I was once again honored to be a guest on Coffee Break With Game Changers Radio today with David Jonker and Ira Berk of SAP — it’s always a delight to have a dialogue on the Internet of Things with these two brainy guys (and hats off as well to moderator/host Bonnie Graham!).

Toward the end of the show, Ira brought up a concept that was new to me: virtual sensor networks.

I’ve got sensors on the brain right now, because I’m frankly worried that sensors that don’t have adequate baked-in security and privacy protections and which can’t be ungraded as new opportunities and threats present themselves may be a threat to the IoT because they typically remain in use for so many years. Ah, but that’s a topic for another post.

According to Wikipedia, Virtual sensor networks are an:

“… emerging form of collaborative wireless sensor networks. In contrast to early wireless sensor networks that were dedicated to a specific application (e.g., target tracking), VSNs enable multi-purpose, collaborative, and resource efficient WSNs. The key idea difference of VSNs is the collaboration and resource sharing….
“… A VSN can be formed by providing logical connectivity among collaborative sensors. Nodes can be grouped into different VSNs based on the phenomenon they track (e.g., rock slides vs. animal crossing) or the task they perform. VSNs are expected to provide the protocol support for formation, usage, adaptation, and maintenance of subset of sensors collaborating on a specific task(s). Even the nodes that do not sense the particular event/phenomenon could be part of a VSN as far as they are willing to allow sensing nodes to communicate through them. Thus, VSNs make use of intermediate nodes, networks, or other VSNs to efficiently deliver messages across members of a VSN.”

Makes sense to me: collaboration is a critical basic component of the human aspect of the IoT (one of my IoT “Essential Truths), so why shouldn’t that extend to the mechanics as well?). If you have a variety of sensors already deployed in a given area, why should you have to deploy a whole new set of single-purpose ones to monitor a different condition if data could be synthesized from the existing sensors to effectively yield the same needed information?

A 2008 article on the concept said the virtual sensor networks are particularly relevant to three categories where data is* needed:

“Firstly, VSNs are useful in geographically overlapped applications, e.g., monitoring rockslides and animal crossing within a mountainous terrain. Different types of devices that detect these phenomena can relay each other for data transfer without having to deploy separate networks (Fig. 1). Secondly, VSNs are useful in logically separating multipurpose sensor networks, e.g., smart neighborhood systems with multifunctional sensor nodes. Thirdly, VSNs can be used to enhance efficiency of systems that track dynamic phenomena such as subsurface chemical plumes that migrate, split, or merge. Such networks may involve dynamically varying subsets of sensors.”

That article went on to propose a flexible, self-organizing “cluster-tree” approach to create the VSN, using tracking of a pollution plume as an example:

“…  a subset of nodes organizes themselves to form a VSN to track a specific plume. Whenever a node detects a relevant event for the first time it sends a message towards the root of the cluster tree indicating that it is aware of the phenomenon and wants to collaborate with similar nodes. The node may join an existing VSN or makes it possible for other nodes that wish to form a VSN, to find it. Use of a cluster tree or a similar structure guarantees that two or more nodes observing the same phenomenon will discover each other. Simulation based results show that our approach is more efficient and reliable than Rumor Routing and is able to combine all the nodes that collaborate on a specific task into a VSN.”

I suspect the virtual sensor network concept will become particularly widespread as part of “smart city” deployments: cash-strapped municipalities will want to get as much bang for the buck possible from already-deployed sensors, without having to install new ones. Bet my friends in Spain at Libellium will be in the forefront of this movement!

Thanks, Ira!

*BTW: if any members of the Grammar Police are lurking out there (I’m a retired lt. colonel of the Mass. State Grammar Police myself), you may take umbrage at “data is.”  Strictly speaking, the proper usage in the past has been “data are,” but the alternative is becoming so widespread that it’s becoming acceptable usage. So sue me…