Aadhaar’s architect discusses what went into world’s biggest biometric repository

“Open digital infrastructure has hugely helped societies access information, products, knowledge, and collaborate better with each other,” said Dr. Pramod Varma, chief architect of Aadhaar, at the Future of Work 2020. He spoke about having built out the world’s largest biometric identity system, which now contains the biometrics of more than 1.2 billion people seven years after its launch, at India’s largest product-design-tech conference.

Aadhaar is the world’s largest biometric identity program, or about ten times more massive than the largest digital identity system there is – depending on whether classified systems for intelligence purposes are factored into the equation – estimated to contain the biometrics of nearly 16 percent of the world’s population. Politically and socially, within India anyway, it is widely seen to represent the growing potential to enhance India’s public services programs as well as advancing a trailblazing pathway toward digital and financial inclusion.

In 2014, only four years after its introduction, Aadhaar already had nearly 690 million entries, making it the largest biometric identity repository than six years ago, accounting for 600+ trillion biometric matches each day, or roughly 3-5 MB per person, for something like ten petabytes to 15 petabytes of data.

Today, there are more than 100,000 laptops with Aadhaar installed used in remote villages. In 2014, there were 150,000 certified operators and supervisors trained and certified to operate laptop-based enrollment, averaging at the time roughly 50 people per day, or 1 million enrollments a day.

“As India’s digital identity program, Aadhaar has successfully covered more than 1.25 billion people,” wrote Ryan Frantz for the conference. “It is also one of India’s best case studies of scale. In addition to core digital identity systems, India’s open digital infrastructure, collectively known as India Stack, also includes electronic payment, digital signature, digital locker, and data empowerment as its core layers.”

Using the comparison that “India has benefited from hundreds of millions of people having access to mobile phones and the Internet,” Varma also observed that “it is essential that from time to time many such digital infrastructure layers are built as public goods to create open access and a level playing field for innovators to provide solutions to large, diverse society. India has been building its own layers of public digital infrastructure spanning from identity, payment, health, and education.”

Varma is also the architect of various India Stack layers such as eSign, Digital Locker, and Unified Payment Interface (UPI), all of which “are now working at population scale in India.”

Varma is also the CTO of EkStep, a not-for-profit creating learner-centric, technology-enabled platform that provides learning opportunities to 200 million children in India.

“To understand the scale and success of Aadhaar, it is important to look at the landscape before the program was introduced,” wrote Frantz. “Back in 2008, financial inclusion was really low – just 17 percent of India’s population had bank accounts. And while around $50 billion was being spent on subsidies, diversion, and leakage was rampant. This caused a ripple effect. Getting a loan was near impossible for citizens from the lower socio-economic rungs. Customers with the lowest default rates had the highest interest rates because of no credible data. Besides helping realize 80 percent of financial inclusion in just six years and proving naysayers wrong, Aadhaar helped disrupt the financial innovation through eKYC, particularly in the mutual fund and neo banking arena.”

Franze claims that centralized system like Aadhaar works better than a distributed system built and maintained by banks. “Collecting the biometric information of a customer is a long and expensive process. The high entry costs associated with a distributed system make moving one’s account between banks a cumbersome process. A centralized system like Aadhaar Enabled Payment System (AEPS) makes it easy for a new financial services provider to plug-in and launch its services. A widespread network of agents associated with different banks, operating on AEPS, would expand the choice set for the customer, increase competition, and improve customer service. Flexibility in using the branch, agent, or biometric ATM to access bank account would put such a customer on an equal footing with the customers who access the existing network of ATMs using a card and PIN for authentication. The choice of authentication system would no longer define the extent of access to the banking network.”

Therefore, “India now runs the world’s largest direct cash transfer program with over $35 billion now directly being sent to the bank account of the beneficiary, resulting in savings to the tune of $5 billion, as per government estimates.”

There have been more than 33 billion Aadhaar-based authentications, more than 7.5 billion e-KYC transactions, 200 million micro-ATM transactions, and over 1.3 billion UPI transactions every month at the time Varma spoke.

“Open digital infrastructure” is “a gamechanger for the industry; India is one of the first countries in the world to start Open Application Program Interface (APIs) for consented financial data sharing, and Pramod spoke about how they worked to keep minimalism and openness at the very core,” Franze wrote, saying Varma’s message is “keep it simple and minimalistic” by focusing on only a few parameters of storage rather than the 50-plus parameters some other her countries are attempting, and which, he said, lead to bureaucratic knots.

“Aadhaar also did not invest in field infrastructure but just the strong, secure, scalable backend exposed via APIs and ecosystem partners invested and participated in the fastest growth of a billion users worldwide,” Franze pointe out, concluding that Varma’s “key takeaways from building the India stack is composed of ten strategies or processes.”

The recommendations are that identity systems be;

• Unified, not uniform;
• Solutions should create and coexist with diverse solutions on shared infrastructure;
• Unbundle by creatin smaller digital infrastructure can be applied successfully to multi-dimensional challenges;
• Design interoperability using specifications that connect and combine microservices which can build solutions;
• Adopt open standards and open source;
• Design built-in retractability and observability telemetry;
• Plan scalability at the outset of the design phase by building in resilience for failure contingencies in every microservice;
• Aim at maintaining “data minimalism” and construct privacy and security “by design;”
• Build trust by design through registries, attestations, and signatures;
• The employment of nearly acute automation in which every repeated task and API testing shall be automated; and
• Simplicity or the “elimination of feature creep,” unless required, and then “layer them.”

According to MapR, the “entire technology architecture behind Aadhaar is based on principles of openness, linear scalability, strong security, and, most importantly, vendor neutrality.” The principles behind Aadhaar’s backbone are “open architecture” through open standards and APIs, and “design for scale” with massive deduplication capabilities support for hundreds of millions of transactions a day, which requires “load balancing and multi-location distributed architecture for horizontal scale.”

Article Topics

Aadhaar  |  biometric database  |  biometrics  |  digital identity  |  identity verification  |  India  |  India Stack  |  national ID