By now, you know IoT sits at the intersection of the digital and the physical, dramatically extending the reach of information technology. As the natural evolution of automation, IoT presents insurmountable opportunity (up to $11T/year of economic impact according to McKinsey: http://bit.ly/213q3VE). Given the hype over the past few years and the fact IoT transcends virtually all industries, one would expect to see mass deployments by now? Yet, not so much. IoT is off to a slow start and there’s a reason - it’s hard. So hard, Gartner predicts through 2018 that 75% of IoT projects will take up to twice as long as planned (http://gtnr.it/1OsQzzR).
The reality is, if it were easy, everyone would be doing it. So, why so hard you ask? Consider this. Assume for a moment your company has committed to embracing IoT, the strategy and business cases are in place, critical process / resource / organizational changes are identified, an executive sponsor is established, a successful throw away PoC has been demonstrated and funding allocated. Now all you have to do is develop the end-to-end IoT system.
The good news is you’ve done your homework, speaking with industry analysts and engaging consulting firms to help you evaluate and select the portfolio of technologies, which at a macro-level include: sensors, modules, antennas, housings, power, gateways, protocols, network, platforms, applications, data bases, UI’s, analytics, security, cloud and enterprise integration. Even if you’re buying an off the shelf solution (limited options exist) customization / integration is critical. For the majority, IoT is not a point solution, meaning you’ll need to work with many organizations and vendors to stich together a coherent end-to-end system. This begets the role of systems integrator, which could be done internally, outsourced to a third party or a hybrid arrangement.
Right Tools, Wrong Hands
With the Spreadsheets, PowerPoint’s and Reports behind you, now for the bad news - the software that is necessary to reach from the sensors to the enterprise is the integration glue and requires deep connected systems experience (skill sets intimate with embedded systems, connectivity, and real time, time series based systems) to develop industrial grade, end-to-end systems. These capabilities are not part of your traditional corporate IT DNA, so if your assumption was you were going to license an IoT platform, provide your enterprises developers a little training and access to the cloud, think again. The challenge is further exacerbated given the lack of industry standards, a battle the titans will surely drag out (http://bit.ly/1Ktk4oo), creating confusion, complexity, risk and lock-in scenarios. From a production ready perspective, I submit to you, one of the significant drivers behind IoT’s delayed traction is the supply-demand imbalance for these constrained connected systems competencies.
History Can Teach
We’ve seen this movie before – for those of you old enough, you may recall the tectonic shift from centralized computing to distributed computing in the late 80’s, early 90’s. The parallels are remarkable, with the same degree of hype, buoyed by new technologies, disruptive solutions, compelling economics and Wall Street darlings. However, the implementation of the systems was not trivial, often requiring tight integration with the corporate mainframes. It was well understood that those who grew up coding in Cobol or Basic and MVS struggled with the transition to Unix and C. The demand for developers with distributed computing chops was only aggravated by the competing technology flavors (Unix Wars: http://bit.ly/1mKJrqQ) and quickly outstripped supply. Ever the opportunists, the System Integrators and VAR’s raced to fill the void, offering one neck to choke, and oh, by the way, teaching their staff the new tools on the client’s nickel. By the mid 90’s, client-server was the dominant model and the dearth of qualified resources was ameliorated, although it took decades before Linux emerged as the de facto standard.
Expect IoT to follow a similar path. With systems integration paramount to bridging the digital to physical divide, and no one company providing an end-to-end offering, many companies will be forced to turn to more agile SI’s and VAR’s who know how to adapt and capitalize on the next big thing (client-server, right sizing / outsourcing, BPR, ERP, web services…). For more context, here’s a 1996 article from CIO Magazine titled ‘Gartner View: Client/Server Payoff‘ that I proffer, foreshadows a corresponding IoT version come 2020 (http://bit.ly/1Kt05Xy).
To be clear, by no means am I suggesting companies should wait to deploy IoT solutions. However, I am preaching caution to the promise of plug and play and drag and drop. If it were that easy, everyone would be doing it. Rather, my counsel, given IoT solutions will often live for decades, is to ensure your system is designed for change and longevity. To that end, consider the following as you roll up your sleeves:
- The IoT Stack – understand the tradeoffs between pure play, Full Stacks versus best of breed, Custom Stacks: http://bit.ly/1oruRpt\
- Protocols – given the paucity of standards and that one-size certainly does not fit all, beware of proprietary protocols with lock-in scenarios, particularly for things in the field. Seek to control your connection by owning your protocol IP.
- Connected Systems Experience – if you’ve got it, great, if not, get it and realize not all experience is created equal. Relatively manual applications with stationary devices that are powered and networked via wire is very different than resource intensive, edge based analytics on ‘things’ that are battery or solar powered and wireless (cellular, LoRA, Wi-Fi…).
- Deployment Models – over time we can expect to see a more intelligent edge so ensure your software choices can support all deployment models (in-cloud, on-premise and on-thing).
- Systems Integrators – SI’s are only as good as the people on your project and true connected systems experience remains very thin.
- Security – design and build in up front. Bolting on after the fact is a flawed strategy. Variables include hardware, network, software and people.
- Dynamic Data Model – uses cases, device models, and external integrations will evolve throughout the lifespan of the system, therefore a dynamic data model is crucial.
- Divorce – these systems can live for decades, so if for some reason you have to change out some of the ‘pieces’ (hardware, software, network) within the end-to-end system, understand your options and points of lock-in.
The promise of IoT remains unchanged, from subtle efficiencies, to disrupting industries, to global economic impact, but it remains early. The technology and know how will improve and every company will need to determine how and when to engage. For those who want to lead, let your mantra be “design for change and longevity.”
Close stream of consciousness. Hope it helps.
Originally published on LinkedIn