On RFID - part 5 (final)

(and last, but not least... In my opinion the best of all the talks I attended at the conference)

Closing session

The speaker of the closing session was Sanjay Sarma from MIT.

He started by presenting a RFID system's functional stack of today, composed of the following layers: tags, readers, middleware, real-time business processes and ERP (Enterprise Resources Planning). Each company has such a system. When it wants to cooperate with a partner company, it uses EPCIS.

Next he discussed what to do when preparing for tomorrow. There are several topics: security, different types of tags, different data acquisition means, new data types, new real-time business processes, new database requirements, new business processes and new exchange mechanisms.

Low-cost RFID is a constant struggle where you can choose, at best, two of the following: cost, range, functionality.

Regarding security, we have requirements for tag authentication, reader authentication and protection from eavesdropping. The latter two require encryption. Cheap and secure tags are, at least, 5 years away, because we need a new class of technologies, like: digital fingerprinting of integrated circuit on tag, read/write fingerprint at manufacturing and verification on demand.

Sensors need power, so new research is looking for power scavenged from vibrations.

Actuators will also be possible, i.e. tags that change state (1 bit) on their code. This can be a cheap communication medium for remote control.

Beyond passive RFID we'll takes to mapping the space with location/identification technologies like: wifi, cellular, wimax - all will live together. We'll have inexpensive radars.
Reading will expand soon, as Application Specific Readers (ASR) emerge, that are cheaper than today's generic-purpose readers, because they take out the stuff you don't need.
Also interesting is the possibility separating the power and signal from the reading process. For instance, a light bulb could power tags and a centralized reader would read the signals.

Concerning health issues, all studies so far have not shown any influence of RFID radiation on pharmaceuticals and other sensitive products.

The RFID system functional stack will have to extended for real-time business processes.
The middleware layer will deal with device management and will be able to detect events: "what", "when" and "where" (literal).
The business process layer the "where" will have to be enriched with additional context, to be meaningful to business and the "why" will also be important, to trigger actions.
The ERP layer will need a new data model, one that takes into account that mistakes are a part of RFID data. Databases will need to be self-healing i.e. correct themselves using basic models of behavior and of the physical world (e.g. "one missing read doesn't mean much in a warehouse stacked with products..."). Databases will also have to act more like streams and less like batches to handle the voluminous data.

RFID will enable new business processes but it's important that the "tail not wag the dog" meaning that business needs should drive RFID and not the other way around...

RFID will enable dynamic, real-time, reactive systems with the ability of just-in-time focus. Discovery will have to accept high-level commands like "Where has this object been?" and "Who damaged it?" with lots of inter company logic. This ability to "touch the physical world" will allow for the invention of new business processes.

In his final remark, Mr. Sarma recommended that "don't paint yourself into a corner" meaning don't limit your future options. As an analogy, he said that we compare the Internet of 2005 with the Internet of today!


Closing speech of the Portuguese Presidency

Bráz Costa from IAPMEI in representation of the Portuguese Minister of Economy and Innovation, talked about the European paradox: good research but bad economic valuation of knowledge. The current Portuguese strategy follows four axis: create knowledge and companies, qualify to manage, grow by innovation and export.

Carlos Zorrinho from the Portuguese Government technological plan urged Portuguese companies to "make excellence a normal state" using their knowledge, technology and innovation, sparked by creativity, with the goal of economic valuation through problem identification and solution building. The goal is a exports-driven economy.

Afterword

This concludes my report on the “On RFID – The next step to the Internet of Things” conference and exhibition that was held at Lisbon, Portugal on the 15^th and 16^th of November 2007.
I would like to publicly thank the organization for giving me the opportunity to participate in the conference. I would also like to congratulate them on a job well-done!