A pet peeve I have with cloud services is that the tutorials and “wizards” are not always practical. The configurations these tools suggest or use by default may be functional, but they also tend to be a bit expensive. There are the best solutions, and then there are the practical ones. It is essential that you understand your requirements to estimate realistic cloud pricing for your solution.
Peaks and Valleys
One of the most common issues I have seen in estimating cloud resources is a focus on peak usage rather than average or typical. You will end up with a worst case scenario and not realistic cloud pricing if you base it on peak needs. The very power of a cloud solution is being ignored in that situation. Embrace the options you have to dynamically size your cloud resources. This not only provides you with more power than on-premise solutions, but it also can reduce costs dramatically.
Look at your typical needs for setting up an environment. A lot of the tools and wizards do not take this into account properly when estimating pricing. A perfect example of this is the AWS WordPress example. They will generate a suitable solution for your solution. However, it will be far above what a typical user may need. This is to the tune of $250/month for the AWS solution created by the wizard as compared to a perfectly suitable solution at $10 or less each month. I even used the free tier for a year with an active blog which is much better than $250 a month.
Typical vs. Average
If your peaks are rare and short-lived while being very high then they can throw off your average. This is simple math but a common mistake. When you price to your typical usage then you are more likely to determine your realistic cloud pricing instead of over-compensating for those peaks. The thing to consider in these instances is that your peak needs may trip a threshold of requiring another server or processor. Instead of having that additional resource on hand all of the time, just bring them online for that short, peak need.
You might also look at how you process as a way to smooth out that peak. This situation often appears with batch processing. Your systems may take up a large number of additional resources when they do nightly or monthly processing. If you have a way of spreading out these processing steps then you may be able to reduce the peak need and keep costs down. Sometimes avoiding a high peak alone can reduce costs dramatically.
Look Before You Leap
The bottom line in determining realistic cloud pricing for your situation is to ask questions and understand the answers. Think of it as a fitting for a custom suit. You need to spend some time and take proper measurements in order to get the best fit. A cloud solution works the same way. Whether you use a consultant, a tool, or your own estimates it is best to measure twice and cut once. The good news is that you are (hopefully) not locked into anything once you make your cloud decisions. Try out your approach for a while and adjust your resources after the reality sets in. Embrace the dynamic nature of a cloud solution to find the best fit for your needs.
