Avoid public cloud repatriation: The need to create a workload placement strategy
The reasons organizations turn to the public cloud are well established.
It’s simple and easy to deploy; it offers agility and efficiency, allowing for more time to focus on your core business; and it relieves the burden of having to manage your IT infrastructure. The public cloud can give you all these things – when it’s the right place to store your applications and data.
Recently, however, more and more companies are realizing that they may have jumped to that conclusion too quickly.
According to a recent IDC survey, over 80% of organizations reported repatriating workloads or data from the public cloud to a private cloud or on-premises data center.
Cost plays a big role in this decision, as the public cloud can be more expensive than anticipated for certain data and workloads. Pricing options may vary depending on the types of storage tiers you use. Data transfer charges can be unpredictable. Running applications that aren’t architected for the public cloud can also lead to increased bandwidth consumption, resulting in higher fees.
This trap can happen to anyone. Luckily, it’s also avoidable. This starts with optimizing workload placement.
Developing a workload placement strategy
An optimal workload placement strategy begins with understanding the cost implications of running a workload on various infrastructure, from the public cloud to your on-premise data center. For starters, ask the following questions:
- What will it cost to stand up your workload?
- How much will it cost to maintain and operate your workload at the level you want?
- What will it cost if you want to move the workload – to another public cloud, a private cloud, or an on-premises solution – in the future?
Don’t make any decisions until you understand the total cost of ownership (TCO) first. You may think running this workload in a public cloud offers you the best savings, but unless you weigh the total potential costs against those of a private cloud or on-premises solution, you may later regret your decision.
Outside of cost, you must also evaluate security. How is your data exposed if it’s stored in the public cloud? What would your exposure be if the data was leaked or stolen, and how would you mitigate the consequences of such an event?
By performing a thorough examination of your infrastructure, determining available options through cost comparison, and identifying any potential security concerns, you can determine the best location for your data.
Consider working with a partner
Many organizations are realizing they can attain significant savings through public cloud repatriation. However, if you don’t have the resources or expertise to evaluate your workload placement yourself, you can always turn to a partner.
We recently worked with a large telecom company that placed a nonoptimized version of an application in the public cloud. Instead of rearchitecting the app, the company moved it to an on-premises solution and reduced what it had been spending by 30%.
By looking at the cost model for workload placements, we were able to help this organization avoid cloud cost overruns and attain maximum value for the future.
Stop the problem before it starts
Not only are over 80% of organizations migrating their workloads away from the public cloud, but according to another report, 41% of IT decision-makers feel that their private cloud implementation costs are less than an equivalent public cloud setup.
However, the only way you’re going to know the truth is by understanding the cost model for a workload placement first. That way, you can avoid any problems before they start.
About the author
Jason Lamb is a Cloud Solutions Specialist at Intel with over 30 years of enterprise data center infrastructure experience. In his role, he advises Intel’s end-user customers, partners, ISVs, and OEMs on best practices for developing cloud and cloud management strategies and architecting and designing data center modernization blueprints with software defined infrastructure projects.