The adoption of cloud computing has quickly become a key driving force for businesses today, as applications are moved out of on-premise data centres in a bid to cut costs and increase agility.
Infrastructure-as-a-service (IaaS) is a model where a third-party provider hosts and maintains core infrastructure, including hardware, software, servers and storage on behalf of a customer. This typically includes the hosting of applications in a highly scalable environment, where customers are only charged for the infrastructure they use.
Early concerns over security and data sovereignty have largely been addressed by the big three public cloud vendors: Amazon Web Service (AWS), Microsoft Azure and Google Cloud Platform, with only the most heavily regulated businesses lagging behind in terms of adoption.
This has fuelled a crowded IaaS market, worth a total of $31 billion in 2018, up from $23.6 billion in 2017 according to research firm Gartner's latest figures.
It is a market that has been dominated by AWS since day one, back in 2006. Synergy Research Group's figures for Q1 2019 have AWS as the clear market leader globally when it comes to public IaaS and PaaS service revenues at 38 percent market share, which is down 2 percent quarter-on-quarter, followed by Microsoft at 18 percent, Google at 9 percent and Alibaba at 6 percent, all up a point respectively.
Despite AWS’s dominance, Microsoft has quickly gained ground under the leadership of "cloud first" CEO Satya Nadella, building a huge global cloud network of its own. Microsoft's recent move to increase the prices of only on-premise Office 2019 packages by 10 percent creates a clear incentive for customers to adopt a cloud-first attitude too.
Then there is the internet giant Google, which has been busy building out its public cloud services and IaaS business under the Google Cloud Platform (GCP).
So what separates the big three western cloud providers? And how can you start to decide which IaaS platform is best suited to your organisation?
Features and services
Selecting one cloud over the others will come down to the wants and needs of each individual customer and the workloads they are running. It is often the case that organisations will use multiple providers within different parts of their operations, or for different use cases, which is called a multi-cloud approach.
However, there are a number of differentiating factors that separate the approaches of the three firms, which can help end users consider which is right for them.
AWS, Microsoft Azure and Google Cloud Platform offer largely similar basic capabilities around flexible compute, storage and networking. They all share the common elements of a public cloud: self-service and instant provisioning, autoscaling, plus security, compliance and identity management features.
All three are investing heavily in their cloud services, and have sizeable parent companies to do so. This has resulted in more mature analytics offerings. For example, support for Hadoop clusters are provided by AWS (Elastic Map Reduce), Azure (HDInsight) and Google (Dataproc).
AWS still offers the largest depth of services with 140 across compute, storage, database, analytics, networking, mobile, developer tools, management tools, IoT, security and enterprise applications. But it has been around the longest.
All three vendors have added machine learning tools and a number of features targeted at cutting edge technology areas like the Internet of Things (IoT) and serverless computing (Lambda for AWS, Functions with Azure and Google), while customers can tap either cloud to variously build a mobile app or even create a high performance computing environment depending on their needs.
Naturally, all three vendors are strong in machine learning as they can draw on deep wells of internal expertise.
AWS launched the Amazon Machine Learning service in April 2015 to help developers create machine learning models. Then in 2016 it announced three new machine learning services for image recognition (AWS Rekognition), text to speech deep learning models (Polly) and the engine that powers Alexa (Lex).
Google offers a Cloud Machine Learning Engine, which helps machine learning engineers build models based on its open source TensorFlow deep learning library. Google also offers a whole host of off-the-shelf APIs for things like natural language processing, translation and computer vision.
Microsoft's Azure Machine Learning Studio allows specialist developers to write, test and deploy algorithms, as well as a marketplace for off-the-shelf APIs.
The recent buzz around containers is catered for as well, with all three providers supporting popular services like Kubernetes (which emerged from Google), service mesh and Docker.
All three providers take a pretty open approach to partnerships, allowing customers to run various apps and services in their cloud environments.
