My BMW is ok, but it’s not new. It’s now of an age that gets me thinking about replacing it, but for the last too many months my marvel of German engineering has hardly moved. I’ve realised that for three months it’s only been used to go and get a takeaway of a Friday evening. Hardly much of a reason to buy something newer. In fact, a quick rough and ready calculation totted up monthly outgoings on the thing of £400 or thereabouts, which meant that my personally owned transport was adding a hefty £100 to the cost of each takeaway. Madness! Perhaps I should use a taxi on a Friday night and avoid all the cost and hassle of ownership, paying only for what I use, as and when I need it.
Of course, taxis won’t make sense if routine usage levels return to anything like they were before. Getting a taxi to the train station and back two or three times a week, plus the 200m round trip to one board meeting each month, and 60m for the other, plus a few other bits and pieces, would push the cost above £750. Then there’s a clear justification for having a private vehicle at more like £400 a month.
As with cars, so too with Cloud.
It sounds like a no brainer to pay only for what you use, but if the cost of that flexibility is high, and you have a routine level of usage that reduces the amount of flexibility you actually need, then the economics just aren’t there. Once the routine usage rises, it becomes too expensive to use a service priced for flexibility. When this happens it pays to shift to a mixed architecture, and put routine, predictable workload where that’s most economic, and put spikey, unpredictable demand out to the Cloud where that spot capacity justifies what it costs. This mixed model is hybrid Cloud.
The two week holiday
Different situations call for solutions that fit. Much as it makes sense to rent a car for a two week vacation, even when the usage could be high and consistent, a project could easily justify Cloud infrastructure, because the flexibility you pay for gives you the freedom to hand it back when the project is over. It’s a financial calculation that determines the right way to go.
The Start-up trap
Naturally enough, many start-ups have little real usage when they begin, and can’t easily predict how usage will grow as they add customers and new markets, because so many of the factors are yet to be proven, such as the actual rate that new customers will come on board. This makes the flexibility of Cloud pricing really attractive, and some brands make it even more attractive by giving some away free at the beginning. Storage is often free or almost free. What could be better?
At the beginning this is all really sensible and completely justified, but that picture can change alarmingly quickly when the realities of growth and real usage begin to bite. Storage might be incredibly low cost but there are charges to get the data back. That’s like having to pay whenever you want to get the car out of the garage. The “free” credits are fine until they end, and then the realities of pricing kick in. By then, what began as occasional bursts of usage and low levels of overall activity, has often grown to be solid usage over time, and uneconomic at prices set for flexibility that just isn’t needed.
The worst aspect of this is that by then it can be too late; data volumes may be too big to move without large charges; particular quirks of the Cloud provider may have become embedded in the service making a move overly complex and risky.
Solo Cloud, Multi-Cloud, Hybrid-Cloud
These architectures all have technical differences, advantages and challenges, but they each have very different economic and business implications, especially relating to the dynamics of price and usage, and the the cost and feasibility of change. Different providers have subtly different pricing models at speed, as well as different structures and variables. These things are all typically in the CFO’s wheelhouse and often obscure to those thinking primarily of technology. Perhaps the CFO is the first port of call when trying to decide Cloud strategy and atchitecture for the future?
Peter is chairman of Flexiion and has a number of other business interests. (c) 2021, Peter Osborn