Mitchell Gee
August 13, 2019

Reservations on Reserved Instances

There’s no denying it. Cost is one of the biggest factors within an organisation’s decision-making process. From the stationery it stocks to the highly scalable Cloud-based IT solution it chooses. With this in mind, you can’t blame key decision makers for being impressed by the cost-saving stats that are offered up with Reserved Instances. Paying up front for ‘up to 72% cost savings versus pay-as-you-go’ is a no-brainer, right? Well, not always.

Here we look at why Reserved Instances (RIs) don’t always deliver the cost savings offered and how the end user can use Microsoft’s Azure offering a little smarter, to get the maximum benefit from their investment.

To start, let’s just recap on what a Reserved Instance is. (Already know? Then skip over this bit.) Put simply, it’s an upfront payment for a fixed period, with the choice of one or three year commitments, that means you don’t have to pay for the computing costs of the a VM (Virtual Machine); so, the CPU and RAM. You just pay for the OS and other supporting costs (storage, networking etc.)

Why 72% is not always 72%

The thought of freeing up two-thirds of infrastructure costs may sound like a simple and effective solution. But it doesn’t always work out to be the most cost-effective solution for every business. In our experience that 72% cost saving is rarely achievable. We’ve found the actual savings associated with RI is more likely to float around the 20-40% mark. Still good, but a disappointment if you’re expecting 72%.

Why is this percentage so much lower? Well, when a business purchases a RI, they’re simply getting a discount on the computer power of the machine. Not the cost of the OS, storage, networking, etc. So, if an organisation has a whole heap of data to store or runs a very large network, the percentage benefits they’re going to reap will be lower.

What’s more, when a business commits to a RI it must assume its needs aren’t going to change within the next one to three years to reap the benefits offered. What happens if it wins a new client? Or a large order comes in? Or worse, its business shrinks in size? Microsoft’s Azure offering claims to be ‘elastic’ and ‘scalable’, yet to reap the biggest discounts with RI, a business must tie itself to its current infrastructure size.

Getting the most out of Microsoft Azure

This may seem like a doom-and-gloom picture, but there are ways to ensure your clients are getting the best financial rewards. Azure can provide flexibility and higher savings if a business considers how it uses its system first.

Here are a few ways businesses can use Azure smarter for maximum benefit

1. Automate on/off schedules

Systems that require a VM running all the time at a steady capacity – such as a database – will see percentage discounts at the higher-end of the scale with a RI. But if they only need it operating at set times of day, they could be paying for the OS and Cloud-storage when it’s not required. Implementing a run book so the VM is only running during specified times, could reduce costs by 50%. For example, if it’s being used for a Domain Controller and staff will only be logging on during set office hours, then ensure its only running at this time.

2. Utilise scale sets

If a machine needs to run night and day, does it need to run at full power all the time? Take an online retailer as an example. It probably operates at a steady utilization, but during sale season the system will be put under greater pressure. It’s possible with Azure for the smaller machine to then automatically scale up or out at these peak times, matching the natural ebb and flow of the business.

3. Use a B-Series VM

B-Series or Burstable VM, does exactly as the name suggests. It runs at a lower CPU than it’s capable of, only increasing its power output when the system requires it. Perfect for workloads with large spikes between low and high usage. The system gives the end user the flexibility of being able to respond to sudden demand, but only paying for the full CPU when it’s required. How? Well, while the B-Series VM is running in the low-points and not fully utilising the baseline performance of the CPU, the VM instance builds up credits. When the VM has accumulated enough credit, usage can be burst up to 100% of the CPU and beyond when needed.

4. Right-size the machine

This may sound like an obvious one, but it’s surprising how many organisations are paying for far more computing power than the business requires. Azure’s monitoring tool is a great way of making sure the organisation isn’t paying for more than they need. And, should that need change in the future, Azure’s offering is scalable enough that they can increase or decrease the VM’s capacity at a later stage.

In a nutshell

I genuinely believe RIs have a place in the world of Azure, for organisations that run a consistent workload 24/7. But many businesses are using them as a cost-saving solution that doesn’t deliver maximum benefit. Utilising the methods mentioned above could help save them far more, while still enjoying the freedom of Azure.

Every business is unique, so there’s never a one-size-fits-all solution. It’s all about finding out what’s right for them and tailoring the offering accordingly. So, if you need a little bit of guidance on finding the best option for your client, don’t hesitate to contact us at Westcoast Cloud. We’ve got the resources and tools to help point you in the right direction.