Machine-to-machine integration may be on the horizon in a big way, Carl Zetie says. Adoption of new technologies will make it happen, but the stumbling block will be people's reluctance to be made redundant.

InformationWeek Staff, Contributor

April 8, 2003

6 Min Read

Machine-to-machine integration might well be the "next big thing" for the connected enterprise. Applied to the right business processes, machine-to-machine integration has the potential to lower costs, improve responsiveness, tighten the supply chain, eliminate inefficiencies and, in the most exciting cases, create whole new business models. Long limited to a few specific vertical functions such as industrial automation or confined to labs as proofs of concept, the potential for machine-to-machine integration in conventional IT organizations is starting to generate interest in the venture-capital community, in a few vendors, and in leading-edge companies. In the same way that the mainstreaming of key technologies beginning about five years ago gave impetus to the subsequent adoption of handhelds and other mobile devices in the enterprise, a similar convergence of enabling technology changes may be happening now with machine-to-machine integration.

The idea is certainly nothing new, at least not in certain industries. In fact, I drew my first IT-related pay packet two decades ago as an apprentice programmer on a building-management system, sometimes also called a building automation system. A building-management system is a classic example of an integrated machine-to-machine system that combines embedded sensors, actuators and activators, and a hierarchy of controlling computers to manage the environment of a building without human intervention. Sensors measure the temperature, humidity, and sunlight throughout the building while actuators operate heating and air-conditioning units, dampers, and humidifiers; open and close windows; and operate shades. Predictive algorithms optimize themselves based on rates of change to keep the environment comfortable while minimizing heating costs. Human intervention is rarely required unless the desired operating parameters change or equipment requires servicing.

This same concept of a "closed loop" of sensors, controllers, and outputs is starting to make itself felt in the broader world. Part of the reason is the emergence of certain enabling technologies combined with the increasing maturity and declining costs of others. The same economic drivers that made PDAs, smartphones and Web cameras so affordable will make embedded devices of similar power and connectivity (but without a human interface) available for enterprise purposes. GPS and cellphone-based location technologies are making it possible to know the locations of assets and people more cheaply and accurately. Radio-frequency identification (RFID) transponders are reaching the price where it's economical to tag and track individual pallets of goods. A variety of wireless technologies, including Bluetooth, Wi-Fi, and cell networks are making it easier and more economical to connect everything to everything else, and in particular to connect devices dynamically as people, products, and assets enter and leave each others' vicinity.

Some of the possible applications are quite simple and offer fairly obvious economies of automation, while others are just emerging and offer more provocative opportunities for change. In many parts of the country you can drive through highway toll plazas without stopping (or in some places, even slowing) as a transponder in your car automatically signals your passing to an overhead receiver, which in turn automatically bills your credit card to top up your pre-paid account if necessary. Not only is this more convenient for the driver, it's considerably better business than paying somebody to sit in a cramped tollbooth, breathing in the vehicle fumes, making change from a dollar for a 50-cent toll. A more-recent addition is cameras that capture the license plate of toll violators, use character recognition to read the plate and determine the registered owner, and automatically send off a citation and fine. More innovatively, self-service checkouts are starting to appear in some stores. For low-value items, self-service means swiping the bar codes yourself; for big-ticket items, it's reasonable to put transponders on each item, which then signal their presence in your shopping cart as you push it through the checkout.

Within the enterprise, the most obvious early applications are in the supply chain. If you know where each truck and driver is, you can use your assets more efficiently--not just the trucks, but the warehouses that are waiting for them to pick up or the recipients that are waiting for them to deliver. In retail, if you can track products more carefully you can dramatically reduce shrinkage, or product that unaccountably disappears. Countering losses due to theft is one obvious benefit of better tracking, but a remarkably large proportion of shrinkage--as much as 18% in some reports--is put down to administration errors that can be targeted with machine-to-machine integration technologies.

Other possibilities are starting to present themselves as the enabling technologies cost less and become more standardized. In the supply chain, some companies are moving beyond tracking each truck to tracking each pallet, giving them far more detailed information and greater control. In other places, processes that were improved once by giving employees mobile technology can be improved again by eliminating unnecessary human intervention entirely. For example, when you return a rental car to the compound at the airport an employee may come up to with a handheld device to complete your check-in. This is much more convenient than the old system where you had to stand in line to return your keys and fill out paperwork--but is that person really necessary? Why not put an "asset tag" on the vehicle that is detected by sensors in the compound as you drive in, and a wireless connection from sensors in the car that transmit the mileage and gas level more rapidly and reliably than an employee reading it off the dashboard? A receipt could be automatically mailed or E-mailed to you. Your return process would be so fast you wouldn't even know it had happened, and the rental car company would need to keep far fewer people around just to handle exceptions such as a damaged vehicle.

As the enterprise and the individual become more "instrumented", ever-more-sophisticated integrations become possible. Consider this professional scenario: you're driving to the airport when the hazard warning system in your radar detector picks up an accident ahead. It notifies your GPS-enabled in-car navigation system, which checks wirelessly with the highway patrol's computer and confirms that there is a delay. Recomputing your route, it sends your new travel time to your PDA. The PDA discovers that you won't make your flight so it wirelessly contacts a server back at your office, which rebooks you on a later flight while alerting the limo service that is booked to pick you up. Within minutes, a voice message is delivered back to your car reassuring you that you don't need to drive aggressively to make up time as everything has been taken care of. All of the technology to make this happen exists today; it just needs to be integrated.

Just as with the evolution of mobile technologies, the impact of machine-to-machine integration will be felt both in relatively mundane incremental efficiencies as well as in radical reinventions of processes--and those reinventions are largely impossible to foresee. The biggest barrier to machine-to-machine integration's adoption and integration into our daily work lives will probably be human reluctance to relinquish decision-making to automated systems. In fact, in some large buildings controlled by a building-management system, individual offices still have thermostats. Often those thermostats aren't really connected to anything, they're just there to make the occupants believe that they still have some control and to smooth the way to machine-to-machine automation.

Carl Zetie is an analyst with Forrester Research.

Never Miss a Beat: Get a snapshot of the issues affecting the IT industry straight to your inbox.

You May Also Like

More Insights