"You are asked to design an application to process mobile phone usage data on behalf of a large global telecom provider. The telecom provider wishes to know their top subscribers by revenue based on each subscriber's data usage. The raw usage data resides in thousands of log files produced daily. Each log file can be up to 256MB in size and they all contain 1-line data records, each representing 1-minute of data usage for a single subscriber. Each record includes 10 numeric fields that measure the data usage across various usage categories. The record also contains information such as the subscriber ID, and timestamp. The files are compressed and stored in an S3 bucket organized by year, month and day. The revenue calculation requires a number of computations to combine and price each of the 10 different usage categories, but each data record can be priced independently. You are going to solve this problem using the Map/Reduce pattern. Submit a copy of this M/R spec template: http://bit.ly/2W2TUVA describing how to implement the various phases of Map/Reduce."
