In the age of IoT, think “Data Centric”

Status:
Accepted

There are two primary approaches to sharing data in the IoT: message-centric and data-centric. The message-centric approach (made popular by MQTT) has become the go to approach when architecting, designing and implementing IoT systems. We believe that there are systems for which a data-centric approach is much more suitable and should be considered.

Unlike a message-centric approach, in a data-centric approach each physical entity is represented as a data object that has:

  1. An identity
  2. A structure
  3. A state
  4. A lifecycle, and
  5. Metadata that characterises it and captures the above as well as provides content-awareness for database like filtering and querying.

Furthermore, there are some data-centric solutions that allow for attaching data qualities to the data. Examples of data qualities are: 

  1. Security
  2. Persistency
  3. Consistency
  4. Reliability
  5. Urgency; and
  6. Importance.

We will look at the primary differences between a message-centric and data-centric approach to sharing data in an IoT system. As part of this the systems or scenarios were a data-centric approach should be considered are explained to show the benefits to using data-centricity. Finally, we conclude by looking at some open standards for a data-centric approach to data sharing. For example, the Data Distribution Service (DDS) from the Object Management Group (OMG) is one such standard.

Session details
Speaker(s): Session Type: Experience level:
Intermediate
Track: Tags:
iot
DDS
m2m