How Will The Internet Of Things Affect The Telecommunications Industry?

19/04/2017 15:00

How will the Internet of Things affect the telecommunications industry? originally appeared on Quora - the place to gain and share knowledge, empowering people to learn from others and better understand the world.

Answer by Shane Greenstein, Professor in the Technology Operations and Management unit at the Harvard Business School:

You are asking a general question, and it is tough to give a general answer. Apologies for the length of this answer. You have to understand what we can do today to understand what we would like to do tomorrow.

First of all, let's consider up to 2020. Even in this short span we should expect to see new applications placing pressure on scale, reliability, and security of the applications. That will press telecommunications operators and equipment providers to the edges of what they can do.

Look, the vast majority of IoT applications (so far) involve industrial applications with low bandwidth, light sensors, widely dispersed networks, and minimal reliability. For example, it is very common for a logistics company to record the progress of a package along a path. How do they do that? With inexpensive readers that send simple signals, placed along the path, automatically recorded by servers at headquarters, alert to simple warning signs of danger, which an operator can investigate. And there are gazillions of these sensors, so if one fails, another can take its place. This has become common in shipping, trucking, rail, and warehousing, not to mention oil and gas networks, and inside of retail firms with large logistical operations, such as Walmart, Amazon, and P&G. These type of applications are capable of large scale deployments, involving more sensors dispersed over large geographic distances.

That also tells you where the challenges will lie, and how it will impact the telecom industry. Acquiring and managing a rather large and geographically dispersed IP network is not beyond human capability today, as long as it contains a few fail-safes if something breaks. But...

1) Reliability. While 100% reliability beats 99%, and 99% is good most of the time, 99% is frightening in many applications. Sure, it is easy to design a pacemaker or blood pressure monitor with modern wireless IoT wireless capabilities, but nobody thinks 99% is acceptable in every location and under every circumstance. This hold back many valuable applications unless (or until) firms find ways to address it.

2) Security. While lax security is not a problem for boxes of razor blades shipped across the country, that is not acceptable for something like an autonomous vehicle. If you don't think so, I can show you scores of rogue programmers just waiting to hold your automobile for ransom. Again, this will hold back many valuable applications.

3) Privacy protection. While privacy rules pose few issues for the temperature gauges on an oil pipeline, not so for devices that read activity at home. For example, just recently a vibrator company got in trouble for sending a little *too much* data back to the company that designed the product. It is funny, but symptomatic of a much bigger issue.

4) System design. Real-time integration of massive data is already feasible, but those types of applications have many potential points of failure. After all, the integration of Waze with Google Maps already integrates huge numbers of low-bandwidth signals into an application that many can access. It may be irritating when it is inaccurate, but you can survive without it if the network goes down. Ah ... but there lies the rub. Nobody is suggesting we integrate airplane traffic with digital weather data, and have the computers do all the routing and land the aircraft. We still need humans because the systems are flawed. To be sure, there are some things we can improve.

Just to be clear, each of these frontier issues -- in reliability, privacy, security, and systems -- requires great engineering in the networks, the devices, and in the software. However, each of these requires more than mere engineering solutions. Law and management have key roles to play. Many frontier applications require some cooperation across firms -- in the writing of protocols for regular procedures, in the legal norms for liability in the event of failure, and in the design of systems to assign boundaries of responsibility and avoid cascades of failure, to name just a few general problems.

Back to the main question: How will the future of IoT impact telecom companies? They need to have teams of lawyers, engineers and MBAs design processes, platforms, and protocols for making applications work across firm boundaries. That is challenging for any organization, big or small, and it certainly will challenge telecom firms too. The big firms will offer turn-key solutions, as usual, and, as usual, buyers will resist "all-in-one-firm" solutions, and ask to mix-and-match from a variety of best-of-breed providers.

Sound familiar? Tech has lived with these sorts of challenges for decades ... and here we go again.