When considering ROI maximization of cloud computing solutions it is crucial to minimize costs of cloud hosting, especially that differences can be very big. Anyone who has tried to compare prices of IaaS providers knows that it is a very tedious task if done manually, since there are different pricing models used, there are subscription plans available, etc. It really requires some time to understand a provider’s pricing model. Probably most of users will give up just after few providers. But there are several ways to deal with this.
Probably the most popular way of finding the cheapest provider is studying articles that try to compare several providers or “which provider is the cheapest” forum questions. These are the most natural way of sharing ones experience but still it is only individuals’ research result and what is the most importantly – matching their
needs. Prices of cloud computing are dependent on many factors, like size of a cloud, proportion of resources need (one may have cheap RAM, while other have storage), bandwidth requirements or if you need on demand or subscription instances. So there is no universally cheapest provider but rather cheapest for the particular needs.
A one step further is Cloud Price Calculator. Unlike the name suggests its approach is to define a Cloud Price Normalization index, which is a sum of resources divided by price. Unfortunately it has just several configurations from few providers, which does not cover full space, and you can only provide your own resources with price (which you need to calculate on your own), to get the index calculated so you can compare it to others.
But the ultimate solution may bring Cloudorado – Cloud Computing Comparison Engine. With this service you provide the resources you need on your cloud (RAM, CPU, storage, bandwidth, OS) while it will calculate the cheapest option for each provider and give you the price for each of them. It will even take into account packages or subscription plans, if you provide maximum plan duration. In advanced mode you can also calculate more complicated models with multiple servers, servers not running instantly or tell the engine to find best combination of servers to get a summary of a resource (scale horizontally). This allows you to really compare prices and find the cheapest option for your specific needs.
Also an interesting option is recently described Python library to calculate cloud costs. It takes the required resources either from monitoring tools or from a specified usage simulation. In this way you can calculate price on your historical data or simulate the future load. Unfortunately you will need to manually program
the library and what is probably even more limiting – currently it will calculate costs only for Amazon and Rackspace.
There is couple of ways of finding the cheapest cloud provider. Regardless the way you choose – manual calculations, articles or any way of automatic price calculation (Cloudorado or the Python library) keep in mind that there is no universally cheapest cloud provider but the cheapest providers for specific needs. And take it really seriously – difference between the cheapest and
the most expensive can easily exceed 10x while the cheapest provider in one configuration might be the most expensive in another!