Japanese semiconductor vendors have mostly stayed away from the maker market, at least outside Japan, as most people would be hard-pressed to come up with a hobbyist development board powered by processor or micro-controller from Toshiba, Sony, Renesas or other Japanese companies, despite the three aforementioned names being in the top 20 semiconductors companies. I can only remember having written about Fujitsu F-Cue 96Boards, as well as Renesas GR-PEACH mbed board since I started this blog 7 years ago. Renesas seems to be the only company to have a real community behind with their “Gadget Renesas” pink-colored development boards, and the latest and seventh board is GR-LYCHEE powered by Renesas RZ/A1LU ARM Cortex-A9 processor, and equipped with a WiFi & Bluetooth module, and a camera.
Renesas GR-LYCHEE board preliminary specifications:
- Micro-processor – Renesas RZ / A1LU (R7S721030VCFP 176-pin QFP) ARM Cortex-A9 Processor @ 384 MHz with 3MB on-chip SRAM
- Storage – 8 MB flash+ micro SD card
- Connectivity – 802.11 b/g/n WiFi, Bluetooth 4.1 LE via ESP32 wireless module
- Audio – 3.5mm audio jack (heaphone + mic)
- USB – 1x USB host port
- Camera – 1x camera interface for VGA (640×480) camera
- Expansion – Arduino UNO headers
- Debugging & Programming – 1x micro USB port, JTAG interface
- Misc – 32.768 Hz RTC clock, 2x user buttons, reset button, 4x user LEDs
- Power Supply – 5V via 1x micro USB port; operating voltage: 3.3 V / 1.18 V
The board is mbed compatible so at launch you’ll be able to use the mbed compiler with the board. The board is still in beta version, documentation is still being worked on, and launch is scheduled for the end of November 2017. While most Gadget Renesas’ users are likely in Japan, Renesas also organized events in India, ASEAN, and Oceania with GR-PEACH board earlier this year as you’ll find out by visiting the community’s English page.
Documentation and more details about GR-LYCHEE board should eventually surface in the product page (in Japanese only for now).
Jean-Luc started CNX Software in 2010 as a part-time endeavor, before quitting his job as a software engineering manager, and starting to write daily news, and reviews full time later in 2011.
Support CNX Software! Donate via cryptocurrencies, become a Patron on Patreon, or purchase goods on Amazon or Aliexpress
I would really like them to release a version of the RZ with 32mb or 64mb of SDRAM/DDR built in. You can just about run linux on the RZ with 10MB if SRAM but it’s pretty tight.
Something that could run linux but isn’t a multi core 1ghz+ thing would do well in the IoT arena. We want linux because we want memory protection, proper libraries etc but it needs to get down to the ~$6 with boot storage, memory and wifi level first.
You could go for a RDA M2M soc http://www.rdamicro.com/Products/Detail_436.aspx
@dgp
How about a 7688-based device such as the Omega Onion 2? Those are $5. VoCore was selling a 7688 device for $4 but seems that was only the Kickstarter price. I bought a no-name 7688 device off eBay for $11 which can be had for $8 in bulk from the company as well as the LinkIt Smart which is $12 and available on DigiKey.
Seems like everyone wants to go higher and higher speed (with a corresponding higher and higher cost) but there are no good $6 low-powered battery-saving Linux devices that can suspend to disk or RAM and wake in milliseconds. Such a device would be a big boost for IoT I would think. A sensor that only updates daily, that I can apt-get updates and runs for a year on a coin cell? Fantastic.
apt-get running on a coin cell? LOL. Each daily update of apt-get on something without few gigahertz, few gigabytes and an SSD of ~1GB/s will run for a week. Meet the modern Linux. Dudes who don’t use Linux for IoT – they aren’t mad, they just want to do something which really works. On a coin cell.
@anon
Why would it need apt-get? IoT would be a yocto style static build. And yes IoT vendors do want linux.
*works for an IoT vendor*
@SlowBro
The Mediatek and RDA parts are close but would be better from a big name vendor.
Something like the im6ull with integrated memory would be perfect too.
> And yes IoT vendors do want linux.
Of course, everyone wants Linux because it means a massive reduction in effort to develop software.
If only companies kept caring after shipping a device with Linux. Too often the issue is that they ship a firmware based on the vendor’s Linux SDK (already horribly out of date) and then never ship any updates after the retail version.
IoT doesn’t have to be a tire fire of insecurity, but hardware companies don’t want to spend more than the bare minimum needed to get their product to market, and so we have shit security.
You make that sound like a bad thing. Imagine this: Everyone comes up with their own crappy implementation of TLS that just about barely works with all sort of implementation mistakes etc vs everyone contributes to one or two TLS libraries that multiple parties are actively auditing etc? Of course there are problems with big issue like heartbleed that end up affecting a lot of people but I think that’s a lot better than hundreds of implementations of the same thing all with bugs as bad as heart bleed.
People seem to think cheap Chinese webcams running Linux equal IoT. I’m not sure why.
Running linux is actually much much more expensive that running everything on a microcontroller that has no real memory protection. If running Linux gets cheaper the potential to have things that aren’t super easy to hack because of some trivial programming error goes up massively.
I didn’t mean it as a bad thing. Obviously it’s better for everyone if the vendor uses Linux and doesn’t roll their own TLS stack. This is exactly what I meant by “a massive reduction in effort to develop software.”
Because “IoT” is like “big data” in that it’s a self-defined term by marketing wanks who want to sell more of their product by attaching the latest buzz words to the spec sheet.
There’s a big difference between consumer “IoT” devices and industrial “IoT” devices. I’m not saying either has better security, I’ve seen some fantastically shitty implementations on expensive industrial equipment (seriously, ditch ModbusTCP, it’s from the 80’s and has no security whatsoever!)
Running Linux will always be more expensive than some 32-bit M0 micro that a company can throw FreeRTOS on (or something worse). From a hardware perspective of course it’s more expensive, the Linux kernel contains vastly more features than any RTOS. Now whether your application requires these or not is a question for the engineers.
This I really have to disagree with. Bad/rushed programmers can make severe security mistakes whether the thing is running a custom firmware or Linux. Just because you are running a Linux kernel doesn’t mean they can’t make a trivial off-by-one or buffer overflow in the application that’s running in userspace (obviously as root).
From the perspective of someone who has to deal with the business end of an attack, it doesn’t matter what OS the device is running, or whether it was the kernel or a user space program that was compromised. Any compromise of the device by a third party for their nefarious use is something of concern.
Just because a device is running Linux doesn’t automatically mean the vendor won’t put their “secure as a screen door” software on it and never update it.