AWS, Azure and GCP Pricing: The True Cost of DevOps

Cloud computing has rapidly become a strong driving factor for companies worldwide, as software is transferred out of in-house data centers in an effort to modernize, reduce costs, and boost agility. Businesses more and more use it as an all-in-one solution, a model in which a third-party supplier comprises and manages a customer’s fundamental infrastructure.

Among the most used and popular DevOps services, namely Amazon Web Services, Azure DevOps services, and Google Cloud services, there is a battle going on in the market. Based on Statista analytics, Amazon Web Services, the most prominent provider in the cloud computing industry, held 32% of the total market in the 3rd quarter of 2021. Microsoft Azure comes in 2nd place with a 21% market share, followed by Google Cloud with an 8%. Therefore, in the 3rd quarter of 2021, these three cloud suppliers are undoubtedly leading within the statistics.

Since all three solutions are actively used by the public and are in demand when developing software you might be overwhelmed by the choice you are facing. Every DevOps services have their pricing plans, estimation of which depends on the type of application, the features it requires, and many other factors. Fortunately, all the DevOps services providers have a pricing calculator, which facilitates the search of products. [Find it here: Amazon Web Services, Azure DevOps services, and Google Cloud services].

So, to make it clearer for you, we conducted research, highlighting the most needed DevOps services for MVP development. Such a set can be called a basis of any application, basically a specific minimum you might need for your future software launch. This ‘basic package’ includes storage service, compute cloud, database, cache, hosting, and content delivery service. Each and every listed service remain on all three cloud computing platforms in different formats, with different names. That is why we selected the most similar technologies by functional abilities and characteristics, in order to compare their pricing in each service category. Besides, to gratify major businesses, we determined the main targeted locations: US West (North Virginia), US East (California), UK (London), to show the variation depending on location. Hence, let’s start!

Storage services

The most standard storage service, that is quite popular within Amazon Web Services is S3 (Simple Storage Service) for object archiving. An equivalent to it is Cloud Storage on Google Cloud, and Block Blob Storage on Azure. Such storage would be enough for simple app development. A little bit about each service:

• AWS S3 – an object storage service that provides scalability, accessibility of information, reliability, and efficiency.
• Azure Blob Storage – a highly scalable and protected object storage service designed for cloud-based services, repositories, elevated computing, and machine learning.
• Google Cloud Storage – object storage fitted from small to large businesses. As an advantage, it offers an unlimited quantity of data that may be stored with the possibility to get it unlimited times.

For the price estimation, we defined the minimum characteristics necessary for MVP running.

Measurement Specifications

• Storage amount: 200 GB per month
• PUT, COPY, POST, LIST requests: 10’000
• GET, SELECT, and all other requests: 10’000
• Data returned 200 GB per month
• Data scanned 200 GB per month

With mentioned amounts, we calculated the minimum sum, that you will need to pay per month for your MVP. For S3 the price in different regions varies, while remaining somewhere around 5.5$: UK – 5.47$, US East – 5.20$, US West – 5.87$. Using Cloud Storage costs equally for all outlined locations 4.65$ per month. The situation with Block Blob Storage, a part of Azure DevOps services, stays the same for US West and East – 3.79$, while for the UK this amount is a little bit higher: 3.97$.

As we can see from the comparison above, Azure with its Block Blob Storage provides the cheapest solution among other market giants.

Cloud Virtual Machines

Cloud computing has seen a significant rise in the tech market sector, with leading brands such as AWS, Microsoft Azure, and Google offering a variety of Virtual Machines. This implies that the VM instances might be viewed as a real machine with its CPU, storage, connection, backup, and so on. Cloud computing is the first to be paid for, during the development, so this is a crucial part of the process.

Azure’s service is called Virtual Machines, not to confuse anyone seeking such a solution on their website. On Amazon Web Services there is EC2, and on Google Cloud – Compute Engine. Let’s take a glance at their functionalities.

The Azure VM delivers virtualization adaptability without any need to maintain the actual hardware that runs it. However, it must be maintained by completing operations such as customizing, debugging, and setting up software. Azure VM enables to build a PC with the particular specifications required to write and test an application.

EC2 service is primarily intended to make web-scale cloud computing easy for developers. This also allows you to purchase and customize capacity with minimum effort, as well as have complete control over your computer resources. Furthermore, EC2 offers a range of processing units, memory, networking, operating system, and purchasing model options.

