What are the representative cloud services “GCP/AWS/Azure”?
Major cloud services such as Google Cloud Platform provided by Google, Amazon Web Services provided by Amazon, and Microsoft Azure provided by Microsoft are the mainstream.
However, each cloud vendor, including these three major clouds, not only provides a large number of functions, but also has become commoditized.
Therefore, in this article, we will compare the characteristics of the three major clouds.
Market share of GCP/AWS/Azure
Here, we will look at the trends in the market share of each of the three major clouds over the past two years.
According to Acrovision, in 2018 and 2019, AWS had a 2018 market share of 32.7% and a 2019 market share of 32.3%.
Azure had 14.2% market share in 2018 and 16.9% market share in 2019.
GCP had 4.2% market share in 2018 and 5.8% market share in 2019.
Features of GCP, AWS, and Azure
As mentioned above, the cloud service market is dominated by the three major clouds.
Therefore, when considering the introduction of cloud services, it is common to choose from these three. However, you may feel that all three big clouds offer similar services.
However, a closer comparison reveals that there are differences in service.So, let’s take a look at the differences between the three major cloud services below.
Two features of GCP
Since GCP is also used as a platform for operating services provided by Google, such as Gmail, YouTube, and Google Maps, it has a wealth of experience in operating services.
Below are two features of GCP.
GCP feature 1: Wide range of AI and machine learning related services
GCP mainly provides five AI and machine learning related services.First of all, “Learn with Google AI” is an AI and machine learning course service. The next “Google Colaboratory” is a free service that allows you to use an execution environment (Jupyter notebook) for the purpose of machine learning education and research on your browser.
The third “TensorFlow” is a library for machine learning, and the fourth “Cloud ML Engine” is a service for executing machine learning.
The fifth “Cloud AutoML” is a service for building machine learning models, which allows you to create your own image recognition model.
GCP Feature 2: Easy to analyze data
Google, which is famous as a search engine company, has great strength in big data analysis and has technology to process huge amounts of big data at high speed.
Since GCP can use this technology owned by Google as a cloud service, it can be said that GCP has an infrastructure that facilitates data analysis.
Two characteristics of AWS
AWS, which was launched by Amazon.com in July 2006 with its EC2 server, has a long history among cloud services and boasts a high market share in the industry.
Below are two features of AWS.
Feature 1 of AWS: Many users in the cloud service market
Amazon lists 10 reasons why people choose AWS.First of all, “zero initial cost / low price” and “continuous price reduction” are the reasons for price.
“Relief from sizing” and “Reduction of operational burden by managed services” are the reasons from an operational perspective, but “Agility not to miss business opportunities” and “Global expansion at any time” are reasons related to business development. .
“The latest technology can be used at any time” and “Improvement of development speed and elimination of individual skills” are the reasons from the perspective of technology development, but the other reasons are “Ensuring high security” and “24 hours a day, 365 days a year, in Japanese.” There is a reason for support.
AWS feature 2: Promotes many standard technologies
AWS provides many standard technologies in interfaces and APIs, so we promote application development based on them.And it’s backed by a Gartner Research evaluation.
The company has positioned AWS as a leader in the Cloud Infrastructure and Platform Services (CIPS) category in its 2020 Magic Quadrant, a research showing the relative positions of competitors in a given market.
And this CIPS is defined as a highly “standardized” and automated type of delivery that complements infrastructure resources (computing, networking, storage, etc.) with integrated platform services.
Two features of Azure
Sales of Azure have grown steadily since Microsoft launched the service in 2010, and as mentioned above, Azure has the second largest share of the cloud service market after AWS.
Below are two features of Azure.
Azure Feature 1: High affinity with Windows
Microsoft, which provides Azure, is working on Windows, which is used not only for personal PCs but also for many IT infrastructures such as corporate servers.
Since Azure is a service based on Windows, companies that have built Windows-based IT infrastructure can use it without any discomfort.
Therefore, it can be said that Azure is a very easy-to-use and highly compatible service for companies using Windows.
Azure feature 2: Easy integration with on-premises environments
With Azure Stack, you can extend Azure services and features to your environment of choice (datacenter, edge location, remote office, etc.).
As a result, administrators will be able to manage the public cloud environment and on-premises environment involved in development from a single screen.
Therefore, it can be easily linked with Active Directory in an existing on-premises environment, so it can be said to be the best platform for embodying a hybrid cloud configuration.
Seven comparisons when looking at GCP, AWS, and Azure from a multifaceted perspective
So far, we have looked at the market shares and characteristics of the three major clouds.
From here, we will compare the three major clouds from multiple perspectives, centering on the individual functions that they actually provide.
Comparison of GCP, AWS, and Azure 1: Comparison from the perspective of storage
For file storage, AWS and GCP use NFS, and Azure uses SMB as a protocol.
Regarding object storage, AWS and GCP are priced according to the frequency of data access, but Azure prices are subdivided according to the type and use of stored objects.
Regarding data transfer services, Azure Data Box Heavy and GCP Transfer Appliance are up to 1PB, but AWS Snowmobile can send up to 100PB using a semi-trailer truck.
Comparison of GCP, AWS, and Azure 2: Comparison from the point of view of computing resources
Regarding bare metal servers, Azure can only be used for SAP HANA. On the other hand, GCP services place physical servers in colocation spaces.
Regarding containers, each company provides Kubernetes, which is an OSS version of Borg, which was developed by Google as a container operation management tool, so the operation load of containers can be reduced.
In addition, AWS Fargate, Azure Container Instance, and GCP Cloud Run are serverless container platforms that enable rapid development.
Comparison of GCP, AWS, and Azure 3: Comparison from a network perspective
Each company provides a private environment for virtual networks, but GCP covers multiple regions with a single VPC.In addition, each company has a load balancer at L4/L7, but AWS ELB distributes the load of on-premises resources with a hybrid configuration.