Google, for example, has announced a range of key partnerships with established vendors like SAP, Pivotal and Rackspace. Read next: Google courts enterprise customers with SAP, Pivotal and Rackspace partnerships
For UK customers worried about data sovereignty, AWS launched its UK region in December 2016, with Microsoft and Google quickly following suit.
Compute, storage, databases and networking
For compute, AWS' main offering is its EC2 instances, which can be tailored with a large number of options. It also provides related services such as Elastic Beanstalk for app deployment, the EC2 Container service, AWS Lambda and Autoscaling.
Meanwhile, Azure's compute offering is centred around its Virtual Machines (VMs), with other tools such as Cloud Services and Resource Manager to help deploy applications on the cloud, and its Azure Autoscaling service.
Google's scalable Compute Engine delivers VMs in Google's data centres. They are quick to boot, come with persistent disk storage, promise consistent performance and are highly customisable depending on the needs of the customer.
All three cloud providers support relational databases - that's Azure SQL Database, Amazon Relational Database Service, Redshift and Google Cloud SQL) - as well as NoSQL databases with Azure DocumentDB, Amazon DynamoDB and Google Bigtable.
AWS storage includes its Simple Storage (S3), Elastic Block Storage (EBS), Elastic File System (EFS), Import/Export large volume data transfer service, Glacier archive backup and Storage Gateway, which integrates with on-premise environments.
Microsoft's offerings include its core Azure Storage service, Azure Blob block storage, as well as Table, Queue and File storage. It also offers Site Recovery, Import Export and Azure Backup.
All three typically offer excellent networking capabilities with automated server load balancing and connectivity to on-premise systems, which brings us to...
One growing trend amongst the hyper-scale public cloud providers in the past year or so has been a growing focus on helping to serve customer's hybrid and multi-cloud needs. This tends to mean where they are deploying across multiple vendors' infrastructure and also need to maintain some applications on-premise. Vendors have responded with a range of solutions to help serve these customers who aren't ready to jump all-in on public cloud just yet, which is of course the majority of large enterprises.
Microsoft has long been the go-to option for hybrid deployments amongst the big three with its well-established Azure Stack. This provides customers with the hardware and software required to deploy Azure public cloud services from a local data centre with a shared management portal, code and APIs for simple interoperability.
AWS signalled its first serious move into hybrid deployments at re:Invent in 2018 with the launch of Outposts, a fully managed service where the vendor delivers pre-configured racks to your premises where AWS services can be run as though it were in their data centre.
Google then made its own push into hybrid in 2019 with the release of Anthos, which is effectively a rebranding of the Google Cloud Services platform and brings together a combination of the existing Google Kubernetes Engine (GKE), GKE On-Prem and the Anthos Config Management console. This promises unified administration, policies and security across hybrid Kubernetes deployments.
Speaking about Anthos at the Google Cloud Next conference in April 2019, new CEO Thomas Kurian said: "Many large companies truly want to train the people once on a set of technology that they can deploy anywhere. None of the other cloud providers are solving that. Today, if you talk to Azure they will say you can run Azure Stack on-premise and on the cloud, Amazon will say you can run Outposts on-premise and in the AWS cloud. They are fine companies, but they're not solving the multi-cloud problem."
Pricing can be a huge attraction for those considering a move to the cloud, and with good reason: there has been a continued downward trend on prices for some time now as the big providers compete.
In general terms, prices are roughly comparable, especially since AWS shifted from by-the-hour to by-the-second pricing for its EC2 and EBS services in the Autumn of 2017, bringing it into line with Azure and Google.
However, making a clear comparison can be tough as all three offer slightly different pricing models, discounts and make frequent price cuts.
All vendors offer free introductory tiers, allowing customers to try their services before they buy, and typically offer credits to attract innovative startups onto their platforms as well as 'always free' tiers with strict usage limits.
