In our first blog for Cloud pricing, we talked about the challenges of estimating cloud pricing for IT. PRICE® Research has developed the IT Cloud Pricing Model based on open source intelligence (OSINT) from six major providers: Amazon, Microsoft Azure, IBM, Google Cloud, Oracle, and Alibaba. The data set contains pricing for Virtual Machines and Storage. This data has other characteristics such as amount of memory (RAM), number of Central Processing Units (CPUs), operational requests, snapshots, and discount rates. Below is the breakdown of over 28,000 datapoints over the following cloud service categories:
Not only is this data available in two PRICE® TrueFindings™ Files (Virtual Machines and Storage), it will soon be integrated as calculators in the Infrastructure Services Operations/New Projects cost objects which covers Infrastructure as a Service (IaaS) costs. As all cost estimators know, traceability of CERs (cost estimation relationships) and models is key. So, the CERs and models used in the cost object can be obtained by looking at the “Findings” section in the TrueFindings data. For instance, to find the CER used for Virtual Machine cost with no local disk in the Virtual Machine TrueFindings file, access the findings section and choose the first row in the section:
Next, head to the “Multicurve Finder”. The CERs in the models are from Lognormal transformations, so choose “LN On Demand Price (Hourly, Linux)” as a dependent variable and “LN CPUs” and “LN Memory (RAM) in GB” as the independent variables. Pictured below is the same CER used in the model:
Users can also choose to focus on the datapoints that are most relevant to their own studies which allows them to generate their own CER. For example, if the user knows they want to use Amazon services, then they can filter the data set so that they only see the Amazon data points and the CER will pertain specifically to this provider.
A Component Database was created and uploaded to the cloud. The database contains data pertaining to a model and is linked to the TrueFindings™ file mentioned above. The cost object is linked to one or more “Findings” which filter the data to show exactly where the CERs and average prices in the model come from. IT Cloud Pricing Component Database: