Reserving Cloud Instances and Capacity: 3 Reasons For and 3 Against

Introduction

Cloud computing is the most prevalent way for businesses to manage and 'fix' cloud costs.

One of the most popular ways to manage costs in cloud computing is to 'reserve' instances and capacity - also referred to 'reservations' or abbreviated 'RI'.

It involves committing up front for period of time with a discount as your reward.

Reservations can help ensure that the computing resources needed are available when needed.

Typically for 1 or 3 years, with a bonus of 20 to 60% discount respectively.

"Well, isn't that nice?" you say.

It just does not come cheap. 

It's really inflexible.

‍

Cloud and fitness

I usually compare it to a fitness subscription. You have great intentions of going into it, doing the work, and getting the body of a greek God.

But then life - the everyday - hits.

All your good intentions don't hold up. Now that subscription is just a hole burned into your pocket

But in cloud, that factor is 10, 100 or 1000 times the size of the fitness one.

And every reservation is tied to every single muscle you want to work.

Not per muscle group, but every single muscle.

Cloud and muscles

In cloud, these muscles are just Virtual Machines (compute), Storage (capacity), Databases, and so on (often referred to as categories).

And every category has it's own set of items, or Stock Keeping Units (SKU's for short).

Take Virtual Machines. In Microsoft Azure alone, you can choose between 720+ different SKUs.

Seven hundred and twenty plus different virtual machines!

For comparison, the human body has around 600 individual muscles.

The entire human body.

Holy cow.

And don't get me started on AppService, Storage, Databases, Functions, and more.

It would be a looooong post *yawn*.

But for the sake of keeping a sober tone and not just rant about the subject, 

Below, my assistant and I have debated three (3) pros and cons of reserving instances and capacity in cloud computing.

To guide you on how to avoid a nasty cloud-deal for your company.

3 Reasons Pro

First, the main advantages of reserving instances and capacity in cloud computing is that it can help ensure that the necessary computing resources are available when you need them.

We see more and more people complain about having to wait for various services 3-6 months into the future.

Cloud is 'just' datacenters. Chip shortages hit them as hard as every other buyers.

Second, reserving cloud resources can help you reduce costs by taking advantage of discounts offered for long-term commitments (1 or 3 years).

For one year comitments, you typically get 20 % off. For three years, you get up to 60 %.

Nice (said in Borat fashion).

Another benefit of reserving instances and capacity is that it can help improve performance. 

By reserving instances and capacity, businesses can be sure that their applications run on dedicated resources, which can help to improve performance and reliability.

It is not always a given. But comitting to usage also comes with an increas service level (SLA), which for some resources is higher than to Pay as you Go.

‍

3 Reasons Con

One of the main drawbacks of reserving instances and capacity in cloud computing is that it limits flexibility. 

Committing to certain levels of computing resources for a specified time limits your ability to scale up or down as needed.

And it does not prevent you against overcharges. So if something unexpected happens - say a misconfiguration or bug - you should expect to see that on your bill.

IT-leaders have complained to me about simple mistakes leading to thousands of euros (or dollars, if that's your spiel) in overcharges.

Unfortunately, those errors are often human.

And you know what the hardest thing to manage is, right?

Which leads us to the second drawback: Reservations can be hard to manage. 

Make sure that the computing resources you reserve are actually consumed or used.

A good rule of thumb is to reserve resources for production ready environments and applications.

It may make sense in non-production (development, test, etc.). But there significantly better ways to manage this, such as auto-scaling, shutting down VMs in off-hours, etc.

Lastly, the upfront cost of reserving instances and capacity can be high.

This ties back to the fitness analogy. Someone had good intentions for the business some while ago. 

But the project or application running on said cloud resources either flunked, people left or something other unexpected event happened.

The business paid the bill and cannot get that money back, contrary to if they had bought physical servers instead.

Conclusion

Reserving instances and capacity in cloud computing can be a really tempting way to save a good buck or two.

However, it limits flexibility, and can be expensive and difficult to manage. 

Businesses should carefully consider the pros and cons of reserving instances and capacity in cloud computing before making a decision.

Either talk to folks who know a thing or two about it *cough* us, Ximplifi *cough*. 

Or someone else.

Surprise fact:

Microsoft, Amazon, Google, and the other cloud providers want you to reserve.

It makes their forecasts and stockholders happy.

But given 30-40% of all cloud resources are wasted - bought but not used - there are good chances your loss is their gain.

Happy choosing. And please, choose wisely.