September 24, 2014


Internet of Things – Connecting Intelligently

Internet of Things (IoT), or in other words, connected things can range from sensor devices to washing machines, refrigerators, vehicles, mobile devices and shop-floor machines. A survey states that there will be more than 10 billion connected devices in around 7 years’ time.

This blog covers the various attributes of IoT and how they work together to create an eco-system.

The "things"

In IoT, it is all about "things". As mentioned before – connected things, which send information somewhere, or that can be controlled by external forces, or that can act as consumption channels.

The “controller”

Things need to be 'connected' in IoT. The connection could be enabled by the controller, in other words - the brain of the eco system. This blog is limited to just one particular network, although it is possible to create a network of networks in IoT parlance.

The controller has the following 'modules'.

  • Device detection
  • Device administration - registration, deregistration, etc.
  • Device runtime monitoring - online/offline
  • Communication     
    • to receive information from devices
    • to send commands/information/alerts/notifications to devices
  • Security – to transact securely
  • Data handling - data storage, manipulation, archival and collation
  • Integration - communicates with enterprise systems and external systems for analysis
  • Intelligence - scoring and decisioning

External systems

The concept of IoT is to make it well distributed so that the load is always shared. This is also applicable for processing information. External systems such as ERP, CRM , analytical engines, and scoring systems are important in the eco system.

The operation

Enterprise business systems adopt a different methodology when it comes to making themselves suitable for IoT. Traditional ways of accumulating data and processing at a later point in time, like batches, may not work well for IoT. IoT requires real time data capture and processing. Depending on the industry and the use cases, real time can be compromised for near real time.

There are two main types of data being collected by the controller - primary data coming from the core devices such as sensors, and supporting data coming from external sources, for example: weather information. Analytics can range from a simple rule such as a particular sensor exceeding the threshold limit, to complex events based on data from machines, weather, historical data, and transport vehicles. A scoring engine can also be involved to improve the accuracy of the decisioning based on what is inferred from the analytics operation.

The IoT landscape includes systems, tools and technologies that can handle large volumes of data transactions, and handle the information securely while processing, storing and streaming the data. Sensors range across sound, vibration, transportation, chemical, electric, magnetic, and more, and are used in multiple industries, but they mostly work in silos. When each such device is connected and the information is centrally processed overlaying other intelligence inputs such as data from other sensors, enterprise data stores, etc., the degree of business applications of IoT will be very high.

The outcome of the analysis can be used for prediction and for post-event activities.