My recent interest and thought process about the data warehouse, BI & related domains led me to a very interesting discussion with one of my long time friends today evening. Must say that it was a productive evening spent after ages!
We were actually discussing the nuances of the data storage and auto-scalability of the database designed for the highly ambitious Unique IDentification (UID) project of Indian government. With more than 1.2 billion people spread across a geographical area of 3,287,263 square kilometers, data collection would be a huge problem no matter how many parallel processes you run. By the time you collect the data of an area, there would already be a new person on-board! Hence you need to get into the depth of analyzing the frequency of birth of a given region (assuming that there would be region wise managers appointed for monitoring and that they would do they job J). A simple query run over a few days (however optimized the query might be) will give them this result; no hi-fi statistical tools are required.
Luckily for them (or may be planned - not sure though), the 2011 census is just on its way. Hence getting the basic information about majority of population should cover up in the cost allocated for census (or have they considered this streamlining - well, not sure again!). But, the major task is to collect the iris details, finger prints and other details which the census would not cover. Now, there is an interesting scenario - Say the above mentioned query is run; even then, they have to re-do this analysis of who is the new born? In a country like India, getting a birth certificate is not a mandate for a lot of communities! Over that, there are a lot of mishaps which happen in the due course of issuance of the birth certificate (due to government job issues etc). So, this leads to a lot of data inconsistency w.r.t a real time physical data. This is where government policies come in handy. One very straight forward solution which strikes me is the mandate of UID at various critical point of contacts (banks, Income Tax, Ration Cards, Voters ID cards etc) in which case atleast some sort of inconsistency is covered. But again, there is still a work around for this. No prices for guessing - CORRUPTION of course!
This is when the equation in my mind got really complex and I asked myself - 'UID - Is it really possible to an extent what it is dreamt about? Are they really going to make it large?’ - Well I really hope they do so; time will anyway answer the question!
PS: I haven’t read any facts about UID as such. Just penned down my thought process for today. Would love to know more - technically, politically and managerially.
