Author Topic: Licensing "As a Service" Metrics for Cloud  (Read 2452 times)

wiredlife

  • Hero Member
  • *****
  • Posts: 934
    • View Profile
Licensing "As a Service" Metrics for Cloud
« on: February 03, 2016, 10:52:16 PM »
Licensing "As a Service" Metrics <-- This is a Note to Self

Having Worked on the C/C++ compilers I was always made to think about the extra machine cycles [CPU Cycles] each line of code took to execute. For example, at a higher level, if you were asked to reverse /  List of 1 Million, I'd switch the pointers not the actual data as it maybe a penalty on the CPU and Memory.

With this background I began to think why not license software based on usage but how ? At the lowest level it could be usage of CPU Cycles / Memory consumed, then on the little higher side actual execution time, time until the instance is in memory / bandwidth consumed, so on and so forth until the highest point dedicated allocation of resource. Measuring each CPU cycle would be a penalty on self it is analogous to running Code Coverage Test where the Software bits are profiled and it takes at least 2x times more to run.

So the actual technology that one could commercialize was not at the lowest end i.e CPU Cycles nor at the higher end [dedicated machines] but took a middle path, re-invented the programming paradigm to ensure scalability and companies developed own metrics to compute usage - Amazon Compute to Google App Engine. So far the latter appears to me as one of the best available at the cheapest price but totally depends on the application that you are developing.

Gist of it, as a Product Engineer I'll have to hack and find this Patch of commercialization in licensing "Technology as a Service".