The concatenation of RFID and relational databases

Posted on May 21, 2008 by Karl Palmås

A note on Katherine Hayles‘ current work on panspectrocist technologies.

In chapter 6 of Super Crunchers, mentioned in the previous post, Ian Ayres charts the technological developments that have given rise to the era of “super crunching”. (He emphasises that it is a matter of technologies, not techniques; regression analysis is hardly novel mathematics.) Apart from the development of more powerful processing power (stated by Moore’s Law), he deems two technological developments crucial:

1) The pervasive digitalisation of everyday life, (making what we call universal modulation more effective)

2) Advances in database compatibility (allowing for data sets to be “mashed up”) and storage capacity (Kryder’s Law: “the storage capacity of hard drives has been doubling every two years”).

Ayres writes:

There’s no reason to think that the advances in database technology will not continue. Kryder’s Law shows no signs of ending. Mashups and merger techniques are becoming automated. Data-scraping programs of the future will not only search the web for new pieces of information [...]
And maybe most importantly, we should see continued advances in the digital domain’s ability to capture information - especially via miniturised sensors. (154)

Whereas Ayres is gleaning into a “not-too-distant future”, where “nanotechnology may spur an age of ‘ubiquitous surveillance’”, where we are surrounded by “smart dust“, Katherine Hayles believes that the currently used RFID technology will yield some substantial results. To her, the building blocks of ubiquitous surveillance are already out there.

In a recent issue of Theory, Culture & Society (vol 24, no 7–8), Hayles discusses her current work on RFID and relational database (RDB) technologies. Thus, while focusing on somewhat different technological actualisations than the ones pointed out by Ayres, she too has singled out the concatenation of universal modulation and database technology as key to the development of new forms of surveillance and control.

Hayles points out that the RFID technology holds one key promise/threat: Cheap and simple, it enables us to give unique identifiers to an infinite number of objects:

Most RFIDs are encoded with a 10-digit number hardcoded onto the chip; if you do the maths with the permutations, a 96-bit chip has enough permutations uniquely to identify every manufactured object on the planet, about 80,000 trillion (or 2 to the exponent 96). So that means that we can now give an individual identity to everything in the world that is built as opposed to being natural. [...]
RFID tags themselves are very simple. They are much simpler than any AI system that I know about. That’s their beauty. They can now be manufactured as cheaply as 2 cents per tag. Many passive tags are about the size of a grain of rice. Hitachi is working on tags that are much smaller. And because the brains are separate from the legs, these tags are cheap and pervasive, and can be embedded in everything. (352)

When Hayles talks about “brains separate from legs”, she is referring to the fact that RFID chips do not have to be particularly “smart” in order for an intelligent system to emerge. The “smarts” of the system emerges though the intermingling of RFID chips and RDBs. In this assemblage, there are three tiers;

you have the tags which are ubiquitous because they are cheap and can be everywhere. You have the readers, which are more expensive but not very expensive. And then you have the back-end relational databases, which are pretty expensive and will be in stationary computers because of their size and complexity.

So, the intelligence of this system emerges at the database/analysis level, where all the data sniffed by “stupid” RFID tags are scanned for patterns - as in the original Panspectron sketched by DeLanda. In other words:

If the relational databases are the brains of the system, RFID tags are the legs. (349)

The result is a form of man-made intelligence, but not in the traditional HAL sense. Whereas traditional means of building Artificial Intelligence (AI) were premised on the idea of “trying to build the equivalent of a person; [...] trying to build the whole cognitive system in one artefact” (353), the distributed RFID+RDB approach creates “low levels of cognition, but very significant cognition”.

What about the practical uses of these panspectric contraptions?

It can absolutely be used for surveillance. It’s already being used for surveillance. But at the same time, like any technology, amazing things have also emerged from these kinds of correlations. There is a medical researcher called Don Swanson and he’s using data mining and text mining techniques to hypothesize causes for very rare diseases that have never been able to be understood because they happen so infrequently and in widely separated locales, so that the aetiology of disease is simply untraceable. But now, with these mining techniques on large datasets, he’s able to make correlations that could never be made before and has successfully hypothesized the causes of several of these rare diseases. (350-351)

The panspectic diagram is not to be seen as neither inherently good nor inherently bad, we can simply join in the sentiment expressed by Deleuze: There is no need to fear or hope, but only to look for new weapons. There is, after all, a huge momentum in the shift towards this form of ordering. It is used for road pricing/congestion charges, it is spurring a revolution in factory production etc. Also, major purchasing agencies now require their vendors to use RFID technology:

Wal-Mart is requiring its major suppliers to tag every product they buy, and so is the US Department
of Defence. (354)

At some point, we may even see RFID tagging of products may becoming a required standard at an EU or US level. The examples of uses are limitless; many interested parties will see the merits of making tagging mandatory. For instance, Hayles talks about how it enables our fat Sunday papers to be customised, preventing consumers from having to throw out “three-quarters of the paper in the trash before they even read it”.

Or, as Ayres writes, electronic sensors “could keep track of how long you hold on to a particular product before using it, how far you transport it” and so on. As soon as goods have IDs, their origins can be precisely traced. Moreover, these goods can be linked to owners (who can be held responsible for recycling them). In the end, it might be eco-friendliness that causes us to make RFID traceability mandatory.

All content copyleft the authors under the Creative Commons 2.5 license • Powered by Wordpress • Using Detour theme