Read time

Should I Use Reserved VM Instances?

By Heather Robinette

More and more, we hear the question from Microsoft Azure customers, “Should I be using Reserved VM Instances?” As cloud adoption continues to accelerate, customers are looking for better visibility, recommendations, and other ways to reduce costs. And the answer to the Reserved VM Instance question is…It depends. The reason there is not a straight “Yes” or “No” answer is because there are things to consider when analyzing fit for your unique deployment that can change breakeven points and level of risk. Today we are going to talk about what a Reserved VM Instance is briefly, and then we are going to discuss some things to consider as you think about whether it makes sense for you or not.

What are Microsoft Azure Reserved VM Instances?
First, let’s take a look at what Microsoft Azure Reserved VM Instances are and how they can benefit you. Microsoft Azure Reservations are a way for a company to guarantee space and possibly save some money. Instead of paying for cloud usage on demand, also known as the pay-as-you-go model, you can essentially commit to a usage term in exchange for a discounted cost. You have the option of 1 year or 3 years of commitment by VM size. The longer the length, the better the discount. That said, you can opt to pay upfront or commit to pay for the RI period and pay monthly. The latter option frees up cash flow, which allows more companies to use Reserved VM Instances. As we will touch on in the next section, knowing your current usage, utilization, and future needs are key factors in deciding whether to apply Reserved VM Instances or not. If you’d like to learn more, see one of our past blogs with more detail on Reserved VM Instances.

Are Reserved VM Instances a Good Fit for Your Azure Deployments?
At first glance, it seems like a no-brainer with guaranteed cost savings, but as you start to look at the variables, you’ll quickly realize there is more to the decision than you thought. For starters, you need to know what VM series you want to use, what region you want to be in, and, more importantly, how many Reserved VM Instances would be applicable. With each Reserved VM Instance, you are paying for that VM (or instance) 24x7x365. So even if you potentially stop (deallocate) that VM throughout the month, you will still pay the same consistent monthly rate. In cases where Virtual Machines run 24x7x365 Reserved Instances will save up to 40% of the VM runtime costs. In fact, 60% is a good rule of thumb if you are considering a 1-year RI option. If you are running a VM (instance) more than 60% of the hours in a month as normal usage, Reserved VM Instances will provide cost savings. But the cost savings do come with a caveat of either 12 month or 36-month lock-in. If you decide to stop using the VM’s prior to your committed time period, you will pay an early termination fee of 12%. So, there is a breakeven point to consider for analyzing risk.

Let’s look at a cost savings example using 1-year Reserved VM Instance on a D4 v3 VM size.

VM

Monthly Runtime

Annual Cost Savings

Breakeven (months)

D4_v3

24 hrs X 7 days (730 hrs)

$678.84

2

D4_v3

18 hrs X 7 days (540 hrs)

$258.36

5

D4_v3

12 hrs X 7 days (365 hrs)

($162.12)

N/A


In this case, the top scenario would make sense as there is significant cost savings and very little risk (breakeven after month 2 with 12% early termination). The second scenario still provides cost savings but includes a bit more risk in that it would take 5 months to find a breakeven (with 12% early termination). And the bottom scenario would NOT make sense with the negative cost savings since the VM runtime is less than the RI cost savings.   

One additional note of complexity is that you can technically benefit from a single Reserved Instance across multiple VM’s. In this case, the VMs shouldn’t run concurrently, but if you have two VMs that are on at different times, you can still gain cost savings because you have that “slot” reserved. An example use of this is if you had VM’s running to support different time zones/regions in the same Azure subscription.

 

VM

Time On

Time Off

VM #1

08:00 GMT

17:00 GMT

VM #2

18:00 GMT

02:00 GMT

According to the example above, you would use VM #1 for 9 hours and VM #2 for 8 hours, making up a total of 17 hours. If these time allotments don’t overlap, then this would be a situation where a Reserved Instance would make sense for these two VMs since they could use the same Reserved Instance.

So, as you can see, Reserved VM Instances are not as straightforward as at first glance. If you really want to analyze potential cost savings of Reserved VM Instances based on historical usage, we encourage you to reach out to us directly and check out other blog posts like 6 questions that can help you decide if Reserved Instances are for you or not.

Want to Learn More about Microsoft Azure?
Download the Ebook

Tags: Microsoft Azure, Cost Optimization

azure cost optimization

Related Articles

Monitoring and Managing Commitment-Based Discounts in Microsoft Azure

More and more, we hear the question from Microsoft Azure customers, “Should I be using Reserved VM Instances?” As cloud adoption...


Topics: Microsoft Azure, Cost Optimization

Protect Your Microsoft Azure Investment with Cost Anomaly Monitoring

More and more, we hear the question from Microsoft Azure customers, “Should I be using Reserved VM Instances?” As cloud adoption...


Topics: Microsoft Azure, Cost Optimization

Mastering Azure Resource Tagging: Best Practices for Effective Cloud Management

More and more, we hear the question from Microsoft Azure customers, “Should I be using Reserved VM Instances?” As cloud adoption...


Topics: Microsoft Azure, Cost Optimization