I have previously pointed out how supply chains for data are an increasingly important aspect of ITS and C-ITS delivery. They are also important for autonomous vehicles because the onboard systems require access to up-to-date mapping and other data, even though this access may not need to be in real time.There are several possible architectures for data collection and management.
The most obvious is for some form of large, central repository – such as a data warehouse – that collects data from a variety of sources. It then manages the storage, curation and perhaps validation of the data and reformats it so that a number of output channels are available for data consumers.
The Internet of Things (IoT) approach is different. Streams of data from devices or primary data collection systems are available through an internet platform and consumers can select the set of streams they need for their purpose. In this case, the consumers may add value to the data by combining disparate data streams, or performing analytical tasks including estimating future data. They may then make their output available as an IoT stream that can be accessed by other consumers or provide a service direct to a specific user base.
While different, both approaches have common issues regarding ownership of data and, more pertinently, responsibility for the data availability, accuracy and validity. There is also the question of metadata. To be able to consume a data stream usefully, you need to understand the nature of that stream. There is a whole host of fixed, or slowly changing, information needed to interpret a dynamic data source. This metadata needs to be owned and maintained, and responsibility for its veracity clearly identified.
In a data warehouse the operator of the facility, which may be a physical entity or a cloud-based arrangement, takes on much of the responsibility. It can be selective about sources of data, provide verification and data cleansing, and perform data curation and formatting. The IoT is more diffuse and open ended. Any owner of data can make it available as a source stream. However, this places consumers of data streams in a caveat emptor situation, having to work out if the metadata is adequate to allow the data stream to be understood, and they are reliant on the supplier for accuracy, reliability and availability of both.
One of the key opportunities is claimed for IoT is the ability to bring together disparate data to create new streams of information that provide value to new, untapped customer bases. Those creating this new value may not be experts in every data stream they access and so are reliant on that stream being fit for their purpose. At present this is a considerable risk, but it does create an opportunity for a brokerage service to act as data manager. Currently reaching a specified data quality seems to be a problem for public sector bodies and the private sector is still reluctant to share data by itself. There are technical challenges in IoT, but they are soluble if there is a business, organizational and contractual infrastructure to match. This should become the immediate focus of government support and needs the involvement of risk capital players before the technology heads off in a different, more commercially attractive, direction.