<img src="https://certify.alexametrics.com/atrk.gif?account=VdU0q1FYxz20cv" style="display:none" height="1" width="1" alt="">

Cloud Cost Optimization

The Definitive Guide to Minimizing Public Cloud Costs

Why Does Cloud Cost Optimization Matter? 

The rush towards cloud computing was fuelled in large part by cost-conscious IT departments that saw the cloud as the natural next step for virtualization and a way to stretch tight budgets, and by the need for IT to provide fully elastic services that can scale in real time. But that doesn’t mean it’s time for IT to take its eyes off the bottom line.


As cloud computing models proliferate and evolve, it's just as important as ever that technology leaders make informed decisions about which products, services, and payment models deliver the best results.

Public clouds are an important part of your hybrid cloud strategy – they’re powerful, scalable and highly available. They can also be very expensive; it’s possible to get a surprise bill at the end of the month that’s several times more than you planned for.

One of the most common concerns is how to track and control public cloud spending. Virtual Machine Sprawl can still be an issue in enterprise data centers, but maturing IT virtualization processes, as well as performance optimization, automation and rightsizing, allow data center admins to identify and remediate these situations. In the public cloud, by contrast, sprawl is exacerbated by the fact that an organization’s initial public cloud consumption is seldom managed by IT. Instead, it’s controlled by various lines of business (LOBs) through shadow IT teams that have little or no expertise in IT processes. Here, we discuss the various ways in which a Cloud Management Platform (CMP) and other strategies can help to minimize public cloud costs and provide cloud cost optimization.


 means ensuring that all of the Virtual Machines in your infrastructure are allocated the correct resources for their workload, with the goal of minimizing costs and maximizing performance.

cloud cost reporting

Over-allocation wastes expensive server resources and decreases ROI for the infrastructure. In public cloud environments, over-provisioned instance types mean higher daily costs. And for private cloud workloads, rightsizing your workloads can result in deferred hardware purchase savings over several quarters, years or even in perpetuity.

Rightsizing reports analyze performance monitoring data for individual virtual machines and then recommend changes to resource allocation. Such a report might, for instance, identify a poorly performing virtual machine and recommend that additional resources be allocated. More often, however, rightsizing reports find that virtual machines have been provisioned with more resources than they need. Reclaiming these resources may make it possible to achieve a higher overall virtual machine density, thereby reducing costs.


In addition to running rightsizing reports, a cloud management platform can also gather performance metrics to determine the average and peak resource consumption of an instance over a configurable period. If these remain below a threshold, the CMP will recommend a change to a smaller, more appropriate instance type.

With configurable rightsizing rules, you can enforce more aggressive rightsizing thresholds for less critical workloads, leading to even more savings. For example, you might configure a threshold for downsizing a less critical workload when CPU usage is less than 50%, but downsize a critical workload only when CPU usage is less than 25%.

It’s important to manage capacity and sprawl in an integrated fashion, such that both IT and the end consumers of IT services have visibility of their workload capacity needs. Providing rightsizing recommendations ensures that the infrastructure is optimized for the real needs of the business. Moreover, giving stakeholders rightsizing visibility empowers the end IT consumers to make knowledge-based decisions in real time.

How Rightsizing and Automation Can Save You Money

Independent VMware vExpert Stuart Burns demonstrates how to use vCommander to reclaim the wastage of unused/underused resources to reduce cloud costs.

Power Scheduling

CMP power schedules let you configure a schedule to shut down and restart instances, so you don’t leave them running overnight or on the weekend, when nobody is using the resources you’re paying for.

cloud power schedules

You pay by the hour for your running public cloud instances, so the less time they’re running, the better. Many types of workloads don’t need to be running 24/7, but it’s easy to forget to power them down. For example, developers may spin up an Amazon EC2 instance that they only need to use for testing from Monday to Friday, during working hours.

A power schedule group that shuts instances down at 7 pm each night and powers them back up at 7 am, and leaves them powered off over the weekend, can allow a 65% savings in compute cost versus an instance that runs constantly.

Power Schedule Group

Ensuring your instances are powered down when they’re not needed can lead to huge savings in large public cloud deployments, and should be an integral part of your cost-savings strategy.

Lifecycle Management

For public clouds, the need for lifecycle management is clear. Powered-off instances hosting workloads that are out of date and shouldn’t be used can be inadvertently powered on, with the potential to increase monthly bills.


Unmanaged virtual machine sprawl can negatively impact the bottom line by wasting resources. One of the best approaches to control VM sprawl involves assigning an expiration or decommissioning date to each virtual machine that is created (except for core infrastructure VMs or VMs that are running mission-critical workloads).

