On my initial post here its a pleasure to submit the following article written by a Tanzanian IT/database expert based in the U.S., who fears there is no sufficient debate on the capabilities of the proposed National ID program. He warns that "We should not go halfway and just create a piece of paper that ends up being sold like a phone voucher on the street and call it a National ID." Read the rest below.....
TANZANIA NATIONAL I.D. DEBATE: A TECHNOLOGICAL PERSPECTIVE
BY A.G. MWINGIRA
I’ve been trying to navigate the issue of a national ID card in Tanzania. I have noticed that in recent months this issue has heavily dominated the media. There are many aspects to this debate. The on-going discussions on who will be awarded the contract, who is vouching for whom, and which leaders will personally gain from the tender award are beyond my sphere of influence. Therefore, I will only base my arguments on the tech/IT realms of it and will leave room for open and constructive discussion on other areas, notably corruption, irresponsibility and mismanagement.
At the end of the day, someone will eventually be awarded that ‘lucrative deal’. Therefore, at this point, what I really care for is how that someone is going to implement the system and sustain it. Unfortunately, so far public interest has not been on debating the benefits of the system, but rather on who will be awarded the project. Clearly, corruption is in our DNA as a society. It links us all in thoughts and instincts, and the consensus among the citizenry is that majority of those in power are corrupt. Ironically, this resentment is somewhat powered by envy - there are enough people out there who praise and talk with admiration on corrupt practices done by some public officials. As Nyerere once said, this is a plague that may haunt us for many years to come.
Moving beyond this, I am a strong supporter of the National ID card system. I think the Government should bring this program to fruition. The benefits in this global age of technology will be astronomical if implemented well and maintained accordingly. If we get serious, Tanzania can have one of a kind National ID card system in sub-Sahara that is systematic and comprehensive. The prospect of a functioning e-government can finally materialize based on this system. From government statistics (administration, taxation, revenue etc), to private usage in banks (loans, credits), hospitals (patient info, billing), by employers (verification, payroll) to the police and Court system (identification, criminal records), and many more, they will find this National ID system crucial in conducting their daily tasks. Don’t forget that non-governmental functions will be charged a fee for data verification. Therefore, this will potentially bring in additional revenue to the government.
I know a thing or two about IT and data management, so I have some worries since we are yet to fully understand the details of the system sought by the Government. As I write this, I’m almost certain that many people think this issue is just a matter of issuing ID cards and that’s it. NO. There is implementation, sustaining, training, security, data access, data backups, and massive storage to name just a few. I believe any ID card system should have the following four key components: an initial individual identity verification system; a database; an ID card and a secondary ID card verification system.
Before a card can be issued, there must be some means of ensuring that the person receiving the card is who he or she claims to be. Any ID system is only as good as its ability to accurately identify people in the first place. Now the question is, how are we going to effectively do this? It is estimated that by 2015 there will be 57 million people in Tanzania, with current growth rate of 1.4% in every 5 years. This will pose a greater challenge in obtaining accurate data at child birth, and taking accurate census of those who are still alive or dying. Most hospitals and clinics don’t keep efficient birth and death data on file. The health centers are generally not computerized and most of the children born in rural areas are being delivered at home with no State documentation whatsoever. Forget about rural, it’s a problem obtaining a copy of a birth certificate even in central Dar es Salaam at the newly restructured “Vizazi na Vifo” State agency. Only a small fraction of the population today carries a passport. BUT we can still do this national ID project if we have a long-term perspective. Yes, we might have false or insufficient data to start the identification process with, but if we put sufficient identification infrastructure in place today at hospitals, clinics, and at immigration centers across the country, going forward the data will slowly become more accurate. We have to start somewhere - government workers will have to be trained and civic education will have to be given on the benefits of good record keeping. If I remember correctly, we succeeded in having a CCM branch in every village; there is no reason why we can’t be ambitious again on a crucial project like this.
The next step after individual citizen identification effort will be to issue the ID cards. Aspects of material, durability and security (photo quality, forgery & tamper-proof) of the card itself will have to be looked into. Point of issuance, whether by Government or private contractors will have to be ironed out as well.
Thirdly, a database with massive storage capacity is the backbone of the system. The government should oversee who implements the best Storage Area Network (SAN). This is a whole different ball game here. A private company could be contracted to manage this SAN, or perhaps the same company that will build the system can be extended to manage it. Whatever the case may be, how we design and manage the SAN will help in the long run to avoid incurring massive operation costs. We need a unique architecture that can manage data inside effectively. We also need a scalable hardware to eliminate the need for time costing upgrades, and powerful software for optimizing it. Also, this company will need to be fully vetted and trusted to manage this sensitive national information. The contractor should be able to set, grant and manage data access for commercial use without compromising national security and privacy rights.
Deriving any value from building enhanced high-tech security measures into a national ID system will require a massive secondary ID card verification architecture. Putting a microchip on an ID card, or having some way of linking it to the SAN database will complete the project. Agencies such as the Police, the CID, the Courts System, TRA, and Immigration, as well as private entities such as Airlines and the banking/ATM/credit card sector, or any person for that matter who should have cause to inspect the ID card should have a machine capable of reading its advanced features. So, for instance, we are talking about high speed computers in every Police station in Tanzania, or computers in their vehicles, that are linked to the national database. We also need to implement a system with forward compatibility. For example, when we finally implement the infrastructure to the police, it should be capable of “plug and play”, meaning it should require little or no hardware upgrade to avoid unnecessary and wasteful future projects. We should train our “finest” police officers on how to operate these computers. We should not go halfway and just create a piece of paper that ends up being sold like a phone voucher on the street and call it a National ID. We really need to stay the tech-course with utilization, automating, replication and speeding up data recovery, or else we decide to put this project on the back burner until we are ready. We can have the system built today and completed tomorrow but the major huddle will be sustaining it.
I hope that the ministries of Home Affairs and Science and Technology respectively are collaborating with the University of Dar es Salaam’s IT department. This is sensitive national data that the Government will have to protect, so whatever IT technology is to be deployed it needs to be secured and robust. Of course, the building inspectors need to make sure the hardware is placed in a fireproof, theft-proof and flood-proof building.
April 26, 2009
Causal inference and random trials
2 days ago