yeah, I'll believe it when I see it: Cisco heavily involved in low-margin products. Can you think of any company less culturally suited to being the $2 sensor in my flowerpot, or the fitbit/tracker in my cat's collar.
Even if 50 billion devices by 2020 is a reasonable prediction, to generate $14.4 trillion in sales means an ASP of $288 per device.
Move the decimal place one position to the left and then the 50B unit volume might actually happen.
I don't know that I could get behind such a pie-in-the-sky prediction as $14 trillion without some sound evidence to back it up.
No question about it, though. It will be a huge market with a huge impact on just about everything. There will be plenty of $2.00 or $0.50 mini systems in flower pots, cat collars, greeting cards and pretty much everything else.
But, as Mark stated, much of that probably won't be made by Cisco. A good deal of it will be assembled in some distant no-name factory where labor is cheap.
The money won't be in the hardware, it'll be in the services provided.
I could see home monitoring subscriptions (like your alarm system today). Equipment service contracts. Cloud computing services to aggregate multiple sensor outputs and perform the necessary calculations that the low power sensors would not be capable of.
Definitely a solution looking for a problem, but if there is enough marketing hype, the problem will eventually be perceived.
The only REALLY likely market to come out of all this is all those electricity, water and gas meters needing to be read automatically at low cost, if only because public utilities generally have easy access to the capital needed to fund large projects. (The electricity meter's needs are somewhat unique because it tends to merge somewhat with general power distribution management or the "smart grid".) However because the data rates needed AT EACH METER are quite low it's unclear how such a network could be done efficiently. Some have envisioned an all-wireless "self-organizing" system (with individual nodes acting as data concentrators) and that no doubt could be made to work, but why would each utility need to fund its OWN network? It's also not clear why smothering individual OSes with a "universal API" helps reduce costs,
if the need could be demonstrated then this industry needs more of its own "Android-like" platform to build on, but it's not clear that the demand exists for even that. Maybe it would make more sense to "build out" these networks based on existing utilities, maybe it would make more sense to envision other models, for example you might consider something that would look more like the old "pager" systems in reverse (ie a central system COLLECTING small amounts of data when polled), maybe even accessed through the existing cable TV or internet distribution access system. In any event a 14 trillion dollar market looks just about absurdly optimistic.
I wonder how security will be provided for so-called "connect everything" IOT network. Nearly 20 years ago (long time, isn't it?). IPv6 designers included IPsec as standard feature for security, but IPsec is notoriously unpopular due to huge implementation diversity (=incompatibility) and difficult configuration.
Will "40Kbyte footprint" IPv6 stack support IPsec and IKEv2, or TLS1.0(likely, because SSL/TLS is selected for 6LowPAN architecture), or nothing at all?
Always being wary around April 1st I double checked the veracity of this $14.4 x 10E12 claim. Numerous other news source also reported Mr. Lloyd's remarks - I thought perhaps he meant Billion with a B as in 10E9 - nope - seven other news outlets say trillion as well and surprisingly, none challenged it. So either this is "the mother of universally coordinated April Fools spoofs" or a new pinnacle of "irrational exuberance". Don't get me wrong, I think IoT will be a very important technology but fourteen trillion+ in a little over 6.75 years? Compared to that the market growth of the iPhone looks like a flat-line dud!
Blog Doing Math in FPGAs Tom Burke 13 comments For a recent project, I explored doing "real" (that is, non-integer) math on a Spartan 3 FPGA. FPGAs, by their nature, do integer math. That is, there's no floating-point ...