30/01/2007 By Dirk 0

(Participatory) Wireless Sensing in GENI

The US is about to ramp up a large scale effort to provide a test environment for Future Internet research. The initiative is called GENI (Global Environment for Network Innovation) and its budget is immense (some hundreds of millions).

The scope for GENI encompasses fixed Internet, wireless Internet and wireless sensing from the hardcore physical layer space over routing to service environments.

But keep in mind, GENI only intends to provide the test environment; it does not perform the actual research to be tested. So you can think of GENI as taking Planetlab and blowing it massively up in terms of scale (so no surprise that Larry Peterson from Princeton University has a central role in GENI).

Within GENI, there is a task force to define the wireless environment together with the wireless sensing part. It is good to see that the current direction of the system specifications envisions the notion of GENI-compliant information gateways towards non-GENI sensor networks. This largely targets scenarios for participatory sensing, apart from rather static gateway scenarios where local deployments are remotely connected to some backend infrastructure.

What does participatory sensing stand for? Take the idea of end users participating in a GENI-like experiments with their own mobile devices (such as their phones), connecting permanently or temporarily (e.g., when passing by) to sensor networks that are deployed in the field, delivering particular information back to other users or dedicated servers in the infrastructure, e.g., providing large scale services for pollution or traffic monitoring.

I’ve been looking into such types of scenarios with the work around NORS, developing a Java-based platform for mobile phones that allows for exactly these types of scenarios. For instance, in a project with Cambridge University (UK), pollution monitoring is one of the objectives in such participatory sensing, involving the community members directly. Another interesting activity in this space is the work of Matt Welsh in CitySense, in which some hundred light poles in and around Harvard University are to be equipped with such information gateways, potentially complemented with mobile devices carried by students. Also Deborah Estrin’s work on multiscaled actuated sensing is very well aligned with these ideas.

As a member of the GENI wireless working group, I’ve been working to marry the concepts of NORS with the specifications for this future networking test environment. Many of the concepts of NORS do map quite well onto the thoughts and work in GENI. But more is needed, keeping the research going in this field of mobile-centric wireless sensing (probably fueling future releases of NORS as well).

All in all, it is an exciting area and shows the increasing possibilities and opportunities of end users directly participating in making services happening by contributing towards a common goal. Creating incentives for end users to participate in such services will be one of the more non-technical key questions to answer in the future. But this is a topic for another entry…

So where is Europe here and its collection of test beds and initiatives, one might ask? There are indeed many European initiatives for test beds in different areas. The ambition to build a GENI type of facility (in particular with the Global ambition for such environment) does not really exist. But there is ample research in Europe around the same problems and any activity, be it GENI or some local test bed facility, will benefit from this research. The EIFFEL initiative is currently trying to create some upwind for more explorative types of research projects. And there is also increasing interest in the European community to extend many of the sensor network deployments towards participatory scenarios (I attended an invited event by the European Commission last year on this topic). So there is momentum, not only in the US, but worldwide when it comes to participatory sensing.