All News | Boards | Chips | Devices | Software | Archive | About | Contact | Subscribe
Follow LinuxGizmos:
Twitter Google+ Facebook RSS feed
*   get email updates   *

Linux-driven IoT gateway hooks up to Google Cloud IoT services

Dec 20, 2017 — by Eric Brown — 1177 views

[Updated Dec. 24] — SolidRun’s “ClearCloud 8K” IoT gateway runs Linux on a new “Single Shot” version of its quad -A72 MacchiatoBIN SBC with 1-, 2.5-, and 10GbE ports, and links to Google Cloud IoT Core.

SolidRun has launched a MacchiatoBIN ClearCloud 8K IoT gateway appliance with built-in software for connecting to Google’s beta-level Cloud IoT Core services (see farther below). The $399 box is built around SolidRun’s new, $269 “Single Shot” version of its $349, open-spec, Marvell MacchiatoBIN Mini-ITX networking SBC.

ClearCloud 8K, front and back
(click images to enlarge)

SolidRun calls the ClearCloud 8K “simple and powerful.” It’s simple only in the sense that from a hardware perspective, it’s nothing but a fanless, 182 x 174 x 44mm enclosure for the MacchiatoBIN mainboard. But the MacchiatoBIN is anything but simple, and it is indeed quite powerful.

The MacchiatoBIN board is based on the headless Marvell Armada 8040 SoC, which has 4x Cortex-A72 cores. The original MacchiatoBIN — now called the MacchiatoBIN Double Shot — is clocked up to 2.0GHz. The MacchiatoBIN Single Shot used on the ClearCloud 8K clocks only to 1.6GHz. Both models provide a single 2.5GbE ort and 2x 10GbE ports, all furnished with fiber SFP+ connections. There’s also a standard Gigabit Ethernet RJ45 port designed for a copper line.

However, the Single Shot board and the ClearCloud 8K lack the additional RJ45 copper ports for the two 10GbE connections. In addition to the clock rate and price, this appears to be the only difference between the two models.

The ClearCloud 8K’s MacchiatoBIN Single Shot mainboard (left) and detail view for almost identical MacchiatoBIN Double Shot (note the missing RJ45 ports for 10GbE connections on Single Shot)
(click images to enlarge)

Aside from the clock rate, both versions of the Marvell Armada 8040 appears to be the same. Networking functions include a packet processor with a security co-processor and extensive network protocol support. SolidRun previously partnered with Marvell when it released its earlier, Armada 38x based ClearFog Pro and ClearFog Base networking boards. The Israeli firm is also known for its NXP i.MX6 based HummingBoard SBCs.

The 170 x 170mm MacchiatoBIN Mini-ITX board ships with schematics and layout files. It offers an open source, mainline Linux 4.4x BSP with support for Yocto 2.1 and SUSE Linux.

The ClearCloud 8K enclosure exposes all the MacchiatoBIN board’s ports, including the Ethernet ports as well as single USB 3.0 and micro-USB ports. Like the MacchiatoBIN board, the ClearCloud 8K ships with 4GB or 16GB of DDR4 RAM plus a microSD slot. The spec list also lists “eMMC,” but without a capacity. Both the 4GB and 16GB RAM versions of the Single Shot board ship with 8GB eMMC.

MacchiatoBIN Single Shot block diagram
(click image to enlarge)

Other specs appear to be almost identical, including 3x SATA, 4x PCIe, 2x UART, and JTAG connections. There’s also a Marvell TDM low-speed module expansion header with SPI, I2C, TDM, and power interfaces. The only major omission we see from the 12V board is the SBC’s USB Type-C port with MoChi MCI interface. (For more details, see the diagrams above, as well as the spec list in our earlier MacchiatoBIN report.)

Google Cloud IoT Core

Google’s Cloud IoT Core managed service was announced in May as a way to “securely connect and manage IoT devices at scale.” The service was released as a public beta in September.

Cloud IoT Core conceptual diagram
(click image to enlarge)

Cloud IoT Core part of a larger Google Cloud IoT platform, which Google describes as “a set of fully managed and integrated services that allow you to easily and securely connect, manage, and ingest IoT data from globally dispersed devices at a large scale, process and analyze/visualize that data in real time, and implement operational changes and take actions as needed.”

Cloud IoT, in turn, is related to Google’s growing stable of Google Cloud Platform (GCP) offerings. These include Google Storage and the Cloud Vision API tapped into by Google’s recent AIY Vision Kit for the Raspberry Pi Zero W, which performs TensorFlow-based vision recognition. When used in conjunction with Google Cloud IoT, you can connect Cloud IoT Core to many of these Google Cloud services, including Pub/Sub, Dataflow, Bigtable, BigQuery, and Machine Learning Engine.

Unlike Amazon’s AWS IoT platform, Cloud IoT Core appears to lack any provision for integrating with local analytics, which Amazon provides via its AWS Greengrass stack for Linux-driven devices connecting to the AWS Cloud. However, with the ClearCloud 8K providing several 10GbE connections, you’re likely to have a nice tight connection with Google’s cloud services.

Cloud IoT Core maintains a logical representation of the physical IoT device, including device properties and last reported state. APIs are provided for applications to retrieve and update device properties and state even when the device is not connected.

For security, you get MQTT and asymmetric key-based authentication per individual device enabled via Certificate Authority (CA) verification. The CA function appears to be enabled on the MacchiatoBIN’s Armada 8040, as Marvell is listed among the first Cloud IoT Core partners.

The open source Cloud IoT Core is free for the first 250MB per month. Standard pricing starts at $0.0045 per megabyte per month for up to 250GB per month volume, with lower per-megabyte prices for higher volumes.

Further information

The MacchiatoBIN ClearCloud 8K is available now starting at $399. The ClearCloud 8K is intended for evaluation use only, and lacks FCC resale compliance for resale. More information may be found on SolidRun’s ClearCloud 8K product and shopping pages.

The new MacchiatoBIN Single Shot SBC is detailed on the standard MacchiatoBIN product page, and is available separately for $269 with 4GB RAM and 8GB eMMC, or $399 with 16GB RAM and 8GB eMMC.

(advertise here)

Print Friendly, PDF & Email

Please comment here...