CAPEX vs. OPEX: Managing the economics of cloud
You don’t have to spend a lot of time with cloud computing to come across debates about cloud economics, but then you’ll certainly hear the word “Capex vs. Opex.” This comparison relates to the fact that maintaining your own data center necessitates a capital investment while utilizing cloud migration services that provide pay-as-you-go service necessitates ongoing operating expenses: thus, the term “Capex vs. Opex.” The next round of the debate instead devolves into a squabble about which option is the most cost-effective.
There have been several debates contrasting the cost of running an Amazon EC2 instance 24 hours a day, seven days a week, versus the cost of operating a server in a company’s data center. Often, people divide the retail sale price of a 1U server by 36 (that number of months in a piece of equipment’s normal planned operating life) to prove that it costs less per quarter than renting from Amazon. As a result, they argue that cloud migration services will still be more costly than self-owned, making them unsuitable for traditional enterprise apps that need 24×7 availability. The faint ad hominem assault that, since cloud services want to make a profit, they were more costly than internal computer systems is often tossed in.
A highly critical aspect revolves around the debate of where the real expense was made. They are then compared to internal data centers against the external cloud. Comparing the monthly cost of an EC2 server to a supposedly comparable piece of hardware in a data center is simplistic since it ignores:
- The direct costs of running a server, such as electricity, floor space, and storage, as well as the IT operations required to handle those resources.
- The server’s indirect costs: memory and communication infrastructure and IT operations to handle the overall infrastructure.
- Server ownership expenses include procurement and accounting resources and a vital resource in short supply: IT leadership and attention.
As these costs are applied to an internal server’s expense, the hosting server skyrockets’ monthly total cost. ACCORDING TO THE LATEST UC BERKELEY CLOUD COMPUTING PAPER, the RAD Lab reports that cloud providers have 75 to 80 percent lower costs than internal data centers (which I addressed last week). Some of this advantage of cloud migration services stems from buying power gained by volume, others from more effective management, and the fact that these companies are run as profitable businesses with a keen eye on cost.
As a result, most cost comparisons between internal data center and cloud service costs are oversimplified and fail to allocate a true cost structure to the equation’s internal data center side. A blog post by Steven Oberlin, Lead Scientist at Cassatt, reflecting on a recent post about cloud TCO, provides another viewpoint on why this analogy fails. He pointed out that these types of cost comparisons of cloud migration services neglect the internal server’s usage with cloud migration services. If it’s operating at 20% utilization, the actual price of a given amount of computing is five times greater than what’s usually assumed in these cost comparisons.
Even so, this does not go into detail on why IT companies and, more importantly, senior standard business governance is involved in cloud computing. This is where we enter CAPEX vs. OPEX territory. Even most cloud proponents who preach about the OPEX benefits of cloud computing miss the full spectrum of factors computing is appealing in this regard. On the other hand, I’ve heard people dismiss the distinction between the two forms of cloud computing expenditures as irrelevant; after all, they argue, it’s all about cash flow, and whether the money is spent on capital good or an EC2 bill, it’s still the same sum of money. (Of course, this ignores the fact that they’re actually not the same sum of money, but you get the idea: the assertion is that, given a fixed amount of money, the sort of “bucket” it falls from is irrelevant.) This is completely incorrect, and it fails to explain why computing would be so common among senior executives.
A better understanding of the importance of capital expenditure in companies should be the essential aspect of cloud computing’s OPEX character. Companies are constrained in the capital expenses they can make by the public markets. Since capital investment is small, businesses typically want to direct their investment into income-generating activities. This is so many businesses rather than buy real estate — they don’t want to tie valuable resources into dead properties.
Rightly or wrongly, many enterprises regard IT as the latter investment type and control it to mitigate its costs provided by cloud migration services companies. That is why in many businesses, IT reports to the CFO.
With both of these considerations, attempting to combat cloud computing by comparing the expense of running an internal server to a cloud-based server’s cost is misguided. Several appealing facets of cloud migration services in the USA would lead top general staff to see it as very enticing until the cloud figures are considerably higher. Identifying judgment requirements for deciding when a specific program could be hosted locally or migrated to a cloud environment would be a far smarter approach. A portfolio review may be carried out using specified parameters to generate a collection of guidelines and an action plan. However, proving that internal servers are cheaper is a failing tactic that reminds me of complaints I’ve learned from HR, legal, and other management groups right before their tasks were outsourced to vendors who provide fixed costs, greater accountability, and greater stability.