The news announcement, that personal data of more than 25 million British people has been compromised, was greeted with the inevitable tut-tutting. “Mistakes by junior officials” meant that two disks, containing copies of the data, went missing when courriered between two Government offices….
But, hold on a minute.
That such a practice can happen in the first place is surely the mistake, not the cack-handed blunder by some poor sod, who is now going to take the rap.
The core problem is that electronic data is still treated all too often in ways familiar to any 19th Century bureaucrat: data and information is stored on a medium – for centuries, on paper – and then passed around as needed, including being mislaid, misdirected, lost, misfiled or intercepted. Nothing seems to have changed, except that the paper has been replaced by digital media.
When will eGovernment join the 21st Century? Data – whether its personal data, company data, or whatever – are not just streams of bytes that should be able to be copied and moved around at whim. They are valued and valuable assets. The whole chain of errors reflect 19th Century processes (copy a file, stick it in an envelope and send it through the mail – I mean, really!) not the “joined up Government” that the the UK’s eGovernment strategy sings about.
Data should be accessible and useable as a service, not as a passive object in someone’s programming code. Computing and information architecture paradigms today allow for much more robust and suitable models for data management in the 21st Century. “Data as a Service” means that access to it is governed by policies, transactional controls, real-time authentication and access control, etc. In the latest fiasco, such an approach if applied would have meant:
- a junion official, or whomsoever, would not be able to simple copy data in this manner. The data would not be “sitting around” anywhere in a way that can be simply scopped up and copied;
- if there were a need for a copy (and a backup would seem to be the only legitimate need), this would be managed and executed according to clear policy and end-to-end management of data access: once a data set is copied onto a CD – as seems to have happened in this case – the data resides on a non-managed and non-manageable medium: it simply should not be possible;
- there simple shouldn’t be a need for a copy: if another service needs access to certain data sets, let them come, be authenticated and granted transactional access as needed to the original data, as one would do with any half-decent SOA-based system. And what about referential integrity? Has that good old-fashioned database principle been forgotten?
There is a more fundamental issue, and which is still the subject of considereble policy debate: who actually “owns” the personal data in question? Ceratinly the state exercises (a sometimes monopolistic) custodianship in many countries but very few would go so far as to claim that personal data actually “belongs” to a public authority. Unfortunately, state immunity from prosecution means that an individual could not take legal action against abusive access to personal data, in the same manner that one might if a financial institution started messing around with your money. But this is no excuse for inaction. The increased importance and relevance of personal data – and other valued information assets – means that this issue will only become more and more important.
See the discussion paper on “Personal Data Services Model” that looks at such an approach