Moreover, the adoption of a cloud model implies changing your organization’s consumption culture from an ownership model to a rental model. Workloads (especially those in development and testing environments) no longer live in perpetuity, but instead exist only as long as they’re required.

A background process should perform a daily scan of the virtual machines to determine which are nearing their expiration date. A message can be automatically sent to the virtual machine’s owner informing them of the impending expiration. If the VM is still being used, the owner can typically extend its lifespan. Otherwise, the virtual machine will expire and be automatically purged from the system so that its resources can be reclaimed.

Lifecycle Management

Ideally, an administrator should also be able to use a reporting engine for capacity planning purposes. Imagine that a private cloud is running low on available memory, and the administrator knows that some new virtual machines will need to be created soon. The administrator can check a report to see which virtual machines will be expiring soon, and which resources are likely to be reclaimed. The availability of such information may allow the organization to postpone hardware upgrades, and so defer additional expenditure.

Cloud Cost Analytics

Reports providing cost visibility, cost comparison and chargeback can be used to educate business units on correct workload placement based on cost and technical requirements, allowing you to compare apples to apples.

icon-cloud cost mgmnt

Enterprise organizations often struggle to compare the cost of provisioning services internally with the cost of sourcing them externally in an effort to establish whether a public or private cloud is the right choice for any given workload. But frequently, the comparison is not so much apples to apples, but that of apples to oranges, and even comparing the prices between the cloud providers themselves is troublesome.

TCO calculators from cloud providers and other companies are ultimately vendor-biased marketing tools that may not provide an independent assessment of a competitor’s offering. A CMP provides cost modeling, enabling you to compare your existing workloads in both public and private clouds and analyze the various costs associated with each—not just compute, storage and memory, but also the sometimes hidden cloud costs of software, support and IT resources.


In addition to associating costs with specific hardware and software items, it’s important to compare costs within your infrastructure, such as production systems, development systems, systems with high-end or low-end compute/storage resources, or even across different groups of your consumers.

Implementing this form of cost modeling can help to temper a common bias in favor of public cloud. 

Infrastructure Cost Visibility - Chargeback/Showback
Easily provide IT chargeback or showback to your users with vCommander’s built-in cost calculation and simple-to-use wizard-based cost models that you can customize for both private and public infrastructure.

AWS Reserved Instances

If you’re using AWS, Amazon EC2 Reserved Instances (RIs) are one of the most obvious ways to control compute costs, allowing you to reserve EC2 computing capacity in exchange for significantly discounted hourly rates (up to 75%) compared to On-Demand Instance pricing.

Image_021 copy

However, Reserved Instances do have some downsides. First, it’s not always easy to predict usage over one or three years, especially if you have no historical data to base your prediction on. Secondly, businesses that are moving to a public cloud because of the attractiveness of the pay-as-you-go cloud model are wary of capital costs that may be taking them back to long-term contracts and sunk costs. Finally, setting up the extra capacity for a particular instance type can be complicated, time-consuming and costly.

The best use cases for RIs are applications with very stable usage patterns. These allow you to achieve the highest savings with rapid ROI, lowering the potential impact of future unused resources. Follow these best practices to get the greatest benefit from RIs:

  • First, buy reserved capacity for instances that are always running.

  • For instances that aren’t always running, but with stable usage patterns, buy the minimum number of RIs to keep the application running, based on minimum or average sustained usage.

  • Consolidate your RI purchases in a centralized AWS account, managed by a single financial team, so that resources not used for one application or team can be used by others.

  • A centralized AWS account can purchase RIs more effectively when multiple teams use instances from the same instance family so that a single RI can be applied across teams as usage varies.

  • For applications with non-stable usage patterns, you may obtain better value from On-Demand Instance pricing, choosing smaller instance sizes and implementing power scheduling.

Reserved Instance Recommendations

By using a CMP to perform the analytics and make RI purchase recommendations, you can have the confidence that your RI purchases are providing the best return on your investment.

Rinse and Repeat

Cost optimization is an ongoing, iterative process. Once you've rightsized, applied power schedules and purchased RIs, it's critical to monitor your infrastructure and your bill, and continually optimize.

Image_023 copy

To get the most out of cloud cost optimization, following a strict process is critical:

  1. Apply rightsizing recommendations. This ensures that you’re using the proper resources for your workloads.

  2. Configure power scheduling. VMs that are running needlessly will affect your usage data.

  3. Analyze usage patterns towards the end of the month.

  4. Based on your analysis, and after discussion with owners of the target instances, purchase AWS RIs.

  5. Monitor and repeat.

If you begin with RI planning, you may end up purchasing the wrong size or quantity of RIs, which can be an expensive mistake.

Embotics vCommander Cloud Cost Optimization Features

Additional Resources