Practically every keynote speech given at technology conferences seems to offer some CEO's vision of the future, often spun in such a way that the keynoter's company appears to be leading the way into Tomorrowland. Now that Embedded Systems Programming is verging on its 15th anniversary, it's time for us to do our own crystal ball gazing.
If you look back over the past twenty years or so and then try to project forward, a couple of things stand out. First of all, increasing semiconductor integration has been the rule and seems likely to continue. Pundits say that Moore's Law still has legs. That being the case, you'll see see processing power and available memory continue to increase.
There must be a corollary to Moore's law relating to the software content of systems as well. Fifteen years ago, 64K was all you could expect to have, and all your code had to fit in that space — unless you came up with some sort of bank-switching scheme. The rule that software expands to fit available memory is as true in embedded systems as it is on the desktop. That trend will surely continue, which means that more of the functionality of systems will shift to software.
The application of processing power to an ever-widening circle of products will continue as well. As processors get cheaper, they can be deployed in more and more products. And for good reason. Applying embedded technology to products makes them cheaper to build, more reliable, safer, more feature rich, and likely to consume less power.
In anticipation of Embedded Systems Programming 's 15th anniversary, we called on several industry veterans to give their perspectives on the direction embedded technology will take over the next few years.
Nick Tredennick and Brion Shimoto report on microelectromechanical systems (MEMS — a name that by itself justifies the existence of acronyms) and conclude that they are “a dream come true” for embedded systems designers.
In “Embedded processors of tomorrow” Jim Turley tells us that in 15 years, microprocessors will contain 4.5 billion transistors and run at 250GHz — and they'll adapt their features while the system is running.
For many, the common platform of the future may very well be FPGAs, according to Bob Zeidman. Not just FPGAs, but big ones with CPUs — maybe several — in them. And new hardware-software codevelopment tools will match their capabilities.
The future will definitely be wireless, says Larry Mittag, but only after the technologies get sorted out. But, he cautions, “The number of wireless protocols is not likely to lessen; nor is the rate of innovation that defines new protocols likely to decline.”
While these visionaries have their view of the future, Bernie Cole thinks the future is yet to be invented. He does think that the future holds all sorts of wireless, Internet-connected, multimodal computing devices. He just hasn't a clue what they'll be.
Since technology builds on technology, the rate of proliferation of embedded technology can only increase. While we have seen amazing advances over the past several years, the most amazing advances are yet to come.