent a server from AWS and use it precisely like a real server, configuring it, protecting it, and connecting to it. The virtual server is distinct because it utilizes an AWS-managed planet-scale network.
Microsoft Azure‘s public cloud platform provides services for analytics, virtual computing, storage, networking, software as a service (SaaS), platform as a service (PaaS), and infrastructure as a service (IaaS), among other things. Your on-site servers might be upgraded or changed.
A package of cloud computing services called Google Cloud, formerly known as App Engine, was launched by Google in 2008. GCP provides platforms as a service, software as a service, and infrastructure as a service to businesses worldwide (SaaS). For instance, GCP focuses largely on providing services for creating and managing unique applications that may later be published from its hyper-scale data centers.
AWS vs Azure vs Google: Market Share
The market share of public cloud services is projected to reach more than 362.3 billion US dollars by 2022. Three of the biggest providers who make up this market are AWS, Azure, and Google Cloud. Mentioned below is the breakdown of the reports shared by these three providers in the year 2020.
AWS
Amazon Web Services (AWS) revenue grew by 29% from Q3 2019 to Q3 2020 ($8.9 billion to $11.6 billion). AWS holds 32% of the public cloud market share and generates more revenue than the combination of Azure and Google Cloud. AWS has had a jump start (released in 2006), which seems to have worked out well in its favor.
Azure
Microsoft Azure revenue grew 48% in the last quarter of 2020. Azure holds 19% of the market share. Having entered into the IaaS marketplace four years after AWS (2010), it has become second in line in the list of public cloud providers.
Google Cloud
Google Cloud generated $3.44 billion in revenue (45%) in 2020. Google Cloud holds 6% of the market share of the public cloud. Google Cloud was released well before Azure but has not managed to go beyond it when it comes to having a share in the Azure marketplace.
AWS vs Azure vs Google: Free Tiers
If you’re just starting to explore the IaaS technologies or have a limited budget, a free tier is a great starting point. While the free tiers may not be sufficient for full production work, they can help organizations get started with some of the IaaS services. There are two types of free tiers for AWS, Azure and Google Cloud: “Limited-time for free” and Always Free.
With the “Limited-time for free” tier, you get specific IaaS products upon registration or first sign-up. However, you can only use these products in limited quantities for a duration of up to 12 months. When this time expires, you are charged for the IaaS products at standard rates. When it comes to the “Always free” tier, you have free access to IaaS products, but you cannot exceed a set threshold in one month.
Below is a detailed comparison of AWS vs. Azure vs. Google Cloud free tiers:
AWS
Amazon, unlike Azure and Google Cloud, doesn’t offer credits. Some of AWS’s “Always free” IaaS products include:
- AWS CodeBuild (up to 100 build minutes per month).
- Amazon DynamoDB (a maximum of 25 GB of storage per month).
- AWS Lambda (up to one million requests and 3.2 million seconds of computing time per month).
- Amazon RDS (MySQL, MariaDB, PostgreSQL and Oracle Database).
- AWS Step Functions (4,000 state transitions per month).
- AWS CodeCommit (a maximum of five users with 50 GB per month).
- AWS CodePipeline (one active pipeline per month).
When it comes to the “Limited-time for free” tier, AWS has the following cloud services:
- 750 hours per month of Elastic Compute Cloud (EC2).
- One million API calls per month.
- 1 GB of Amazon Cloud Directory.
- 30 GB of Elastic Block storage.
- 5 GB of S3 storage.
- 40 hours of Amazon AppStream 2.0.
- Access to Machine Learning (ML) products such as Lex, Rekognition, Polly, Transcribe and Translate.
Azure
When you register or first sign up for an Azure account, you automatically get a US$200 credit to spend on Azure products within the first 30 days. Some of Azure IaaS “Always free” tier products include:
- Azure Cosmos DB (up to 400 provisioned throughput rack units per second with 5 GB of storage).
- Azure App Service (a maximum of 10 mobile, web or API applications with 1 GB storage).
- Azure Functions (up to one million requests and 400,000 GB of resource usage).
- Event Grid (a maximum of 100,000 operations for event publishing and delivery).
- Azure Active Directory (up to 50,000 stored objects with single sign-on (SSO) and Multi-Factor Authentication (MFA) to all cloud applications.
- Azure Service Fabric to implement microservice applications.
- Azure DevOps (first five users free).
- Azure Data Factory (up to five free low-frequency activities).
With the “Limited-time for free” tier, you get the following cloud services from Azure:
- 750 hours of Azure B1S general-purpose Virtual Machines (VMs) for Windows Server.
- 750 hours of Azure B1S general-purpose VMs for Linux OS.
- 5 GB of locally redundant storage.
- 250 GB of SQL Database storage.
- 15 GB of bandwidth for outbound data transfer.
