Businesses operating a hybrid cloud should be aware of their primary cost drivers and take advantage of any opportunity to increase cost efficiency. However, there may be some items on our hybrid cloud cost efficiency checklist that businesses are unaware of.
If a business operates a hybrid cloud, its IT costs are comprised of on-premises capital expenses and cloud-based operating expenses. Since on-premises capital expenses are mostly fixed they are comparatively easy to track. However, cloud-based operating expenses are highly variable and can quickly spiral out of control if they are not carefully managed.
Most cloud service providers supply cost management tools that monitor utilization and make recommendations about rightsizing or terminating certain resources. However these tools do not cover every type of resource or every opportunity to reduce. Trying to identify “zombie resources” can be very difficult unless the business has total visibility of its hybrid cloud environment.
Even with total visibility, it is not always obvious where to look for cost efficiencies. Consequently we have compiled a hybrid cloud cost efficiency checklist that businesses can use to find hard-to-identify wasted spend and either terminate the resources or reallocate them. If you have any questions about our hybrid cloud cost efficiency checklist, do not hesitate to contact us.
Learn more about hybrid cloud in our ebook “The Enterprise Guide To Hybrid Cloud”.
Our hybrid cloud cost efficiency checklist
Rightsize all VMs and Instances
Over-provisioning has been a common practice in the data center since systems administrators needed existing infrastructure to handle peak demand. Even with tremendous improvements in utilization driven by virtualization, resources are often idle during slow periods. This doesn’t cause major headaches on-premises but it’s one of the most common areas of wasted spend in the public cloud. It’s a best practice before migrating a workload to either the private or public cloud infrastructure of a hybrid cloud, that you assess its utilization and rightsize all VMs.
Use the most appropriate block storage
When users launch a VM or instance, it is often the case that premium-cost SSD block storage is attached by default or users forget to select the most appropriate block storage for the workload. Many workloads do not require SSD block storage, and businesses can significantly reduce costs by downgrading block storage volumes to HDD storage. Needless to say, use this opportunity to identify unattached block storage volumes and, after taking snapshots of them, terminate them.
Delete snapshots that are no longer required
Individually, snapshots do not have a significant impact on hybrid cloud cost efficiency; but, if the business is paying for obsolete snapshots that are no longer required, it's not a very efficient use of their hybrid cloud budgets. Usually only the most recent snapshots are required for recovery purposes, so a best practice is to create a policy that looks for aged snapshots that are outside your retention policy and automatically delete them.
Release unattached network IP addresses
Network IP addresses have a unique pricing structure inasmuch as when VMs and instances are running, the IP addresses are free of charge. However, if businesses retain IP addresses once the VMs and instances to which they were attached have been terminated, they are charged for them. The costs can accumulate surprisingly quickly, so releasing unattached IP addresses can save a considerable amount of money.
Find and terminate any other zombie resources
The term “zombie resources” relates to any unused resources that, while idle, are still incurring charges. Unattached IP addresses is one example of a zombie resource, but there are many more—for example idle load balancers, provisioned compute instances, and block storage volumes. It can often be worth scrutinizing test environments for zombie resources to identify projects that may have been started and then abandoned without every component of the project being terminated.
Apply start/stop schedules to non-production resources based on utilization
Many businesses are aware that cost efficiencies can be achieved in test environments by applying start/stop schedules to non-production resources. However, applying a default 8.00 a.m. to 8.00 p.m. schedule (or similar) still results in a considerable amount of waste. It is usually worthwhile to spend a little time finding out when the non-production resources are actually being used and scheduling start/stop times based on utilization.
Make sure Committed Use Discounts are being fully utilized
Committed use discounts (such as Reserved Instances) are meant to reduce costs but inefficient reservation management results in missed opportunities for cost savings and wasted spend. Managing reservations has become more complicated since the launch of Convertible Instances (AWS) and the Azure hybrid benefit. Cloud providers offer tools for managing their reservations but they are only useful for their cloud. For hybrid and multicloud environments, admins need a tool that can optimize reservations across their cloud ecosystem.
Move infrequently-accessed data to lower cost storage tiers
Despite the emergence of intelligent storage services, it may still be worthwhile reviewing data access metrics to identify opportunities to move infrequently-accessed data to lower cost storage tiers - particularly the colder tiers which tend to be ignored by intelligent storage services. Reviewing data access metrics might not only identify cost-saving opportunities, but is a good idea to ensure data is stored in compliance with data protection regulations.
Ensure you have total visibility across your hybrid environment
In order to check off the items on our hybrid cloud cost efficiency checklist, it is important to have total visibility across your hybrid cloud environment. Some of the items are notoriously difficult to find using only Cloud Service Providers´ native tools, and native cloud tools don’t provide visibility into on-premises resources (except in the case of VMware Cloud on AWS which provides consistent operations for VMware Cloud resources deployed on-premises and on AWS).. To ensure you have total visibility, speak with us about the CloudHealth cloud management platform.
CloudHealth not only provides total visibility, but also has policy-driven automation capabilities that can be configured to deal with such tasks as releasing unattached IP addresses and terminating zombie infrastructure. CloudHealth can also be used to efficiently manage committed use discounts and infrequently-accessed data, and to provide user-defined recommendations for rightsizing and scheduling. Contact us today to find out more.