https://orcid.org/0000-0002-9097-2246
[This post builds upon introducing Aadhaar,its size,current way of handling the data sets,discuss its problems and subsequently followed by proposing Blockchain as a solution]
1. When Aadhaar was originally introduced around 2009-10 by the Unique Identification Authority of India (UIDAI),it would not have envisaged the kind of Data juggling,analytics and security threats it would be subjected to in times to come.And here we are around the third quarter of 2018,wherein Aadhaar is central to so many authentications in the country ,being exploited in so many public utility services and also at the same time being subjected to all kind of threats and claims of data theft and leaks.For a record,it is estimated that around 1.2 billion citizens record are held in the CENTRAL servers and thus forms the worlds largest bio-metric identity repository in the world.UIDAI claims that the same is protected by layers of state of art cryptography in central servers located in the country.
2. Now in the world of IT,wherein claiming to be 100% secure is likely to remain a myth for ages ahead,can something like un-hackable really exist on this earth? We may harden something,we may actually add layers of security, we may do every possible hard encryption on this earth,but can we imagine a fool-proof IT domain anywhere. The question here attains severe importance when a Bio-metric repository data of 1.2 billion plus population of a country is at stake.
3. Now what do we have on the platter here,if we consider the size of data,we can have the following assumptions :
(a) Per person biometric data size : 4-6 MB (Maximum I take)
(b) Approx data populated for : Around 1.25 billion plus ie 1,250,000,000 count
Total data ie to say 6 MB x 1,250,000,000 = 7500000000 MB Data ie around 7.5 Petabyte.....that's it...extrapolate the same with on-site backup and mirrors around...disaster recovery sites...we may just be discussing around 20 PB of data.
Even if we consider,augmenting data with the remaining population and generations ahead,we will be at max around 40-45PB of data to suffice around next few decades.That's all from point of view of the scalability of data and size.
4. Now for this amount data, what are our security options in the present scenario.
Firstly we keep doing permutations and combinations and applying layers of hard coded security to the central servers that we have at various locations mirrored to each other.This presently includes the following : [SOURCE : http://www.cse.iitd.ernet.in/~suban/reports/aadhaar.pdf]
- 2048 bit PKI encryption of biometric data in transit. End-to-end encryption from enrollment/POS to CIDR.
- Trusted network carriers.
- Trusted network carriers.
- Effective precaution against denial of service (DOS) attacks.
- HMAC(keyed-hash message authentication code) based tamper detection of PID (Personal Identity Data) blocks, which encapsulate bio-metric and other data at the field devices.
- Registration and authentication of AUAs.
- Within CIDR only a SHA-n Hash of Aadhaar number is stored.
- Audit trails are stored SHA-n encrypted, possibly also with HMAC based tamper detection.
- Only hashes of passwords and PINs are stored.
- HMAC(keyed-hash message authentication code) based tamper detection of PID (Personal Identity Data) blocks, which encapsulate bio-metric and other data at the field devices.
- Registration and authentication of AUAs.
- Within CIDR only a SHA-n Hash of Aadhaar number is stored.
- Audit trails are stored SHA-n encrypted, possibly also with HMAC based tamper detection.
- Only hashes of passwords and PINs are stored.
- Biometric data are stored in original form though.
- Authentication requests have unique session keys and HMAC.
- Authentication requests have unique session keys and HMAC.
- Protection against replay attacks.
- Resident data stored using 100 way sharding (vertical partitioning).First two digits of Aadhaar number are used as shared keys.
- All system accesses, including administration, through a hardware security module (HSM) which maintains an audit trail.
- All analytics carried out only on anonymized data.
- Resident data stored using 100 way sharding (vertical partitioning).First two digits of Aadhaar number are used as shared keys.
- All system accesses, including administration, through a hardware security module (HSM) which maintains an audit trail.
- All analytics carried out only on anonymized data.
From the IT guys perspectives,don't we actually know that above are all individual knitted layers and tools of security wherein we are creating a very complex network of solution for ourselves which might get even more complex to handle and manage in times to come with more severe security threats in pipelines.
At the same time, above all solutions and knits combinations are looking and bracing for external threats while we take the insider threats as negligible or taken for granted any day.
So do we have any other ecosystem of architecture that turns the tables upside down from the security and immutability point of view while OFFERING A MORE ROBUST SECURE IMMUTABLE AND TRANSPARENT ARCHITECTURE...whether BLOCKCHAIN can be a solution?
So,we have the above scenario which discusses what do we have on the platter and what are we actually doing to negate the threats....the next post will discuss how BLOCKCHAIN can assist to negate the security threats Aadhaar faces as on date.