Google Cloud
Google Cloud offers US$300 to first-time account holders. However, while Azure requires users to spend their credit within the first 30 days, Google Cloud allows for a spend period of up to 12 months. Some notable examples of services under Google’s “Always free” tier include:
- Google App Engine (up to 5 GB of Cloud Storage with 28 front-end and 9 back-end instance-hours per day).
- Google BigQuery (a maximum of 10 GB storage and 1 TB of querying per month).
- Google Cloud Build (up to 120 build minutes per day).
- Google Cloud Functions (up to two million invocations for both HTTP and background per month).
- Google Cloud Source Repositories (a maximum of five users with 50 GB storage).
- Google Cloud Storage (up to 5 GB of regional storage in the US with 5,000 Class A and 50,000 Class B operations).
- Google Compute Engine (one f1-micro VM in US regions).
AWS vs Azure vs Google: Security
When it comes to cloud security, cloud providers consider two factors: physical security, which involves protecting physical datacenters at geographic locations, and infrastructure security, which involves authentication and authorization, firewall security, data encryption, identity management, and cloud services protection. The table below highlights AWS vs. Azure vs. Google Cloud security aspects:
Security Service | AWS | Azure | Google Cloud |
Physical Security | Amazon has many highly-diversified data centers spread across the globe to ensure redundancy, availability, and capacity planning. | Azure uses 58 carefully-selected regions spread across the globe in 140 countries/regions to ensure sovereignty, data residency, resiliency and compliance. | Google Cloud has many data centers spread across 22 carefully-selected regions and 61 zones to avoid single failures and achieve data residency. |
Authentication and Authorization | Identity and Access Management (IAM) protocol | Active Directory (Azure AD) with Single Sign-On (SSO) support | OAuth 2.0 protocol with SSO support |
Firewall Security | Web application firewall | Application gateway | Application gateway |
Data Encryption | Key Management Service (KMS) | Storage Service Encryption (SSE) | KMS |
Identity Management | Amazon Cognito | Active Directory B2C (AD B2C) | Unified Management Console (UMC) |
Cloud Services Protection | Shield | Distributed Denial-of-Service (DDoS) protection service | DDoS protection service |
AWS vs Azure vs Google: Storage Comparison
The most common cloud storage technologies employed by AWS, Azure and Google Cloud include:
- Block storage. Block storage is a form of persistent disk storage and is used in conjunction with VMs. There are two forms of block storage: traditional magnetic-based Hard Disk Drives (HDDs) and contemporary Solid-State Disks (SSDs).
- Object storage. Object storage is an elastic and flexible format storage system designed for storing unstructured data within the cloud. There are three forms of object storage: Hot for accessing instantaneous data, Cool for infrequent data and Cold for archival materials in the cloud.
- File storage. File Storage is a relatively nascent cloud storage technology that resembles a traditional Network File System (NFS). With file storage, users can mount files easily to their VMs and read and access their records.
AWS, Azure and Google Cloud also employ various database services, including Relational Database Management Systems (RDBMS), NoSQL Key-Values and NoSQL Indexes.
Here’s a summary of AWS vs. Azure vs. Google Cloud storage offerings:
Storage Service | AWS | Azure | Google Cloud |
Cloud Storage Technology | |||
Block Storage | Elastic Block Store (EBS). There are 3 forms:
|
Managed Disks. There are 2 forms:
|
Persistent Disks (PDs). There are two forms:
|
Object Storage | Simple Storage Service (S3). There are 2 forms of S3:
|
Azure Blob storage. There are 3 categories:
|
Google Cloud Storage (GCS). There are 2 forms of GCS:
|
File Storage | Elastic File System (EFS). | Azure File Storage (AFS). | Lacks a native file storage offering. Uses Filesystem in Userspace (FUSE). |
Database | |||
RDBMS | Amazon RDS | SQL Database | Google Cloud SQL |
NoSQL Key–Value | Amazon DynamoDB | Table Storage |
|
NoSQL Index | Amazon SimpleDB | Azure Cosmos DB | Google Cloud Datastore |
AWS vs Azure vs Google: Processing Power
Elastic Compute Cloud (EC2) is AWS’s flagship infrastructure for scalable computing on demand, competing with Azure’s Virtual Machine Scale Sets and Google’s Compute Engine (GCE). The table below compares the host’s offers for EC2, Virtual Machine Sets, and GCE in terms of VMs and VM scalability:
Processing Feature | AWS (EC2) | Azure (Virtual Machine Set) | Google Cloud (GCE) |
Virtual Machines | |||
Clock Speed | 1.6 GHz to 3.3 GHz | 2.7 GHz to 3.7 GHz | 2.0 GHz to 4.0 GHz |
Maximum vCPUs | 128 | 128 | 224 |
GPU Acceleration | Yes | Yes | Yes |
Maximum vGPUs | 4 | 4 | 4 |
Maximum Memory | 244 GB | 208 GB | 448 GB |
Temporary Storage Limits | 48 TB | 3 TB | 4 TB |
VM Scalability | |||
Autoscale | Yes (via clone building) | Yes (via presetable group) | Yes (via clone building) |
Auto Resize | Yes | Yes | Yes |
Pricing
Pricing is tricky when contrasting AWS, Azure and Google Cloud. This is primarily because the costs adjust frequently, and the price models differ slightly. For a fair price comparison of the three, you need to understand their pricing schemes.