Each company provides an authoritative DNS service, and name resolution can be performed for both public and private networks.
In addition, each company can use virtual servers and container infrastructure at no additional cost, and reduce the access load with a CDN set on the origin server.
Amazon VPN and Azure VPN Gateway provide site-to-site and server-to-client VPN, while Azure allows authentication using Active Directory.
Since GCP is only between sites, access to servers without global IP addresses is done via Cloud VPN or using a stepping stone server.
Comparison of GCP, AWS, and Azure 4: Comparison from the perspective of database services
For relational databases, each company comes standard with replication to multiple locations.Amazon Auroa has fault tolerance and self-healing functions for optimized cloud environments, and can automatically grow up to 64TB per instance.
On the other hand, Azure’s SQL Server Stretch Database is a service that allows you to extend your on-premises SQL Server to Azure.
Cloud Spanner is a serverless relational database management system that enables horizontal scaling while maintaining transactional consistency.
For data warehouses, AWS stores data in S3 and uses Amazon Redshift Spectrum to run SQL queries directly on that data. On the other hand, Azure’s Synapse Analytics can analyze with Spark in addition to SQL.
GCP’s BigQuery provides a data warehouse using a serverless method, making it easier to operate than other companies. You also get built-in machine learning capabilities.
Comparison of GCP, AWS, and Azure 5: Comparison from a security perspective
AWS provides SSL certificates for free for services that work with ACM (AWS Certificate Manager), but there is a charge for private CAs.
In addition, Azure provides managed certificate services through App Service and GCP through Cloud Load Blancing, but they do not have the expiration monitoring and automatic renewal functions of ACM.
For encryption key management, each company can cooperate with a managed HSM. Therefore, security can be improved without taking time to operate.
For network security services, each company can use DDoS protection and WAF functions in addition to the firewall attached to the VPC function.In that WAF, each company can not only use the managed rule set, but also customize the rules themselves.
Comparison of GCP, AWS, and Azure 6: Comparison from the perspective of applications and tools
For code management, each company allows unlimited repositories and is free for up to 5 users.Also, for CI/CD, each company can incorporate third-party tools for each process, but GCP can be used by importing existing Dockerfiles.
As for the IDE, Azure’s Visual Studio includes extensions to Azure, so users who build applications with it will be able to use that environment as-is.
On the other hand, GCP’s Cloud9 is a cloud-based IDE with a rich editor, debugging, and AWS CLI pre-installed, so you don’t need to build an environment.
Comparison of GCP, AWS, and Azure 7: Comparison from the perspective of fee structure
The fee system is a pay-as-you-go system in which you pay only for the amount used by each company, but the first 1GB is free for AWS and the first 5GB is free for Azure.
However, when comparing the transfer charges when using 10TB and when using 40TB, the transfer charges for both are almost the same. GCP is about 20% more when using 10TB, and about 30% more when using 40TB.
Regarding the instance fee, if we select an instance with a similar number of vCPUs and memory amount from the burst type and compare it by hourly unit price (1 dollar = 109 yen conversion), AWS’s t3.small and Azure’s B1MS are both 2.9648 yen/hour. GCP’s g1-micro is 3.5098 yen/hour.
However, GCP automatically applies sustained use discounts. Also, Azure has a one-year reservation discount and AWS has a reservation discount called Reserved Instances. Therefore, this ranking may change depending on the period and discount rate compared.
Three Future Prospects for GCP, AWS, and Azure
Competition continues to intensify among the big three clouds to meet the needs of development and operations teams, such as growth analytics, and the use of datasets when building new applications and systems.
Here, we will look at the future prospects of the three major clouds.
Outlook for GCP/AWS/Azure 1: GCP
Although GCP lags behind AWS and Azure, its growth rate through 2019 is very high, growing by 87.8%.
GCP continues to develop new business customers and build a network of channel partners.
In order to further increase market share in the future, it will be necessary not only to refine our expertise in big data analysis and machine learning technology, but also to improve the disadvantages of the lack of Japanese region and Japanese information.
Outlook for GCP/AWS/Azure 2: AWS
AWS has been provided in the cloud so far, but in 2018 we announced a service called AWS Outposts that realizes the same service as the cloud even in an on-premises environment.
And in December 2019, Outposts began to be used in Japan as well. Therefore, AWS would like to expect this to further expand its market share.
Outlook for GCP/AWS/Azure 3: Azure
Azure has three strategies for the 2021 fiscal year: “cloud-native application development and modernization of existing applications,” “migration of existing applications to Azure,” and “standardization of cloud adoption processes and cultivation of CCoE culture.”
The first strategy is to work with a wide range of partners, especially in specialized fields such as data analysis.
The second strategy is to migrate data center servers to “Azure Stack HCI” (a service for hyperconverged infrastructure) with OEM hardware partners and SI partners.
The third strategy aims to provide Microsoft’s Cloud Adoption Framework and expand the specialized team (CCoE) that promotes the digital transformation of client companies.
Furthermore, in the third strategy, we are renewing the “ESI (Azure Enterprise Skill Initiative)” to improve Azure proficiency and providing free online courses through the AI business school “Education with AI Business School”.
We are also planning to deploy a virtual facility “Virtual Azure Base” on the Internet.
Understand the differences between GCP, AWS, and Azure
So far, we have seen comparisons and prospects for the three major clouds, AWS, Azure, and GCP, in terms of market share, features, and individual functions. All cloud vendors seem to have equivalent functions, but there are differences depending on the environment that can be used and the services that are linked.
And if you don’t know the difference, it will be difficult to get the performance you want.Therefore, make sure you understand the differences between GCP, AWS, and Azure.
