The true power of analytics-driven insights is only realized when the practice of experimentation and learning becomes so deeply ingrained in the organizational culture that it starts guiding every step of the company’s strategy and operations.
This is an article written by Ashish Gupta, Head of IoT, Jigsaw Academy. Ashish helps organizations become data smart by ensuring the appropriate training programs are designed and delivered, tailored to the needs of professionals working in different job functions:
1. Internal Technical teams (IT | BI | Data Science)
• Upskilling people from lateral roles or augmenting existing data skills with the latest analytics tools and techniques
• Technical (basic to advanced) – R, Python, Machine Learning, Big Data, AI, Deep learning.
2. Data Enabled roles
• Data Science and analytics learning for mid-management w/o requiring them to write code, helping them become data savvy
• Storytelling with Data, Data visualization with Excel/Tableau/Knime, domain-specific applications like Finance, HR, Marketing, Retail, Supply chain & Manufacturing.
3. Senior leadership
Updating business leaders about how analytics is being used in their industry, how are their competitors leveraging data and how can they make their organizations ‘Data Smart’ by upskilling people and using data enabled techniques for decision making in more and more departments.
As we know, IOT is not a single piece of new technology but rather a confluence of various complementing technologies which have come together to enable new use cases which are helping businesses in various areas like production optimization, predictive maintenance, asset monitoring and management, remote connected operations and new products and services.
In fact, according to a recent report published by IOTAnalytics.com, there are at least 40+ new and emerging technologies which will play an important role in various IOT use cases.
Hence, while planning and architecting an IOT solution, there are numerous choices to be made. Choices at the hardware level – choosing different sensors, node device platforms, gateway device platforms, network infrastructure etc. Then there are choices at the device software level – Which RTOS to use? Which gateway software platform to use? Should we do it in-house based on some open platform? Or use an off-the-shelf proprietary solution instead? There are a plethora of choices at the networking layer – Should we use a cellular network like NB-IOT or Cat-M or plain simple 2G sim connections? Or is it wiser to go for a LPWAN technology like LoRAWAN or Sigfox? And if we’re going for a LPWAN technology, should we deploy our own private network or use a public network provider? It doesn’t end there, similar choices confront us while planning for the cloud deployments – How much data we need on-premise v/s hosted on the cloud? If it’s cloud, should we deploy our own cloud software based on an open platform using IaaS from one of the cloud providers or should we instead go for a PaaS or SaaS system? If it’s the latter, which PaaS or SaaS system should be used! There are tons of options and competing standards.
Often in life, the main source of confusion is too many choices. The more choices we have at hand, the more decisions we have to make, causing a lot of anxiety and confusion. And in the case of IOT, all the various business use cases, choice of hardware options, software options, choice of networking technologies and the choice of cloud platform options lead to complexity, confusion and bad decisions in many cases.
To simplify this let’s try to find out the common thread that binds all these components and choices together. What is the core element that stitches the various patches of IOT together? After all, hardware, software, networks and servers are not new. “Data is the biggest change”. Now, we’re not only looking at the data and acting on it, but also predicting future events based on data analysis. And if we look at it closely, an IOT solution can be described as the journey of data, from acquisition by a sensor to being communicated by low power node to a bigger computer or gateway using a communication link, to being aggregated and curated by the gateway and being relayed to the cloud, to being stored safely, reliably and optimally on the cloud system, and finally to being processed and converted into KPIs and business insights using various analytics algorithms.
Data acquisition by node -> data communication to the gateway -> data aggregation and transmission to the IOT network -> data ingestion, storage and retrieval at the cloud layer -> Data preparation and Analysis -> KPIs and Business insights
Hence, next time while planning and architecting an IOT system, I suggest we figure out the details of the data first. All other details will follow through naturally from there. We must first ask ourselves the core questions about the data:
As we start answering these questions, we will start developing a much clearer list of requirements for each of the other pieces of the big IOT puzzle. Answers to these questions will lead to a clear choice of hardware sensors to be used, node device capabilities, network technology to deploy, network architecture to be adopted, cloud computing architecture to be implemented and a sense of data processing and analytics work to be done.
In fact, the planning and designing process for an IOT solution could be viewed as the exact reverse of the journey of data through the IOT system.
Business Goal -> KPIs -> Raw Data -> Cloud System -> Network -> Gateways -> Communication Link-> Nodes -> Sensors