Life in the Cloud (sans Harp and Halo)


  • Page 1 of 1
    Bookmark and Share

Sometimes a minor observation can lead you to realize that the world around you is fundamentally changing. Last week I went in for my yearly physical, which of course always involves an EKG. Usually, the nurse wheels in a cart with a cool-looking EKG machine and I obviously wonder whose boards and modules I might find if I opened it up. This time was different. The nurse wheeled in the cart but there was no EKG machine. Instead there was a normal, everyday laptop PC and a cluster of probes. The probe wires went into a small hub that connected to the laptop’s USB port. The entire functionality of the EKG machine was now in software on the PC.

Now, this is not in itself a major development; it’s actually fairly predictable. But it showcases two major things that involve the use and nature of embedded systems. First, the processor side of most designs is becoming commoditized and driven by the semiconductor manufacturers. Whether the processor is a laptop PC or an Atom or VIA-based COM module, OEMs mostly do not design the CPU side, they select it from what is available on the market. The corollary, then, is that the major design challenge is on the I/O side. You still can’t buy a set of EKG probes at Fry’s Electronics. The same goes for the I/O subsystems of a vast number of embedded controllers, instruments, data acquisition systems and more. Many, many of these require detailed engineering to enable them for their specific applications.

The second, equally obvious realization is that the other big design challenge is in the software. Older EKG machines would spit out a strip of paper with the waveforms on it, which the doctor would examine, perhaps discuss with the patient, make a few notes on and then fold up and stick into a folder. Now that same EKG data is in memory and disk, and although it is presented in the visual form familiar to the doctor, the data is also available for analysis such as flagging an alert or giving quantitative values for diagnosis. But the biggest thing is that the data is available online—to quickly send to a specialist for consultation—or for whatever purposes may lie beyond such as statistical analysis with the results of thousands of other patients, for correlation with other medical data like diet, medications. The possibilities are endless and that’s the point.

Now extrapolate to other small medical devices like O2 meters, blood analyzers and more. MRI machines are not based on PCs, but their basic data is digital. But now let’s multiply this by everything, by all the embedded controllers and instruments out there that have their I/O tendrils into the real world. They are not just pasting labels on bottles, running milling machines, industrial processes, transportation systems and climate controls. They are also collecting data. At first the data was intended for use by their own internal software, but thanks to today’s increasingly universal connectivity, that data is potentially available for use by applications both immediate and yet unheard-of. Today, embedded systems represent not only devices for control and automation; they are also the means for connecting the real world to “the cloud.”

For example, an organization wishing to monitor and analyze data as diverse as machine power usage, lighting, plant environmental conditions and more will not have to install separate data collection devices. Mostly the data will be available as a byproduct of the already installed control and automation systems and can be used and correlated across different domains at the IT level to implement all manner of management functions. Already transportation and fleet management systems are pulling together data from GPS, passenger counters, fuel consumption monitors, sensors in bearings, RFD monitoring of freight and traffic conditions for efficient routing, maintenance scheduling and payload tracking by pulling different aspects of this data together in different applications.

The convergence of the small embedded device with “the cloud” will open up huge new areas of innovation at the macro level as more and more devices are connected at the micro level. Software at the middle level—running on all those powerful processors we sometimes think have nothing to do—will have the ability to prepare data in preprocessed values and formats to make it instantly usable at the IT level. And have no doubt that clever minds at the IT level will be thinking of ever more creative and useful ways they can use that data to make their own overall operations more efficient and cost-effective. Break out the harps and halos.