feature article
Subscribe Now

The Twitter of Things?

Developments in Embedded WiFi

The internet has been a massive game-changer for humanity. It started as a way for people to get information, then a quicker way to communicate, then a way to do business. And now… well, perhaps we’ve come full circle back to getting information. But it’s become clear that there is such a thing as too much information. Thanks to ubiquitous access to the many ways of keeping in touch with those people whom you know are interested in your every move, your every thought, your every… synaptic firing, we can all practically live in each other’s brains.

We can now know when you started breakfast, what you had for breakfast, whether you liked it, when you finished, and whether you cleaned the dishes after or just put them in a heap for later. We could, if you wanted, receive a record of every chew. We know when you’re excited, when you’re bored, when the weather is great, when you trimmed that carbuncle, when that ingrown hair started festering, when you learned that… well, it might just get to be a bit more than we need. But as far as you’re concerned, we’re all just dying to hear all this stuff.

Thanks to this spray of information – and you know? Let’s stop calling it information. It’s data. Information is data made useful. Most of this is not useful. So let’s start again: thanks to this spray of data, once-useful (or somewhat-useful) web sites like MySpace or FaceBook or even LinkedIn have decided they need to compete with Twitter and cater to the self-obsessed to the extent that any valuable nuggets are drowned in an flood of meaningless drivel and advertising. We can know everything and nothing about anyone. And yet, that long-promised ability to know something useful about something, like the fact that the garden is getting too dry and needs water, has not yet been widely enabled on the internet. It’s an internet of people. It is not yet, to use the words of ZeroG, an internet of things.

What do things have to say? Well, they could say things like “you’re out of milk” or “the heating oil is running low” or “I’m overheating and need to shut down” or “I’ve just moved from the office to Bob’s car.” They could say a lot of things, just like people can. But there’s one big huge difference: they can say only what they’re made to say. Nothing more. A sprinkler could be made to say, “I think I’m springing a leak.” And if set up to say only that, you’ll never hear it start saying, “OMG, that other sprinkler just sprayed me!” “Oh jeez, he did it again!” “Aaaargh, the dog just peed on me!” So the concept of allowing things to communicate bears the promise of some long-needed peace and quiet.

Giving things a voice

So how do we give these things a voice on the internet? They’ve been able to talk to each other for a while now using a variety of media, like Bluetooth and infrared and RFID. But that’s all proximity-based, and communication is direct. The internet plays no part. To get onto the internet, there are really only three ways: through a wired network, a wireless network, or the cell phone infrastructure.

Wiring? Uh-uh. Might be feasible for something fixed like the water heater, but requires punching through walls, is a general pain, and is just silly for anything that moves. Cell phone system? Um… that seems pretty well loaded already, and did you notice your last bill? Also, how long does it take for a connection to go through? How stable is it? Yeah, I think we’ll pass on that.

That leaves wireless, which, today, means WiFi. (I’m not going to get into the WiMAX debate here… you can’t make me…) WiFi has mostly been implemented as a way for computers to connect without a cable. It tends to be associated with the heavyweight protocols that a full-fledged computer can handle easily. It hasn’t been conceived of as a cheap capability that can be added to all kinds of things.

A number of companies are out to change that. The basic idea is to craft single-chip (or nearly-single-chip – we’ll get back to that) WiFi SoCs that can reduce the size and cost of adding WiFi capability to things. Three such companies have been making some news over the last while: G2 Microsystems (whom we covered briefly before), GainSpan, and ZeroG. These companies have articulated their target applications somewhat differently – G2 Microsystems started by focusing on asset tracking, GainSpin claims to focus on sensors, but, by and large, they are all broadening their attack.

Bursts of power, unforced integration, and speed

One of the big differences between standard computer WiFi and WiFi that’s embedded into things is the fact that most things aren’t powered. WiFi needs power. So embedded WiFi solutions have to be optimized to use as little power as possible. This is helped by the fact that most things don’t need to chatter constantly. Thankfully.  Finally, something with a sense of restraint! That means that, during the time they are saying nothing, they can be powered down. So two of the key metrics touted by these vendors are the sleep power and the wake-up time. Wake-up is important because, not only must the power come up and the circuit reactivate, but the WiFi must re-associate so that it can carry on a conversation.

GainSpan claims 1-5 µA sleep power and 6-7 ms wake-up time; G2 says 4 µA and 11 ms. Sleep power and wake-up time are, to some extent, a trade-off: the harder something falls asleep, the longer it takes to wake up, as any of you with teenage sons will know. And there can be various stages of sleep. So, for example, ZeroG outlines four sleep levels: “sleep,” with 220 µA and 0.5 ms; “deep sleep,” with 70 µA and 15 ms; “hibernate,” with 0.1 µA and 50 ms; and “off,” with 0.01 µA and 75 ms.

