What Is Spare Cloud Capacity?

This is part of a series of articles about Cloud Optimization

In this guide, you will learn:

Excess Capacity at AWS, Google, Azure and Alibaba Cloud

What is excess capacity?

The promise of cloud computing is that whenever you want to ramp up a new instance (or hundreds of instances), you can do so with just the click of a button.

For cloud providers, this presents a capacity challenge.

Cloud providers need to ensure that whenever you need compute capacity, they have it ready for you to use. In other words, cloud providers always need to have unused servers ready and waiting for you whenever you choose to click that button. From a business standpoint, spare capacity is never a good thing. Just like any company doesn’t want to be stuck with excess product they can’t sell, cloud providers want to be monetizing every last piece of hardware they own and run. Not to mention the fact that they still need to pay for these servers. Rent, electricity, the hardware itself, setup, and maintenance all cost money.

At the same time, cloud providers can’t be monetizing all of their servers or else customers wouldn’t be able to ramp up their compute when they needed to. So how can cloud providers both leave servers open for new customers AND monetize them?

 
“Everyday Amazon adds enough compute capacity to its global fleet to power Amazon.com.”
– James Hamilton, VP & Distinguished Engineer, Amazon.com.

The dawn of spot instances

Amazon Web Services was the first cloud provider to address this challenge, and did so with a brilliant offering called “spot instances”. Announced back in 2009, spot instances were launched to solve this very problem. spot instances (aka spare capacity at AWS) is how AWS packages their spare capacity to customers – by selling them at a massive discount (usually around 80% off on-demand prices). This attractive pricing comes with a major caveat. Once AWS needs your spot instance for an on-demand customer, your spot instance will be terminated with just two minutes notice. By doing so, AWS was able to monetize their spare capacity without closing them off to new customers

 
 
“On Average, every week AWS customers are using more compute capacity on Amazon EC2 spot instances than customers 2012 were running across all of Amazon EC2.”
Google launch Preemptible VMs, Azure launch Low-priority VMs and Alibaba Cloud (Aliyun) launch ECS Spot Market

Other Cloud providers like Google and Azure followed suit by releasing their own spare capacity offering at a reduced price, called Preemptible VMs at GCP and Low-priority VMs at Azure.  Names aside, it’s all basically the same – the clever way cloud providers monetize their spare capacity. Since then, Alibaba Cloud (Aliyun) has also released their ECS spot instances offering and we’ve started working with other major compute providers to help them unload their spare servers at attractive discounts.

Unfortunately, most companies focus on the pennies rather than the big bucks when it comes to reducing their cloud costs.

At the end of the day, leveraging the spare capacity that cloud providers offer can be the most effective way to cut costs. Rather than spending tons of time and effort optimizing CPU Utilization to save around 20%, you can leverage spare capacity to save up to 90%. The risk of interruptions and lack of any SLA, however, are why Spot.io was born – enabling our customers to leverage all of the savings spare capacity has to offer with none of the risk.

Read more about AWS EC2 spot instances.

Related content: Read our guide to cloud scalability, cloud automation, cloud orchestration, or cloud infrastructure.