A multicloud cost-efficiency checklist is exactly what you would expect it to be, a list of potential cost drivers that businesses operating in multicloud environments need to keep on top of in order to maximize cost-efficiency. However, keeping on top of these cost drivers isn’t all that straightforward due to a lack of visibility.
In 2018, IBM’s Institute for Business Value produced a whitepaper entitled “Assembling Your Cloud Orchestra” based on a survey of more than one thousand businesses in nineteen different industries across twenty countries. 85 percent of businesses surveyed were already operating in multicloud environments and a further 13 percent planned to do so within the next three years.
What was particularly interesting about the results of this survey was that in every category of motive for adopting a multicloud environment, a reduction in costs was rated the most important. Yet, when asked what the most significant challenges were to multicloud management success, the top-rated answer was managing and optimizing cloud costs.
IBM attributes the concerns over cost management to a failure to implement tools suitable for managing a multicloud environment. Fewer than 40 percent of respondents had cloud management tools that provide information about resource configuration and relationships between resources, or tools that provide total visibility of the environment.
The lack of visibility resulted in 59 percent of respondents reporting that independent cloud adoption by business units has already created a de facto multicloud environment. IBM noted that these “shadow environments” were only the ones respondents were aware of. In all likelihood, the true percentage is likely much higher.
Visibility is key to multicloud cost-efficiency
One reason businesses suffer from a lack of visibility is the mindset that traditional on-premises monitoring tools work in the cloud. In reality, they don’t. Techniques like SNMP, flow, or packet captures don’t have access to servers, switches, and routers inside Cloud Service Providers’ data centers, so they can’t provide the full information required to understand relationships between resources.
Similarly, on-premises techniques can’t provide information about resource configuration or what cloud services are being utilized by whom. To get a full picture of a multicloud environment, it’s necessary to implement a multicloud management solution such as CloudHealth that uses agents and makes API calls to identify exactly what’s going on where, when, and how. Therefore, the first item on our multicloud cost-efficiency checklist is to get total visibility of your multicloud environment.
With visibility, cost optimization is more effective
Total visibility not only gives you a better understanding of how resources across multiple clouds communicate with each other, but it also helps shine a light on de facto/Shadow/Line of Business IT environments which need to be brought under the umbrella of central IT. Then it’s possible to better manage the environment and optimize costs more effectively.
Cost optimization is more effective with increased visibility because you’re able to rightsize resources, terminate zombie assets, and manage committed use discounts holistically. It’s also possible to get a clearer picture of non-production environments and schedule on/off times for development and testing resources based on the utilization of these resources. Consequently, the second item on our multicloud cost-efficiency checklist is to optimize holistically.
Uncovering hidden costs that affect cost-efficiency
One further advantage of total visibility is that it enables you to uncover “hidden costs”, costs relating to premium services you don’t need or services you no longer use that are difficult to find using Cloud Service Providers’ tools. An example of this is premium storage attached to instances by default, when standard storage is adequate. Other examples include:
- Unattached storage volumes
- Obsolete snapshots
- Disassociated IP addresses
- Idle load balancers
- Over-provisioned databases
- Infrequency accessed data stored in premium volumes
By downgrading premium services to standard services, and terminating services no longer required, you can significantly reduce multicloud costs. The important thing then is to prevent these costs from re-occurring, which can be done by taking advantage of the policy-driven automation capabilities of a cloud management platform such as CloudHealth.
How policy-driven automation works
The concept of policy-driven automation is quite simple. You give the platform permission to monitor your cloud environment and apply policies for when specific events occur. The policies could be to alert you to a significant cost trend, a budget projected to be exceeded, or a significant increase in storage costs. These events could then be investigated by the necessary stakeholders to see the reasons for the alerts and determine the best course of action.
Other events can be taken care of automatically by the platform. For example, if a storage volume is unattached for two weeks, CloudHealth can be configured to take a snapshot of the storage volume, terminate it, and notify you by email of the event. Similar policies can be applied to other resources and to events such as expiring committed use discounts or over-provisioned databases.
Multicloud cost-efficiency checklist recap
To recap, there are three main things businesses need to do in order to maximize cost-efficiency in their multicloud environment:
- Gain total visibility
- Optimize holistically
- Automate cost management
CloudHealth can help you achieve all three items on our multicloud cost-efficiency checklist, and if you would like to know more about our multicloud management platform, or how to purchase a cloud management platform in general, check out our buyer’s guide.