Accelerating the
Industrial Internet of Things

Internet of Things (IoT) and IoT Analytics

Published on 11/18/2016 | Technology

348 1

Hemang Dave

I bring an unique blend of technical and business acumen. Over last twenty years in the IT industry, I have held various highly visible technical and leadership positions at various companies including IBM. I have worked with large multinational Strategic Outsourcing customers and have provided guidance on their IT strategy and road-map. I have managed team of architects performing over $1B of solution work. I am a strong leader with very deep technical skill-set.



The Internet of Things (IoT), also referred as Internet of Everything, where objects are provided with unique identifiers and has the ability to generate and transmit data over the network without human-to-human or human-to-computer interaction. I have purposefully titled this blog IoT and Analytics. I strongly believe that in order to be successful at IoT, besides capturing data through IoT, we must drive actionable insights against the data we have gathered. Our Chairman Ginni Rometty has repeatedly stated that successful companies of 21st century will need to be successful at mining all sorts of data. These actionable insights are the key to next generation of business growth. Humanity will see technology transform in a dramatic way by employing IoT Analytics. We are just scratching the surface of such things right now. Let me give you an example: 2015 IoT Analytics market is estimated at $4.8 Billion. However, by 2020, it will grow to a $17B business. The purpose of this blog is to introduce you to concept of IoT and IoT Analytics. It is not my intention to do a technology deep-dive in this blog writing.



Over past 25 years, technology breakthroughs have allowed IoT and IoT Analytics to be formed. Wireless technology, for example: RFID (Radio Frequency Identifier) and others also have played crucial role. In past 8-10 years, wireless technology has become more reliable, but more importantly, cheaper to deploy.

Since 1990s, back-end Database technology has evolved to a great length. Most notably, introduction of Big Data technology. Big Data allows us to store both structured (traditional databases containing records) and unstructured data (pictures, free form text and movie footage etc.) Graphic on the left shows time-frame of evolution.

Increased deployment of connected devices, sensor analytics, predictive analytics, machine to machine communication, other IoT analytics platforms and user friendly applications (both traditional and mobile) will fuel growth of IoT and IoT Analytics. As you are already aware, we at IBM play in all areas - Software development (SWG), Consulting (SWG, S&D, GBS, GTS) , Solution (GBS, GTS, SWG)  and Services (GTS). In GTS, we will play a critical role for Consulting, Solution and Services. 

Let us now take a look at how a typical IoT Analytics deployment topology would look like. Note that this is a high level concept picture and deployment of IoT Analytics will be dependent on use cases, client needs, technology deployment and various other factors.

IoT and Sensors deployed in the field

This is the place where data of IoT and sensor originates. This data can come from various many sources: Connected cars, Sensors embedded in devices, Wi-Fi generated data etc. The purpose here is to create sensory data which can be used in variety of ways. When deploying such technology, one has to be very careful on what needs to be measured or captured. Today's connected cars, for example, can create more than million data points every minute via senors in the car and can store into On-Board Computer (OBC) system. Security during data creation and during transmission to other areas of IoT and Iot Analytics is of utmost importance. In recent months we have seen hackers hacking into connected cars and have taken over control of vehicles remotely. Security breaches can and do occur with sensory data.

Gateway Analytics

Once data is generated, Gateways first determine quality and integrity of data being generated. This is of importance as we need to ensure that we are not passing erroneous data down-stream. More importantly, this is very useful when real-time or near real-time analytics is required. For example, analytics can't wait more than few minutes to be acted upon. Additionally, this type of data becomes stale very fast as well. In other words, data reaches its end of life in matter of minutes and no longer useful for further analytics and insight. At this point, one can choose to discard data or can forward to either public cloud based data store for further retention or in rare cases if additional analytics is needed. Just like previous stage, security of stored and analyzed data in public cloud is vital. Most companies choose to store data which can be of non-confidential in nature. Example of this data can be weather data or any type of public domain information. Companies choose this option as public cloud storage and public cloud based Analytics can be cost effective and can be consumed as - Software as a Service. We are in process of deploying public cloud based analytics services which are based on our Softlayer cloud along with Splunk, Sumologic and our own set of Analytics tools such as Big / Predictive Insights.

Corporate Network

Inside the corporate network, one can deploy a very high degree of customized and sophisticated set of solution for company's needs. Any combination of the following can be sent to corporate network for sophisticated IoT Analytics: Data generated through sensors deployed in the field, Gateway Analytics Data, public cloud store data and Public cloud based Analytics generated data. This data comes through corporate firewall using various protocols and firewall ports. Data is stored in a mechanism called Data Lake, also referred to as Data Reservoir. Data Lake is based on Big Data technology. Both structured and unstructured data can be stored in Data Lake. Various sets of Analytics tools / engines can go through this data and produce actionable insights. Again, this can be real-time, near real-time or any frequency based analytics depending on business needs.

By looking at these various components of IoT and IoT Analytics, you can summarize that today's businesses require very complex and sophisticated sets of technology to make IoT happen. Lot of start-up companies are working towards creating next generation of technology to support IoT and IoT Analytics. 

Please provide your thoughts, comments and feedback below. Please note that the views and opinions provided here are mine and not that of IBM.

This article was originally posted on LinkedIn.

Feature New Record