Compute Engine is a safe and customizable compute solution that allows you to create and run virtual machines on Google’s infrastructure. Google Compute Engine provides one-of-a-kind services, including predefined machine types for executing pre-built and ready-to-use setups, and confidential computing that is used to encrypt your most sensitive data while it is being handled.

All three services’ setup required little different parameters to enter. Have a look at the core we have determined.

Measurement Specifications:

• Operating system: Linux
• vCPUs: 2
• GPUs: NA
• Memory (GiB) – 4
• 1 instance
• 30% inst. for usage per month
• 20GB storage per instance
• Server running: 8 hours per day, 7 days per week

The most expensive prices are for EC2: US West- 14.49$, US East – 12.16$, UK – 13.71$ per month. Compared to AWS, slightly cheaper is Azure VM, with 12.05$ for US West, 10.11$ for US East, and 11.42$ for the service running in London. At the same time there is Compute Engine, with somewhat more acceptable pricing: US East- ​​9.62$, US West – 10.27$, UK – 10.98$.

With the range of possibilities EC2 and Azure VM bring, even though they are less affordable, they are both worth considering purchasing, the same as Google Compute Engine.

Database services

The number of relational database services available in the public cloud is continually increasing. In the future, we could expect to see ongoing updates and upgrades to these services, as well as price cuts caused by cloud companies’ fight for market share. But, whenever new database services enter the market, there still will be reliable time-tested solutions. So here goes the comparison of market giants: Amazon RDS for MySQL, Google Cloud SQL for MySQL, and Azure Database for MySQL.

Azure Database for MySQL is a managed database solution for software deployment and implementation that enables you to set up a MySQL database in minutes and grow on the fly – all on the cloud you know and trust.

Amazon RDS provides you access to the functionality of a well-known database engine such as MySQL, Oracle, or Microsoft SQL Server. This implies that you may use the same code, apps, and tools you’re already using with your existing databases with Amazon RDS. Amazon RDS automatically fixes the database server and supports it, maintaining backups for a user-specified time duration.

Google Cloud SQL is a robust database that is quick, doesn’t run out of capacity, and provides your app with redundant, dependable storage. With it, MySQL databases can be easily implemented on the cloud.

First, take a look at the parameters we set up, before prices rival.

Measurement Specifications:

• 1 instance
• vCPU: 2
• Memory: 4 GiB
• 1005 utilization
• Multi-AZ deployment
• 100GB storage
• 500GB Backup storage

Google Cloud SQL earned a leading position in this battle with US East price – 146.05$, US West – 163.67$, and the UK – 163.75$. Marginally more expensive appears to be Azure Database for MySQL: for US East – 227.25$, US West – 189.40$, and for the UK it is 219.68$. A middle ground here is Amazon RDS with the following pricing: US West – 203.58$, US East – 169.78$, UK – 187.56$.

Cache services

Redis as a storage system can function as a basic database, a cache, or a sub-system. Managed Redis services are available from Amazon Web Services (ElastiCache), Microsoft Azure (Azure Cache for Redis), and Google Cloud Platform (Memorystore for Redis). The offered versions and features differ from one supplier to the next.

ElastiCache boosts online application speed by letting you obtain data using efficient, controlled in-memory caches rather than slower disk-based databases. ElastiCache works with Memcached and Redis. But for the proper comparison, we estimated the price of Redis integration, which also has Google Cloud services, and Azure DevOps services.

Azure Cache for Redis offers a low-cost way to increase the read and write throughput of your data layer. By using a cache-aside technique, you may save and distribute database query results, session states, static content, and more.

Cloud Memorystore for Redis is a fully managed in-memory data storage solution built on Google’s infrastructure. Cloud Memorystore may be used to create application caches that give sub-millisecond data access, and it is Redis protocol compatible, allowing for simple conversion with no code modifications.

The parameters for estimation, in this case, are limited. Besides, each service provides a different minimum cache size. That is why we computed beyond the defaulted pricing calculators, how much it will cost to purchase for 500 MB Cache.

