Google Cloud Apps Admins

Finding the Best Public Cloud VM Provider in 2017: AWS, Azure or Google Cloud?

Written by Christopher Alghini (Senior Cloud Strategist) | September 12, 2017

Adoption of cloud technology is one of the biggest factors driving businesses forward these days. However, as several companies look to reduce costs and minimize risk during their cloud adoption, the introduction of VMs or Virtual Machines is starting to change the landscape. Today's CIOs want to "test the waters" with their cloud solution before committing to substantial changes, and that's where VMs come in handy.

Today, if you're looking for a cloud-based VM infrastructure, there are 3 main competitors to choose from, those are Google Compute Engine, Microsoft Azure, and Amazon Web Services. Since all businesses struggle when choosing the right providers for their technology needs, the best way to make sure that you make the right decision is to evaluate the opportunities provided by each service, side by side.

 

The World of Virtual Machines in Public Cloud

Virtual machines are "IaaS" or "Infrastructure-as-a-service" offerings. This means that any VM instance you create should perform just like a real computer, complete with memory, CPU, storage, and networks. With that in mind, one of the first things you'll need to compare when choosing a VM provider is the functionality offered by each company.

  • AWS: While Amazon is currently storming ahead of some competitors as the market leader, it's worth noting that it offers its customers a somewhat outdated technology stack. With Amazon, it's impossible to create your own custom VM, which means you need to choose from a pre-defined list. In other words, you can't select the functionality that works best for you.

  • Azure: Although Microsoft is making strides in the world of open-source computing, their platform always remains focused on other Microsoft offerings. This means that it's increasingly difficult for customers to use a hybrid cloud system if they choose to work with Microsoft. Unfortunately, as a vast percentage of enterprises and companies are looking into hybrid systems for their cloud needs, this restriction could damage user adoption for Azure.

  • Google: Unlike Amazon and Microsoft, Google Compute Engine is a far more flexible solution to the creation of virtual machines. Google is currently leading the market in terms of new technology, with open-source software including Tensorflow and Kubernetes. Customers can create VM platforms that suit them, rather than having to select from an existing list, which makes custom adoption easier.

 

Virtual Machine Cost and Scalability

Once you've begun to use your VM solutions, there's a good chance that you'll find you need to scale them up and down according to the changing needs of your business. The primary benefit of moving to the cloud is that you're adopting a solution that's inherently scalable. Of course, there's more to scaling than simple functionality, you need to make sure that the pricing suits your budget too. In terms of scale:

  • AWS:     Amazon can resize and auto-scale their virtual machines. This means that the system can keep up with your preferred number of instances, terminating and starting them according to your needs. Auto-scaling can be useful, and resizing can be done with a couple of clicks. Unfortunately, the full process has a few compatibility issues to consider, which means that it may be easier to simply back up your VMs and create new instances.

  • Azure:    Microsoft also offers auto-scaling through availability groups. You can set the number of instances to scale with, and select criteria such as queue size and CPU workload. The Azure web-interface can help you to resize instances quickly, but there are some limitations when it comes to tier-changing.

  • Google:    Finally, Google Cloud provides a solution for horizontal auto-scaling that works by removing and adding new instances through a managed VM group. You can adjust your needs according to any metric, according to custom criteria. Any instance can be resized according to your specific needs.

When it comes to cost, there are also several factors to think about, for instance:

  • AWS:     Amazon comes pre-packaged with complicated pricing models. The hourly rate for your Amazon "EMR" services will depend on the instance type in question, including factors like high memory, storage, high CPU, and more.

  • Azure:     Microsoft offers a per-minute pricing solution, similarly to Google. However, the cost is much more expensive overall than that provided by Google. That means that you're accessing the same functionality, for less cost thanks to the Google network.

  • Google:    Google Cloud offers a flexible pricing model that comes with sustained-use discounts and preemptible VMs which can reduce costs by up to 80%. According to evaluations of Google Compute Engine vs other VM models, Compute Engine is much cheaper than Microsoft. Running an 8-core VM on a windows server, for instance, would cost $833.28 with Azure, and $506.00 with Google.

 

Virtual Machine Reliability

Finally, it's crucial for CIOs to ensure that the cloud provider they choose to build their VMs with can offer reliability and sustainability in this competitive business environment. If a virtual machine goes down, this can lead to serious problems for any company, which means that you need to choose a provider who can protect your data and functions as much as possible:

  • AWS:    Amazon requires their users to reboot entirely for VM maintenance and security breaches, which means that planned outages can be necessary. At the same time, Amazon will only provide a discount on your VM services if your VM is down for more than 0.05% of the time for the entire month.

  • Azure:    Microsoft Azure customers frequently report that they experience a lot of indeterminate start-up time problems for their VMs. Additionally, the VMs that they do create can disappear and crash at random. When it comes to compensation for availability, Microsoft's discounts are the same as Amazon's, with a cap on the amount of credit that can be received.

  • Google:   Thanks to live migration, there are no planned outages with Google Cloud. At the same time, Google VMs can perform consistently, start quickly, and have an impeccable 99.95% uptime. Google also provides credit to users if the VM was unavailable for more than 5 minutes consecutively, which means that customers are always covered.

Choosing the Right Virtual Machine Platform

Ultimately, choosing the perfect virtual machine for your needs will be a subjective process that depends on your specific requirements. Each cloud vendor offers unique features, along with their own distinctive pros and cons. Whichever solution you choose, make sure that you're fully aware of the competitors in the space, and that you've planned accordingly to help you take your VM creation and cloud migration to the next level. 

For customers using AWS, Azure, and Google, multi-cloud is more beneficial today than ever before. As the name suggests, "multi-cloud" is the process of using cloud services accessed from multiple heterogeneous providers, such as AWS, Microsoft Azure, and the Google Cloud Platform, all linked together. In a wider sense, multi-cloud also covers the use of private systems and hybrid deployments that utilize multiple vendors.   Multi-Cloud Curious?  Read about The Undeniable Value Proposition of a Multi-Cloud Strategy.