top of page

Sprinting towards 1.2 billion

The latest from India presented both by Daon’s Cathy Tilton at last weeks BCC in Tampa and reported in the news is that

One year after launching the gigantic enrollment process, 37 million people have been issued biometric identity numbers. 60 million others have enrolled and will be given the numbers shortly. And starting next month, one million people are expected to register every day for the biometric ID.

While I appreciate the optimism expressed, 37 million ( or even 97 million if enrolled but not issued are included), this is a far cry from 1.2 billion. It is not just the challenge of maintaining a consistent level of 1 million “quality” enrollments a day for nearly three years non-stop that gives me pause.  Or successfully performing the unprecedented number of  biometric matches required to de-duplicate this data. It the inevitable “breakdowns” that occur within any IT system that relies on exponentially growing databases of this size.

An IT system can hum along and then reach certain data levels at which things just break. Maybe it is 200 million, or 500 million, or 1 billion. Or all three. It could be something simple requiring a relatively quick software fix. Or it could be a substantial flaw requiring significant system reengineering.

This type of inevitable system issue is easily exacerbated by the demands of  biometrics. There is no model for a  biometric database of this magnitude. What if half way through the process, a quality issue arises that limits the ability to successfully continue to complete biometric comparisons?  What if natural limits on biometric differentiation emerge?

While I do believe that biometric identification is inevitable in India and ultimately globally, the UID biometric system is based on untested theory so it might be prudent to temper enthusiasm for this mammoth undertaking. The harsh lessons learned from global industry experience  with delayed, severely delayed, or even failed biometic ID programs ought to provide a healthy dose of skepticism. 3 years to build a reliable, de-duplicated  database of high enough quality to ensure expected outcomes for more than 1 billion people is a sprint where clearly a  marathon is in order.

7 views0 comments


bottom of page