Also, you must familiarize yourself with how each provider defines its prices for on-demand VM instances, reserved VMs, and storage tiers. This table provides insights on essential price parameters that are directly comparable and those that are not when it comes to VMs:
Price Parameter | AWS | Azure | Google Cloud | Comparable? |
Rating Frequency | Pay-as-you-go and based on per-second billing, with a minimum of one minute | Pay-as-you-go and based on per-second billing, with no up-front costs | Pay-as-you-go and based on per-second billing, with a minimum of one minute | Yes. The pay-as-you-go scheme is uniform across all cloud providers. |
Instance Types/Machine Types | General-purpose, Compute- Optimized, Memory-optimized | General-purpose, Compute- Optimized, Memory-optimized. In addition, Azure provides Storage- optimized,
GPU-optimized and High-performance compute machine types. |
General-purpose, Compute- Optimized, Memory-optimized | Yes, AWS, Azure and Google Cloud have the same instance categories. |
On-Demand VMs | Price depends on the number of vCPUs and memory capacity | Price depends on the number of vCPUs and memory capacity | Price depends on the number of vCPUs and memory capacity | No. Cloud providers use a different combination of vCPUs and memory capacities during pricing. |
Reserved VMs | Discount of up to 72% for a one-year or three-year commitment. | Discount of up to 82% for a one-year or three-year commitment. | Discount for a one-year or three-year commitment. | No. Pricing is comparable only if you are paying monthly. |
As an example, let’s contrast AWS vs. Azure vs. Google Cloud using on-demand instances as a pricing parameter.
Instance Parameter | vCPUs | RAM | OS | AWS (per-hour price)* | Azure (per-hour price)* | Google Cloud (per-hour price)* |
General-purpose | 2 | 8 GB for AWS, 8 GB for Azure and 7.5 GB for Google Cloud | Linux (Ubuntu) | US$0.100 | US$0.117 | US$0.107 |
Compute-optimized | 2 | 3.75 GB for AWS, 2.0 GB for Azure and 1.8 GB for Google Cloud | Linux (Ubuntu) | US$0.100 | US$0.117 | US$0.813 |
Memory-optimized | 2 | 15.25 GB for AWS, 15 GB for Azure and 13 GB for Google Cloud | Linux (Ubuntu) | US$0.133 | US$0.0992 | US$0.134 |
* Prices are based on the cloud provider’s current on-demand pricing structure and are amenable to changes.
Ultimately, AWS appears to have a price advantage over Azure and Google Cloud for both general-purpose and compute-optimized on-demand instances. However, Azure has an overall advantage when it comes to memory-optimized applications.
Parallels RAS Supports the Leading IaaS Providers
There is no perfect, one-size-fits-all cloud IaaS provider for migrating an organization’s IT infrastructure to the cloud. It would be better if organizations focused on selecting the best-suited provider in terms of their business needs. This demands an effective multi-cloud strategy.
One way to adopt the best multi-cloud strategy is to ensure you run your workloads where they fit best. If you find Google Cloud is better-suited for security-intensive applications, then continue running those applications there. If you find Azure is appropriate for compute-intensive applications, then run your workloads there. On the other hand, if AWS is cheaper for general-purpose applications, there is no need to switch to another provider.
Whichever cloud provider you choose, Parallels® Remote Application Server (RAS) can help you achieve your bottom line. Parallels RAS is a multi-cloud-based VDI solution that delivers Windows desktops, applications and data to a broad spectrum of platforms. These include macOS, Linux, Chrome OS, iOS, Android and any other HTML5-ready browser.
Parallels RAS supports on-premises private cloud deployments, hybrid cloud deployments and public cloud deployments. Examples of where public cloud deployments are possible include AWS, Azure, and Google Cloud. With inbuilt security features like multi-factor authentication (MFA), granular filtering and client policies, among others, organizations can efficiently deliver virtualization to any device in any location.
Check what you can achieve with your favorite IaaS provider by downloading your FREE 30-day Parallels RAS trial!
Growth Opportunities for MSPs, ISVs, VARs and SIs in the Post-Pandemic Era
The Road from VAR to MSP: How to Successfully Transition from One-Off to Recurring Revenue