For example, Google offers free usage up to 1GB of Google Cloud Datastore capacity, 28 instance hours per day for Google App Engine, one micro sentence per month for Google Compute Engine, 5 GB-months of Google Cloud Storage (regional only), 2 million Cloud Functions per month, 50GB of logs with Stackdriver for monitoring, as well as limited access to products like: Google Cloud Natural Language, Cloud Vision API, Kubernetes Engine and more.
In July 2018, Microsoft announced changes to the pricing plans for its on-premise and cloud products that came into effect in October 2018. These changes include setting one price for web direct for online services (OLS) and eliminating programmatic volume discounts in Enterprise Agreement (EA)/EA Subscription, MPSA, Select/Select Plus, and Open programs (Open, Open Value, Open Value Subscription).
A high-profile user base may not be the main reason for choosing your cloud provider, but it can help more cautious organisations understand how the public cloud is benefiting others in their sector.
This is clearly a strong point of AWS. It has increasingly taken on large customer deals. For example, although the US Central Intelligence Agency eventually signed a contract with IBM, it awarded AWS a contract to build its private cloud in a one-off deal in 2013, which could be seen as a symbolic moment for potential buyers.
A longstanding AWS customer is Netflix, which eventually decided to shut all of its data centres in a final move to the cloud in 2016. But aside from web pioneers, AWS has been truly successful in convincing more traditional businesses to move to the cloud.
Other major customers include: AstraZeneca, NewsCorp, AirBnB, Aon, Channel 4, Financial Times, Dow Jones, Kurt Geiger, Lonely Planet, Nasdaq, Nike, Nisa Retail, Pfizer, and the Royal Opera House. A full list of AWS customers can be seen here.
Microsoft perhaps has fewer high-profile Azure users, with most of the messaging from the vendor appearing to be around its widely used software-as-a-service (SaaS) tools. But the Redmond firm has also notched up some notable customer wins such as Pearson, Ford, NBC News and Easyjet, to name but a few.
In a bid to turn this around Microsoft cut around 10 percent of its global sales force in July 2017, as part of a broad reorganisation to focus on selling its cloud services under the Azure brand.
The new selling strategy at the company was revealed in a leaked email, which was obtained by the Wall Street Journal. In it, Judson Althoff, executive vice president for worldwide commercial business, outlined how Microsoft wants to focus on targeting businesses instead of specific industries or market segments. He said he wanted to increase the "technical depth and better align sales and services to solution areas" at the company.
Google is in a similar position, but has notched up some key wins in recent years, especially when the streaming giant Spotify completed an all-in migration to GCP in the summer of 2018.
UK bank HSBC has also opted for Google Cloud for its analytics and machine learning capabilities. However, HSBC is taking a clear multi-cloud approach, partnering with all three providers for different workloads.
Snapchat parent company Snap also spends a great deal with Google for IaaS, but it also spends with AWS. During the social network's IPO process it was revealed that the company is committed to a $2 billion five-year deal with Google for cloud services, as well as a $1 billion deal with AWS over five years.
Home Depot and Disney were also named as Cloud Platform customers during Google's 2017 Cloud Next conference.
AWS pros and cons
As mentioned before, the reasons for picking one vendor over another will differ for each customer. But there are aspects of the competing clouds that will offer benefits in certain circumstances.
The breadth and depth of the AWS offering is seen as a plus for AWS.
AWS had a head start on the competition, building out its suite of cloud services since 2006. All of these are built to be enterprise-friendly so that they will appeal to CIOs as well as its core audience of developers.
The vendor ranks highly on platform configuration options, monitoring and policy features, security and reliability. Its partner ecosystem and general product strategy are also seen as market leading, and its AWS Marketplace has a large number of third-party software services.
Another of the benefits of the AWS cloud is its openness and flexibility. For example, Transport for London - which also relies on Azure in other parts of its operations - has used AWS to meet spikes in demand for its online services such as its Journey Planner tool.
However, one area AWS falls short to some degree is with its hybrid cloud strategy, where it has tended to be dismissive of the benefits of on-premise private clouds. As outlined above, the vendor is slowly coming around to the idea though.
