Putting the IoT cart before the horse

The hype surrounding IoT platforms is reaching a crescendo, and at 500+ IoT platforms (by some estimates), we have a dizzying array of choice to contend with. Most of these platforms offer varying levels of functionality across the entire spectrum of the IoT value chain. Some operate in distinct niches (connectivity, edge, analytics, etc.), while the bigger ones seem to be targeting an all-encompassing proposition.

One thing remains common, though, across all these IoT platforms. The focus and fascination with getting the sensors to spew out their emotions (aka measurements) constantly and the ability to visualize that real-time live “in the moment” continues unabated. Customers too are driving this frenzy and the demonstrable use cases they are demanding focus on pulling streaming data feeds off the sensors, ingesting the same and visualizing the same real time with basic trending and exception analysis. Clearly there is a lot more to the IoT business value than real time-data acquisition and visualization that seems to be getting all the attention in the IoT revolution.

In our eagerness to get IoT tech adopted, the table stakes of IoT seems to be obfuscating the real business value equation. Are we really missing the forest for the trees? Is our obsession with connectivity and visualization causing us to miss out the big picture on how IoT will really generate business value? In short, the key question is “Are we putting the IoT cart before the horse?”

Gartner predicts that by 2020, 80% of all IoT projects would have failed at the implementation stage. The obsession with getting data across the finish line in the hope that it will shine a spotlight and create an IoT utopia seems clearly misplaced. Sensors have been there for a long while. No doubt they are getting cheaper, abundant and more capable owing to the integration, compute and bandwidth advancements, but the real power of the IoT platforms would be better tapped in delivering actionable insights on an ongoing basis which delivers the real business value. Clearly there is no one single silver bullet, but a multitude of factors in long road ahead toward the business value realization from IoT.

So, what is the ultimate holy grail of IoT-driven business value? What is it in the medley of data, connectivity, compute, algorithms, talent or process that would enable us to harness the true business potential of IoT? To answer that question, let’s explore each of these factors and the value each brings.

Data: Data has indeed proliferated exponentially, and so has the ability to capture that data meaningfully at a fraction of the cost. According to some estimates, the amount of data actually put through analysis prior to the large-scale advent of the IoT revolution was a mere 2%. Technology is definitely getting better at amassing and analyzing a higher percentage of this treasure trove of data, and definitely as data science would tell us the more the data, the better. Data may be the new oil, but the texture and chemistry of this oil is becoming ever more complex with varied unstructured data sources constituting the lion’s share of data processing. This new data oil will also need massive refining, blending and distribution (to the right stakeholders at the right time in the right context) to deliver the goods.

Compute: If data is the oil, compute is the power to keep the engine humming. Compute has seen rapid advancements and the mainstream adoption of GPU-based architecture and massively distributed parallel processing has ushered in a revolution in compute capacity. Compute has also enabled a level playing field through availability of cloud-based environments on tap which eliminate any potential hindrances and scale related challenges for even a startup with a brilliant idea.

Connectivity: Connectivity still remains a spaghetti of sorts, but constitutes the essential plumbing to keep the IoT engine humming. Multiple standards proliferate both at the radio and network level, and there is no one size fits all. We are however blessed to be in an era where connectivity is no longer a constraint though. Multiple competing standards and protocols may delay our journey towards standardization, but connectivity still is the umbilical cord which binds all the elements together.

Algorithms: Algorithms are purportedly the icing on the cake and the most hyped of the lot. Algorithms are the cool stuff, the ultimate nirvana in IoT land, but algorithms rest on a solid foundation of data and compute and draw their power from the same. Algorithms also need to deal with multiple aspects like precision, recall, false positive/negatives, and despite the hype surrounding them have multiple challenges to contend with. Developing algorithms in an isolated data science classroom-based environment is one thing, deploying them in a production environment and using them in context to derive and then apply insights is yet another. Context also changes too fast and algorithms need to keep pace through continuous refreshes to stay ever relevant.

Talent: Like in any field, having the right talent is the absolute bedrock for true success in the IoT world too. Easier said than done, though, since we need a multitude of very diverse skills to successfully deploy IoT in a production landscape and deliver the goods. IoT needs a cross-disciplinary approach and an amalgamation of engineering, domain, IT, OT, data science skills and a really strong cross-domain program management perspective to integrate the multiple disciplines. Good talent (especially the cross-disciplinary ones to help integrate diverse perspectives) remains a challenge at least in the foreseeable future.

Process: IoT still remains a relatively new field, and the processes and methodologies for production-level deployments at scale are still evolving. There are arguments both in favor of and against adoption of agile-based methodologies for IoT deployments which have proven so successful in the IT realm. The IoT methodologies and best practices development remains a complex endeavor with multiple moving parts and entirely different legacies and philosophies on the IT and OT camps.

IoT platforms thus need to broaden their horizons and evolve into richer IoT ecosystems which can orchestrate these diverse aspects aptly and provide integrated value-driven technologies. For the real potential of IoT to be unleashed and for it to realize the vision of trillions of dollars of business impact in the coming decades, it will clearly take a “systems thinking” approach to integrate all the above perspectives into an integrated whole. That is how we will get the IoT horse before the cart!

All IoT Agenda network contributors are responsible for the content and accuracy of their posts. Opinions are of the writers and do not necessarily convey the thoughts of IoT Agenda.

Data Center
Data Management