The Center for Data Innovation spoke with David Knight, CEO of Terbine, a startup building a digital marketplace for physical-world data. Knight discussed the challenges of working with the large amounts of disparate data sources in the Internet of Things, as well how businesses could benefit from accessing this data.
Josh New: Terbine is attempting to build a marketplace for data generated by the Internet of Things. Why do you think there is a need for this market, and who benefits?
David Knight: So far, the Internet of Things has been discussed mostly in terms of implementing sensors and basic intelligence in consumer, industrial, and governmental areas. Many software companies are starting to offer platforms and online services that help companies capture, store, and utilize their own data. There are also websites that allow the unfiltered uploading of sensor data that anyone can access, typically on an unpaid basis. What hadn’t appeared was a bona fide commercial grade marketplace that curates sensor data from a very large range of sources, considers the provenance of those sources, and enables the processing and sale of the data the way that big-scale data providers such as Bloomberg do today. Many entities would like to access physical world data, but to make it actionable, it needs to be offered in a form that’s codified and indexed, so we developed Terbine to answer this need.
The beneficiaries of these efforts are anyone who wants to buy or sell physical-world data. Terbine’s architecture allows sellers to be paid for their data based on a range of variables, and in many cases this will provide primary or secondary revenue streams for companies and government organizations that have invested in installing sensors. Entities who add value to data can also place it back into the marketplace for sale. On the buyer side, Terbine makes it easy for customers to access data readily, and through our metadata will get a thorough picture of the origins of that data, its timeliness and other important factors.
New: Could you explain Terbine’s Global Sensing Network and how it will manage the high volume of data from the Internet of Things?
Knight: The Terbine marketplace will intake data from a broad range of sources to include publicly available data generated by government agencies across the world, corporate entities who may initially resist the distribution of their data, but for whom Terbine intends to become the trusted curator, and eventually consumer-generated data, which can even include automobiles and other devices related to individual or group environments. This ecosystem of sources combines to create a Global Sensing Network. To handle the enormous volume of data that will be generated and has to be managed, Terbine has developed an architecture that produces metadata describing inbound sensor information so it can be readily and rapidly searched by customers’ software or human end users. From the start, the architecture was designed to be highly distributed to allow task processing to be done at the edges of the system as well as in the cloud. It’s worth noting that quite often the senders and recipients of data will be entirely machine-based. For example, field sensors capturing infrared information to determine oil pipeline flows might feed into an analytics package running on top of SAP. In fact, we anticipate that the majority of sensor data production and consumption will be done via automation and will not involve humans interacting through a web interface. This represents an evolution beyond how other marketplaces such as eBay and Amazon are typically accessed.
New: How will Terbine correlate data from so many types of inputs?
Knight: The metadata that is automatically associated with inbound sensor information is designed in a framework that can theoretically accommodate any type of data, from CO2 levels to vehicle movements, ultraviolet light, radar scans, ocean salinity, sound, and many more. We’ll collect data from all altitudes to include fixed sensors on the ground, trucks, ships, drones, and satellites in low earth orbit. It’s important to note that Terbine will not be deploying these sensors, but inputting from sensors deployed by any number of third parties, although we may elect to launch our own hyperspectral satellites to guarantee near-realtime global coverage. We will be applying distributed machine learning technology to facilitate the guarantee of data quality. Our mission is to be the overarching curator of the physical world’s data, whereas analysis and correlation will be done by our customers.
New:What are some of the regulatory challenges associated with packaging and selling all of this data?
Knight: Last summer with the help of some very bright, eager university interns, we took a look at where and how issues such as data security, privacy, and other salient matters that fall under regulatory umbrellas, are evolving. One thing that stood out immediately is that these issues are very much in flux worldwide, and there are marked variances in the approaches and tone of regulations in different regions. To avoid these becoming stumbling blocks, in particular those that have yet to be solved, we have ensured that our system is capable of dynamic rules-based filtering which can deal with not only legal and regulatory shifts that may be topical, but greatly facilitates tracking chains of provenance for data as different parties add value to a data stream, so we can provide revenue back through that stream.
New:You have an interesting background, from aerospace to virtual reality. How has this led you to working with the Internet of Things?
Knight: I feel like everything I’ve done in my life and career has led to the founding and building of Terbine. On a daily basis I draw on my training and experience in a variety of areas. The most interesting and useful (besides startup experience) is prior work I’ve done in wave theory and other physical world sciences. Plus, having been very involved with the birthing of the private spaceflight industry starting with the XPRIZE in 2004, I can say that I enjoy being a part of things that others say can’t be done. Terbine has the potential to become something large and lasting that can impact not only business, but also people’s everyday lives. I am certain that it will be incredibly challenging to build Terbine at such a huge scale, but when it achieves critical mass, we will be able to, quite literally, take the pulse of the Earth. How cool is that?