Overprovisioning & Always-On Resources Lead to $26.6 Billion in Public Cloud Waste Expected in 2021

Overprovisioning and leaving cloud resources on are two enormous sources of wasted spend.

Wasted spend drags down IT budgets — of particular importance as we enter 2021. The Flexera 2021 State of Tech Spend report found that the biggest change in key IT initiatives from 2020 to 2021 was in cost savings, with the percent of respondents ranking cost savings as a top initiative tripling year-over-year.

It’s important that this is being recognized. Based on data collected by Gartner, we estimate that wasted spend will exceed $26.6 billion this year.

Where the Wasted Cloud Spend is Coming From

Gartner estimates a total market spend of $304 billion on public cloud services end-user spending in 2021, as broken out in the table below. Their estimate for the proportion of that spent on Infrastructure as a Service (IaaS) is $65.3 billion. While wasted spend can be found in any area of cloud spend, customers tend to see the largest amount in these two areas, as well as finding it easiest to identify.

Idle Resources

Cloud resources can be considered “idle” when they are running while not being used. For example, when development servers are left running overnight and on weekends when they’re not needed. Since compute resources are paid for by the minute or second, that’s a large portion of the week they’re being paid for but not used (and yes, this applies even if you have reservations.)

Our data shows that about 44% of compute spend is on non-production resources. If we estimate that non-production resources are only needed during a 40-hour work week, the other 128 hours (76%), the resources are sitting idle.

Applying that to the Gartner IaaS number, we estimate that up to $14.5 billion will be wasted on idle resources this year.

Overprovisioned Resources

Overprovisioning occurs when a larger resource size is selected than is actually needed. There is a mindset of safety behind this, as of course, no one wants their applications to be under-resourced.

But the overprovisioning occurring is far beyond what is necessary, given the elasticity of the cloud. About 40% of instances are sized at least one size larger than needed for their workloads. The cost can be cut in half by reducing an instance by one size, while downsizing by two sizes saves 75%.

Many of our customers show a large percentage of their resources are oversized, but bringing this to a conservative estimate of 40% of resources oversized by one size, giving us a savings per resource of 50%, we estimate that up to $8.7 billion is wasted due to overprovisioning.

Orphaned Volumes and Snapshots

Another significant source of waste is orphaned volumes and snapshots. These are resources that have been detached from the infrastructure they were created to support, such as a volume detached from an instance or a snapshot with no volume attachment.

Our customers spend approximately 15% of their bills on storage, and we found that about 35% of that spend is on unattached volumes and snapshots. Applying that to the Gartner spending numbers, we estimate that up to $3.4 billion could be wasted this year on orphaned volumes and snapshots.

Reducing Wasted Spend

Altogether, this gives us an estimate of $26.6 billion to be wasted on unused cloud resources in 2021. This waste estimate is just based on the three prominent sources of cloud waste. It does not include wasted spend on Platform as a Service (PaaS), which makes up $55 billion in cloud spend according to Gartner’s estimates, nor from SaaS, unused reservation commitments, inefficient containerization, and other areas of the bill.

Attacking the three problem areas above is a great area to start for nearly all public cloud users. Here at ParkMyCloud, we’re on a mission to do just that. See how and try it out today, to do your part in reducing wasted cloud spend.

Originally published at www.parkmycloud.com on January 21, 2021.

--

--

--

CEO of ParkMyCloud

Love podcasts or audiobooks? Learn on the go with our new app.

Recommended from Medium

The coin change problem in Haskell

Ternary tree depicting coin change problem for simple example

5 Ways to Measure the ROI of Your Robotic Process Automation Deployment

The importance of estimating tasks in the software world

Version controlling your dot files

Laptop open with code

Polarity.Exchange version 1.1.2 patch notes

Share the joy of Javascript

notes — 5 to 11 — unit 1 — mitx — 11.154x — 31st jan 2017

picturesque Green Amethyst Silver Green Ring handmade L-1in US 5678

Get the Medium app

A button that says 'Download on the App Store', and if clicked it will lead you to the iOS App store
A button that says 'Get it on, Google Play', and if clicked it will lead you to the Google Play store
Jay Chapel

Jay Chapel

CEO of ParkMyCloud

More from Medium

An example for Workload Migration from on-premises to Cloud using AWS Migration Service

Inside investigation into who Tech2SDE is

How Prancer can detect secrets in IaC templates

AutoScaling Status check vs Health check? Which one is suitable for u?