We won’t go into a line-by-line comparison of all of the other specs, but if we kinda squint to look at the broad differences between the offerings, it comes down largely to integration and data rate. From an integration standpoint, each company has a different approach. While each chip has an integrated processor for handling the WiFi stuff, GainSpan includes a separate processor for applications, G2 makes the WiFi processor available for application use, and ZeroG specifically relies on an off-chip processor or MCU for applications.

While integrating the apps processor makes obvious sense from a reducing-the-chip-count standpoint, there are two reasons why ZeroG decided against that. From a customer standpoint, they believe that their potential customers have established relationships with a particular processor or MCU vendor and have an associated code base in place that they’d rather not change. By working with any processor, ZeroG believes that it can make it easier for these customers by letting them keep the processor and code they already know and, presumably, love. From a business standpoint, ZeroG won’t be competing with the processor guys and so can benefit from the established sales channels that can bring them opportunities.

G2 appears to have taken a middle ground here; while less demanding applications can be implemented using the on-chip processor, they can also hook to external processors for more grueling work. G2 requires that Flash memory reside off-chip (as does ZeroG), while, consistent with the integrated apps processor, GainSpan has Flash integrated on the SoC.

The other integration aspect is inclusion of sensor support; this is relevant, since so many applications involve some kind of sensing. Both G2 and GainSpan have sensor interfaces built-in; ZeroG doesn’t,because the processors they connect to typically have them, and, since those processors would be handling the sensor data, it wouldn’t make sense to bring the data onto the ZeroG chip.

With respect to data rate, the first thing everyone agrees on is the fact that the data requirements are small, so blazing fast speed isn’t specifically a requirement. However, G2 claims that even though an 802.11b device can work on a faster 802.11g network, its presence slows down the entire network, so they have chosen to implement the full 802.11g standard, giving them up to 54 Mbps of carrying capacity. GainSpan and ZeroG are both in the 1-2 Mbps range.

The other practical matter with respect to WiFi is that building a module using one of these chips requires FCC approval in the US (and presumably the corresponding agency approval in other countries). So pre-approved modules are available for all of these chips: G2 and ZeroG provide their own, and RFM just announced a module using the GainSpan chip.

Obviously there’s a good mix and match of different philosophies and capabilities; everyone has a good reason for whatever they have or don’t have. And so they’ll duke it out. We all know that features aren’t always the prime determinant of who wins, anyway. The one thing that is clear is that WiFi is likely to find its way onto things around you. And hopefully they won’t be too verbose. Unless…

<nightmare> Personal WiFi sensors… a mood detector, constantly blaring “Happy!” “Happy!” “Bummed”… A hairpiece that can talk: “OMG it’s sooooo windy!”  “Oh no, I’m having a bad hair day!” Or perhaps personalized Homeropathic medicine, detecting useful remedies as you pass by them: “Mmmmm… beer!” From the internet of things to… the Twitter of things! </nightmare>


G2 Microsystems




Leave a Reply

featured blogs
Aug 15, 2018
https://youtu.be/6a0znbVfFJk \ Coming from the Cadence parking lot (camera Sean) Monday: Jobs: Farmer, Baker Tuesday: Jobs: Printer, Chocolate Maker Wednesday: Jobs: Programmer, Caver Thursday: Jobs: Some Lessons Learned Friday: Jobs: Five Lessons www.breakfastbytes.com Sign ...
Aug 15, 2018
VITA 57.4 FMC+ Standard As an ANSI/VITA member, Samtec supports the release of the new ANSI/VITA 57.4-2018 FPGA Mezzanine Card Plus Standard. VITA 57.4, also referred to as FMC+, expands upon the I/O capabilities defined in ANSI/VITA 57.1 FMC by adding two new connectors that...
Aug 15, 2018
The world recognizes the American healthcare system for its innovation in precision medicine, surgical techniques, medical devices, and drug development. But they'€™ve been slow to adopt 21st century t...
Aug 14, 2018
I worked at HP in Ft. Collins, Colorado back in the 1970s. It was a heady experience. We were designing and building early, pre-PC desktop computers and we owned the market back then. The division I worked for eventually migrated to 32-bit workstations, chased from the deskto...
Jul 30, 2018
As discussed in part 1 of this blog post, each instance of an Achronix Speedcore eFPGA in your ASIC or SoC design must be configured after the system powers up because Speedcore eFPGAs employ nonvolatile SRAM technology to store its configuration bits. The time required to pr...