The Bureau of Labor Statistics Producer Price Index (PPI) reported a 3.9% month-over-month decline in the cost of host computers and servers. Cloud services saw prices increase by 2.3%. This happened after the third quarter of 2022. The overall PPI declined 0.3% in May as prices for goods dropped 1.6% and service fees increased 0.2%. The Bureau tracks cloud computing within the data processing, hosting, and related services category.
Hitting the reset button
My latest book explains that cloud computing should never be the go-to consumption model for all applications and data sets. Many companies are hitting the reset button for some systems already on cloud platforms, repatriating them back to traditional data centers for more attractive costs.
Ten years ago, the reasons for moving to a public cloud consumption model were solid. We avoided the costs of hardware and software, the pain and expense of maintaining these platforms, and leveraged computing and storage as a utility. The sell was operational cost savings, first and foremost by avoiding many capital expenses (capex versus opex). The secondary benefits of agility and speed to deployment were icing on the cake.
Of course, things are never that simple when it comes to new technology. We encountered many unexpected challenges, namely, the freefall of prices for data center hardware at the same time cloud computing service costs ticked upward. This is enough to make CFOs pause to rethink the wisdom of moving all IT assets to the cloud if there is no significant cost advantage to outweigh the migration risks. Today’s business cases are more complex than they were just a few years ago.
Weighing the advantages and disadvantages
This is more complex than just looking at the costs of each deployment model, even with the new lower prices for hardware. If you make decisions strictly based on actual expenses, you’ll likely miss a lot of business value that is much harder to understand. Also, if you examine each application and data set only as to its specific requirements, you’re likely missing larger strategic forces.
Do you have applications and data sets with repeatable patterns of compute and storage usage? Are you unlikely to rapidly scale up or down? These are excellent candidates to reside on traditional servers you own and maintain.
We now have examples of companies that started on public cloud providers and relocated back to traditionally owned hardware. The successful transitions had well-understood compute and storage requirements and easy-to-plan demand. These specific patterns take advantage of lower-cost owned hardware and use the recent price drops to a business advantage.
However, not all applications and data sets fall into this category. The more dynamic the storage and compute requirements for an application or set of applications, the more likely public clouds are a better option. Scalability and seamless integration offered by cloud services are critical for these types of data and computing, as well as the ability to quickly expand and build on public cloud services with other native services.
There may be a short-term advantage to owning hardware for some applications. However, as application requirements shift and expand, the public cloud becomes cheaper in the long term. Regardless, the amortized costs of traditional on-premises hardware solutions are more affordable. The advantages of using your servers for computing and storage can diminish rapidly when the requirements for these resources change frequently, which is often the case.
What to do?
Cloud computing should never be a default solution. After considering your business requirements, you should question the advantages of any technology solution. As mentioned earlier, there are benefits to utilizing resources in your own data center but also reasons to use public cloud providers.
As the prices of traditional servers decrease, it is important to consistently reevaluate our objectives for utilizing these systems. There will be ongoing questions about the ability of traditional servers to adapt and integrate with other service types. Some evaluations will point to public clouds as the platform that will return the most value to the business, but that won’t always be the case. Keep those evaluations up to date.