New advances in IoT edge computing frameworks
Open source project EdgeX Foundry aims to develop a standardized interoperable edge computing framework for the Internet of Things.
In April, the Linux Foundation launched an open source project, EdgeX Foundry, to develop a standardized interoperability framework for edge computing in the Internet of Things. Just recently, EdgeX Foundry announced the addition of eight new members, bringing its total membership to 58.
These new members are Absolute, IoT Impact LABS, inwinStack, Parallel Machines, Queen's University Belfast, RIOT, Toshiba Digital Solutions Corporation and Tulip Interfaces. Its original members include AMD, Analog Devices, Canonical/Ubuntu, Cloud Foundry, Dell, Linaro, Mocana, NetFoundry, Opto 22, RFMicron and VMWare And other companies or organizations.
The EdgeX Foundry project builds on dell's earlier FUSE iot middleware framework based on the Apache2.0 protocol and includes a dozen microservices and more than 125,000 lines of code. After FUSE merged a similar program, AllJoyn-Compliant IoTX, the Linux Foundation worked with Dell to create the EdgeX Foundry, The latter is a project spearheaded by existing EdgeX Foundry members Two Bulls and Beechwood.
EdgeX Foundry will create an ecosystem of interoperable, plug-and-play components for iot edge computing. The open source EdgeX stack will coordinate various sensor network protocols with various cloud and analytics platforms. The framework aims to exploit interoperability code across modules such as edge computing, security, systems administration, and services.
The key benefit for project members and their customers is the ability to integrate a variety of pre-certified software into many IoT gateways and smart edge devices. In an interview with http://Linux.com, IoT Impact LABS principal Engineer Dan Mahoney said, "In reality, the EdgeX Foundry lowers the challenges we face when deploying multi-vendor solutions."
With the Linux Foundation still incorporating the AllJoyn specification under its AllSeen Alliance project into the IoTivity standard, why launch another iot standardization project (EdgeX Foundry)? For one thing, EdgeX Foundry is different from IoTivity, which addresses industrial iot issues, while EdgeX Foundry aims to address the full range of consumer and industrial IOT issues. More specifically, EdgeX Foundry aims to be a general-purpose middleware for gateways and intelligent terminals. Another difference between EdgeX Foundry and IoTivity is that the former wants to shape a new product with pre-certified terminals, while the latter addresses more interoperability between existing products.
"IoTivity provides the protocol to enable seamless connectivity between devices, while EdgeX Foundry provides an edge computing framework," said Philip DesAutels, Senior Director of IoT at the Linux Foundation. EdgeX Foundry is compatible with any protocol devices such as IoTivity, BacNet, EtherCat, and others, enabling a common edge computing framework that integrates multi-protocol communication systems. The goal is to reduce uncertainty and time to market as we build an ecosystem of interoperable components. To better generate economies of scale."
Last month, the IoTivity project, sponsored by the Open Connectivity Foundation (OCF) and the Linux Foundation, released IoTivity 1.3, which adds ties to its one-time rival, AllJoyn Spec, Also added an interface to the UPnP device discovery standard for OCF. Further integration of IoTivity and AllJoyn is expected in IoTivity 2.0.
DesAutels told http://linux.com that IoTivity and EdgeX are "highly complementary" "due to the fact that several members of the EdgeX Foundry project are also IoTivity or OCF members, This reinforces the IoTivity and EdgeX partnership."
While IoTivity and EdgeX both claim to be cross-platform, both in TERMS of CPU architecture and OS, there are some differences. IoTivity was originally designed to run on the Linux platform and is compatible with Linux operating systems such as Ubuntu, Tizen and Android, before gradually expanding to Windows and iOS. The EdgeX was designed to be cross-platform, perfectly compatible with all CPU architectures, Linux, Windows, Mac OS, and, in the future, real-time Operating systems (RTOS)."
EdgeX's newest member, RIOT, offers an open source iot oriented project, RIOT RTOS. RIOT's chief maintainer, Thomas Eichinger, said in a recognition speech: "Since RIOT was originally designed to address issues where Linux didn't fit, it was natural for the RIOT community to be interested in participating in and supporting open source organizations like EdgeX Foundry, an edge computing community."
Simplification of sensor integration
IoT Impact LABS (Impact LABS or simply LABS) is another new EdgeX member. The company has launched a unique business model designed to help smes get through the trial phase of iot solutions. Most of the company's clients, including several EdgeX Foundry project members, are dedicated to building smart cities, re-using infrastructure, improving food safety, and addressing the challenges of society's lack of natural resources.
"We spend a lot of time at LABS reconciling the differences between pilot customer solutions," says Dan Mahoney. "EdgeX Foundry minimizes the effort to deploy edge software systems, enabling us to deploy high-quality solutions faster and better."
This framework is especially advantageous in scenarios involving multiple vendors and multiple sensor types. "Edgex Foundry will give us the ability to quickly build a gateway that can control all deployed sensors." Mahoney added. Sensor manufacturers will use the EdgeX SDK to burn application-layer protocols to drive edge devices that are compatible with multiple vendors and solutions.
Building edge analysis capability
When asked how his company would like to see EdgeX Foundry evolve, Mahoney said, "One of our favorite goals is to have more effective industry protocols for device services, a clearer path to edge computing."
Edge computing is growing in both industrial and consumer iot. In the latter, we've already seen edge computing analytics integrated into several smart home systems such as Alexa's smart voice and video analytics. This reduces the computing load on the cloud service platform, but also introduces security, privacy, and service disruptions due to vendor outages or latency issues.
For industrial Iot gateways, latency becomes the primary problem. As a result, there are some expansions in iot gateways that resemble cloud services. One solution is to integrate some security assurance applications on cloud services into embedded devices with containers such as RIOS and Ubuntu kernel snapshot mechanisms for security purposes. Another option is to develop an IoT ecosystem that migrates cloud capabilities to edge computing. Last month, Amazon released AWS Lambda implementing the AWS Greengrass iot protocol stack for Linux-based gateways. The software enables AWS computing, message routing, data caching and synchronization capabilities on networked devices such as Internet of Things gateways.
Analytics is a key feature in EdgeX Foundry's evolution. One of the founding members is Cloud Foundry, which aims to integrate its major industrial application platforms into edge devices. Another new entrant, Parallel Machines, plans to use EdgeX to bring AI to edge devices.
EdgeX Foundry is still in the early stages of the project, the software is still in phase ALPHA, and its members just had their first all-hands meeting last month (June). There are already some initial training sessions for new developers, as well as more information.
By ERIC BROWN, Translator: Penghuster, proofreader: WXY
This article is originally compiled by LCTT and released in Linux China