Measurement Specifications:

• 1 instance
• 730 hours
• ~500 MB cache

The prices are considerably different. The low-cost variant is the AWS service – ElastiCache. It will cost US West – 16.06$, US East – 12.41$, and the UK – 13.87$. On the other hand, Azure Cache for Redis offers the most high-valued service: US West – 32.12$, US East – 32.12$, UK – 40.88$. In the middle remains Memorystore for Redis: US West – 20.80$, US East – 17.88$, UK – 32.85$.

Even though the value range varies a lot, the functionality of all services is quite different.

Hosting

Each user’s visit to a website begins with a series of DNS requests. Slow page loading on the customer’s browser might be caused by poor DNS performance. Cloud-based DNS services improve DNS performance, resilience, and scalability, allowing you to provide better customer experiences throughout the world. Hence, it is an important part of development.

DNS is available on Azure and Google Cloud with the corresponding names, while on AWS it is called Route 53. Comparing this trio, there is a huge difference in multiple criteria. Route 53 supports the major part of possible platforms and OS. However, Google Cloud DNS and Azure DNS are limited in these terms. But when it comes to managed features, Google Cloud DNS and Azure DNS definitely win the race. Despite this, you had better choose the cheapest solution, if we are talking about the basic featured MVP.

Measurement Specifications:

• 1 hosted zone
• 100 000 queries

With such constrained parameters, we noticed that the price doesn’t differ depending on the location. Furthermore, Google Cloud and Azure even do not have such specifications among other characteristics. Thus, the comparison is facilitated. Route 53 and Azure DNS cost equally 0.54$ per month, whereas Google Cloud DNS is only 0.24$.

Content Delivery Network

The CDN plays an important role in ensuring that the data is transmitted to the user as promptly as possible. A CDN’s operation is influenced by three major aspects: the user’s location, the location of the hosting server, and the location of the content delivery server. Thus, we picked up the US Zone as a target, as it is a highly popular destination.

Dividing into CloudFront (AWS solution), Google Cloud CDN, and Azure CDN, they are all secure and solid providers to choose from.

CloudFront provides a developer-friendly environment that allows them to create features to fit specific needs. What makes this service a leader in providing content delivery networks, is that it is internally integrated with the other Amazon Web Services. This is making it even more flexible during workflow when the urge for additional products appears.

Moving to Google Cloud CDN, we need to mention that sites that utilize it, do not need a regional DNS because they receive a single IP address that can be used anywhere. As the other benefit, Google Cloud CDN secures site content with SSL/TLS and there are no additional fees for this service.

The third option is Azure CDN, which enables users to reduce load times, increase performance, and conserve bandwidth. With its characteristics, it could be a perfect solution for controlling apps, websites, IoT endpoints, streaming games, and other media.

Amazon CloudFront charges for data transmission out and HTTP/HTTPS requests. We selected 1 region for estimation: the US. Selecting other regions would depend on the software specificity.

Measurement Specifications:

• 3 GB transfer
• US Zone

Even though AWS offers the possibility to define your location, the cost for the selected areas doesn’t differ and remails 0.29$. Meantime, Google Cloud and Azure again don’t require specification for that criteria and set the same price: 0.24$.

Find more information about the other services’ alternatives here.

To Sum Up

Clearly, in total, the ‘basic package’ is significantly more cost-efficient with Google Cloud services. Considering, that the market statistics show that it backs down its alternatives. But with such pricing and the variety of functionalities it provides, Google Cloud services might become a more significant corporate rival. But there is a long way to go before it can be considered a viable enterprise alternative.

The second place goes to AWS. We can say, that with such middle-ground pricing and definitely a great wide functionality range, AWS continues to lead the way. Its extensive set of products and services, together with its enterprise-friendly features, make it a compelling offering for large organizations.

However, it looks like Azure DevOps services, with their 2nd place on the market share, offer the most expensive set of on-premise products. Apparently, Microsoft Azure is not the best option for the starters, developing an MVP. But this is a perfect reliable solution for large-scale organizations.

We also want to mention two other options, worth considering for beginning: Firebase and Heroku. Even though these two didn’t fit into the comparison, because of another set of products, with completely different measurements. But they offer a lot of free minimal functionalities, which would be enough for MVP launch. For example, Firebase provides free cloud storage for 5 stored GB, 30GB for transferring, and more than 2 million operations. Besides, with Firebase there is the possibility to get 1GB stored and 10GB transferred data in Realtime Database entirely for free. So, take a look!

Also published here