Another downside to AWS is the scale of its offering. While this is an attraction in many senses, it can be difficult at times to navigate the large numbers of features that are on offer, and some see AWS as being a complex vendor to manage.
Microsoft Azure pros and cons
The big pull for Azure is where Microsoft already has a strong footing within an organisation and can easily play a role in helping those companies transition to the cloud. Azure naturally links well with key Microsoft on-premise systems such as Windows Server, System Center and Active Directory.
In addition, while both AWS and Azure have PaaS capabilities, this is a particular strength of Microsoft's.
One of the downsides, however, has been a series of outages over the years. Gartner analyst Lydia Leong has recommended considering disaster recovery capabilities away from Azure for critical applications hosted in the cloud. AWS isn't immune to downtime, though, suffering a major S3 outage of its own in March 2017.
As part of its 2017 IaaS global Magic Quadrant, Gartner states that its clients have had issues with "technical support, documentation, training and breadth of the ISV partner ecosystem" - but the company has been steadily working on these areas.
Whereas AWS provides users with many options for supporting other platforms, Azure can be somewhat restrictive in comparison. If you want to run anything other than Windows Server then Azure might not be the best solution, but Microsoft has been willing to embrace open source platforms, if a little slowly. For example, the company has been busy extending its support for Linux operating systems in 2017.
Google Cloud Platform pros and cons
Google has a good track record with innovative cloud-native companies and has a solid standing in the open source community, but has traditionally struggled to break into the enterprise market.
Its go-to-market strategy has been focused on proving itself on smaller, innovative projects at large organisations, rather than becoming a strategic cloud partner. Increasing the breadth of its partnerships and supporting pre-cloud businesses and IT processes will need to become focus areas if it wants to attract more traditional enterprises.
This could all be set to change now with the arrival of new CEO Thomas Kurian in November 2018.
"We have a clear vision of what we want to offer customers in a number of industries who are going through digital transformation," he said on stage during his first Google Cloud Next conference in June 2019. "We want to give them global scale, distributed, secure infrastructure; a digital transformation platform that that helps people build innovative digital transformation solutions, and then, industry specific-capability for digital transformation in a number of industries."
The ex-Oracle executive also laid out his plans to grow "our go-to-market function and our work with partners to deliver the right solutions for customers, which means adding people in sales and customer service and customer engineering, which is our technical teams, as well as hiring people with deep industry backgrounds."
He also laid out a commitment to openness, mirroring his rival CEO at Microsoft, Satya Nadella, by saying: "Our goal has always been simple: you have to have the best technology and the easiest to adopt solutions that give customers choice. If you do, they will choose you, if you don't, eventually they will be unhappy."
The company should continue to bet big on its machine learning tools though, with the firm's internal AI expertise and popular TensorFlow framework as selling points in what is set to become a key battleground.
It has also proved itself more than an AWS copycat, launching innovative features in the machine learning space as well as its BigQuery analytics engine, and the Cloud Spanner distributed database.
It is also worth noting that Google has the smallest footprint of global instances of the big three.
In very broad terms, AWS continues to lead the way in terms of offering the widest range of functionality and maturity. It continues to be the clear market leader, but the gap is closing.
Its expansive list of tools and services, along with its enterprise-friendly features make it a strong proposition for large organisations. Meanwhile its huge and continuously growing infrastructure provides economies of scale that enable aggressive price cuts.
But it appears that Microsoft has started to bridge the gap between the two, and will continue to do so with its ongoing investment in building out the Azure cloud platform and further plans to strengthen ties with its on-premise software.
For organisations already heavily invested heavily in Microsoft in terms of technology and developer skills - of which there are undoubtedly many - Microsoft Azure will continue to be a strong proposition.
Then there is Google, which could prove a more serious enterprise competitor under its new leadership. It was already making good progress with certain customers, especially with its Kubernetes and machine learning expertise, but has much more work to do to prove itself a viable enterprise option.