When on-premises IT was the only game in town, corporate IT had to invest long-term. With the public cloud, there is another option. While there are many positives in shifting to public cloud, some organizations are better served by maintaining (and expanding) corporate IT asset portfolios.
OpEx vs CapEx debates are about public cloud spend
A conversation about OpEx or CapEx for IT investments is a proxy for the choice between using on-premises infrastructure or the public cloud. Traditionally, project costs fell into CapEx spend with attendant depreciation showing up in OpEx: the CapEx budget was the investment budget. Public cloud options have upended IT investment approaches.
Satisfying project demand for virtual compute resources with public cloud passes service delivery responsibility to a third-party vendor and limits the financial commitment to the length of the project. Organizations now only need to plan (and pay) for what they need, and no longer worry about how they will maximize an asset over its usable life.
»Related content: 5 capabilities of effective cloud cost management
Public cloud changes the role of I&O
Infrastructure and Operations (I&O) responsibilities are altered by OpEx-fueled public cloud spend. An on-premises IT footprint needs dedicated resources (people, hardware) to build, optimize, and maintain capacity. Technical, focused work has always been in I&O’s remit—and is the reason many I&O professionals were drawn to the profession in the first place. But with public cloud, I&O leaders are pulled away from those operational responsibilities to become more focused on vendor management.
There’s nothing wrong with this transition, but not everyone in I&O wants, or has the skill set, to embrace this shift.
When IT Finance discusses OpEx or CapEx choices to fund innovation, I&O hears a debate about the very nature of their work.
A public cloud investment strategy only works when I&O understands and can advocate for those changes. OpEx-fueled innovation fails without the support (and the operational data) of I&O.
»Related content: Demystifying cloud costs
Start-ups love OpEx
CapEx investment is defined by the long-term: a depreciation schedule for a server is at least 36 months. But many organizations—think start-ups—struggle to think three years ahead about anything. If the goal is to grow market share from zero, you aren’t looking to be distracted by locking capital into fixed assets which, if the company sinks, will be sold off in a fire-sale.
OpEx investments align to projects with compressed timelines and organizations with short-term priorities, which is a perfect fit for organizations focused on growth.
»Related content: Optimize your IaaS footprint by focusing on business outcomes
Define operational goals or over-pay for public cloud
When organizations use on-premises, CapEx-fueled infrastructure, asset costs are already defined. The public cloud is different—usage drives cost.
Public cloud bills cause organizational turbulence when IT costs aren’t connected to consumption. Take archive storage. On-premises archive backup storage utilization rates may vary, but there isn’t a financial impact—the 36-month depreciation on the general ledger is predictable and non-contentious. Compare that with public cloud. There are standard /GB storage rate, different retrieval service options (e.g., expedited, standard, and bulk) and a /GB upload rate for data—any combination of which could be right for your organization.
Agility, a compelling value prop of public cloud, is tempered if you over-pay for a service you don’t fully understand.
It’s in your interest to work out the best options before signing up for the service.
»Related content: Embracing a hybrid mindset
Retain strategic independence
When your organization plans for the long-term, CapEx investments protect market share and competitive independence. Cloud service providers are building massive infrastructure footprints to satisfy public cloud demand. Every service you utilize is increasing their market share and making you more dependent on a third-party to deliver value to your customers. This is the same with any managed service/service provider relationship, but public cloud services can become so embedded in business processes that the promise of agility and lack of long-term commitments turn out to be empty.
Applications must be suited to the pricing model of public cloud. Operational metrics (uptime, disaster recovery SLAs) between on-premises IT and public cloud services are equitable; apples-to-apples comparisons between application portfolios are more complicated. For example:
- Existing legacy applications may not run in public cloud environments—or not at a unit rate compelling enough to justify migration.
- Compute resources may suffer from bursts in usage patterns (e.g., ecommerce sites around Black Friday/Cyber Monday) that are covered by existing on-premises capacity but generate excess usage fees in the public cloud.
- Standards of latency (i.e., response time between a request for access and response from the IT resource) will compromise legacy application performance in the public cloud.
The public cloud gives organizations an OpEx-fueled IT investment option they didn’t have before. Agility and the lack of long-term commitment are compelling, but when public cloud becomes embedded into business processes, you may pay for it 24x7 at (possibly) more prohibitive unit rates than if your apps and services were run on premises.
There is no correct blend of public cloud and on-premises IT to deliver IT—your mileage will vary based on usage, application portfolio, and use of underlying infrastructure.