At ITEXPO last week I had a number of meetings with companies in the M2M space as well as the leaders of some of the forums.
I met with Gemalto, Omnilink and Inmarsat to understand their M2M plays. Each of these three companies is a part of the emerging M2M eco-system, and together they show the complexity of our emerging machine world. Gemalto, in addition to being a major supplier of SIMs to the cellular industry, has focused on the MIMs and other modules of the M2M market. The company seems to have a good range and is working on the edges of the network to deliver M2M value. So they are a device enabler.
But the devices may be part of multiple applications. Inmarsat delivers a satellite network capability that enables devices to communicate when outside of the cellular coverage area. In talking with the Inmarsat folks I realized a couple of things. While the original Law of Bandwidth had only three types of bandwidth (desktop, nomadic, and cellular wireless), satellite is a fourth. It is interesting to contemplate the value of this, especially when the uplink includes aggregation (a hot spot).
The point is clear - while the cellular network may have a footprint that covers me 99 percent of the time, the forthcoming "Internet of Things" will have many devices that are outside that coverage area.
The next generation of satellite networks seems to deliver the right capabilities for many of these applications. The last company, Omnilink is delivers vertically integrated M2M systems such as offender monitoring (criminal tracking), vehicles and other things. Where there are lots of companies building components, Omnilink is building vertically integrated systems, but they are only for single purposes and only viable for applications where the revenue/value in that one application is adequate.
What became clear is that for this new M2M world to truly explode, it cannot be just a bunch of tightly linked vertical apps, it needs to be a combination of sensors and sources that support many apps and apps that can utilize a wide range of sensors.
The other meeting I had was with representatives of the one M2M and OMA (Open Mobile Alliance). In our discussions, an area that came up was how applications would get data from the aggregation points that will aggregate data in these networks. As we talked, the concept that M2M and WebRTC may have much in common became clear.
For example, an aggregation point may gather data from a number of sensors (in your house, in a business, on a vehicle, etc.). If the data is all moved to the cloud, then the apps must get the data from the cloud. An alternative is to treat the aggregation point as a cache and have an amount of data stored there. Then the apps could get the data they needed from the aggregation points. For example, an app tracking location could get the location reading every five minutes if that is the location resolution required, while another app might get a higher resolution for another purpose. What is interesting is that if the policy server that allows apps to connect to aggregation operates using similar protocols and APIs to WebRTC, an interesting "Web of Things" is created. The policy engine could "getdata" like the WebRTC "getmedia" at the aggregation point, and a peer connection could be used to send the data to the app.
The key to this concept is that we can now have an unlimited number of data aggregators (think of them as "data servers") and an unlimited number of apps (think of them as "data browsers"). This network then looks like the Web and has all of the same characteristics of openness and innovation. And when you realize that a significant percentage of the aggregation points may also include real-time human interaction, it becomes even more interesting. And if the protocols are integrated, then the deployment and use models can build off of each other. Anybody up to talk Web M2M?
Edited by
Rory J. Thompson
