Computing at the Edge of IoT – Google Developers – Medium
We’ve seen that demand for low latency, offline access, and enhanced machine learning capabilities is fueling a move towards decentralization with more powerful computing devices at the edge.

Nevertheless, many distributed applications benefit more from a centralized architecture and the lowest cost hardware powered by MCUs.

Let’s examine how hardware choice and use case requirements factor into different IoT system architectures.


  1. Tomi Engdahl says:

    New Architectures Bringing AI to the Edge

    As artificial intelligence (AI) capability moves from the cloud to edge, it is inevitable that chipmakers will find ways to implement AI functions like neural-network processing and voice recognition in smaller, more efficient, and cost-effective devices.

    The big, expensive AI accelerators that perform tasks back in the data center aren’t going to cut it for edge node devices. Battle lines are being drawn among various devices — including CPUs, GPUs, FPGAs, DSPs, and even microcontrollers — to implement AI at the edge with the required footprint, price point, and power efficiency for given applications.

    To that end, a pair of intriguing architectures created specifically for implementing AI at the edge are being introduced at the Linley Processor Conference on Tuesday by Cadence Design Systems and Flex Logix Technologies. Both focus on bringing AI functionality into edge node devices with an emphasis on reducing the memory footprint.

  2. Tomi Engdahl says:

    Startup Shifts Cloud Services to IoT
    Mimik drives microservices to end nodes

    Startup Mimik released software that enables the equivalent of cloud services to run on devices including end nodes in the internet of things (IoT). The so-called edgeSDK aims to lower response times and open up new use cases.

    The code lets any device running a popular operating system host the kind of microservices usually managed by the likes of AWS or Microsoft Azure. Early users include companies in gaming and health care, as well as Lime Microsystems, that will bundle the SDK with its open-source base stations.

    Carriers in the Facebook-led Telecom Infra Project aim to use the software “so they don’t have to go to a cloud service for communications in villages in Africa where connecting to a remote data center over an expensive satellite link would not be viable,” said Fay Arjomandi, Mimik’s co-founder and chief product officer.

    Facebook Likes $1K Base Stations
    Carriers to test open source hardware in 2018

  3. Tomi Engdahl says:

    AI Edges to Factory Floor

    Deep neural networks are crawling toward the factory floor.

    For several early adopters, neural nets are the new intelligence embedded behind the eyes of computer-vision cameras. Ultimately, the networks will snake their way into robotic arms, sensor gateways, and controllers, transforming industrial automation. But the change is coming slowly.

    “We’re still in the early phases of what’s likely to be a multi-decade era of advances and next-generation machine learning algorithms, but I think we’ll see enormous progress in the next few years,” said Rob High, chief technology officer for IBM Watson.

  4. Tomi Engdahl says:

    APAC firms look to edge for faster response but worry over data security

    Edge computing is being sought out for faster response and cost savings, but there are concerns about security and latency when large volumes of data are processed on such platforms.

    Organisations in Asia-Pacific are seeking out edge computing in search of faster response and cost savings, but they also have concerns about security and latency when large volumes of data are processed on such platforms.

    A primary, and often cited, benefit of edge deployments are the rapid response times that would not be possible if data is sent back to a centralised network for processing.

    Taiwan’s Taoyuan City, for instance, turned to edge technology in rolling out smart streetlights in its Qingpu district, using HPE’s Edgeline EL10 Internet of Things (IoT) Gateway.

    The Taiwanese city has ambitions of becoming a smart city and is looking to deploy and integrate multi-sensor information from edge products into a centralised platform to deliver better citizen services.

    “Certain citizen intelligence applications and services require an almost immediate response time [and] this cannot be achieved if data needs to be transmitted back to a centralised cloud for processing,” a spokesperson for Taoyuan City Government’s Public Works Department told ZDNet.

    To address customer concerns about outdoor or physical attributes, vendors such as HPE have designed their products to withstand various external factors such as dirt, humidity, temperatures, and vibration.

    When asked about the initial concerns that the Taoyuan government may experience when deploying the edge technology, the spokesperson pointed to the need to closely monitor such systems.

    “Intelligent edge solutions typically require massive data processing and network connectivity. Hence, ensuring regular system updates as well as stability of the various decentralised devices is critical,” she said.

    “Furthermore, as citizens increasingly rely more on such services, we need to ensure the data collected from multiple sensor devices is stored properly and securely.”

    Tan noted that HPE’s edge systems supported unmodified enterprise software from its partner community, including Citrix, SAP, GE Digital, and Microsoft. This meant that enterprise customers could use the same application stacks at the edge, in datacentres, as well as cloud.

    Key considerations before going to the edge

    Taoyuan City’s streetlight management edge deployment is still currently in its pilot phase and the government has plans to deploy more streetlights over the next few phrases of the project, according to the spokesperson.

    She noted that the city government is hoping to introduce more innovative services by analysing the data collected in the deployment, spanning parameters such as air quality, climate indicators, and image analysis processing.

    In deciding the volume and type of data that should and should not be analysed at the edge, she said the Taoyuan government assessed the network transmission bandwidth of the field device as well as the data management centre.

    It also considered the immediacy of the application service, whether it required real-time processing and feedback, and whether edge computing could support the required speed and security, she noted.

    She added that, compared to traditional datacentres, outdoor environments are harsher and edge deployments in such situations would need to consider factors such as weather, dust conditions, temperature as well as stability of power supply to the device.

  5. Tomi Engdahl says:

    Cadence’s Paul McLellan shares highlights from five presentations all discussing what’s behind AI’s movement to edge devices, the vast amount of investment going into the area, and where a few of the foreseeable challenges lie.

    Bagels and Brains: SEMI’s Artificial Intelligence Breakfast

  6. Tomi Engdahl says:

    5 Reasons Why Azure IoT Edge Is Industry’s Most Promising Edge Computing Platform

    Out of the top 5 public cloud platforms – AWS, Azure, Google Cloud Platform, IBM Cloud and Alibaba Cloud – only Microsoft and Amazon have a sophisticated edge computing strategy. Other players are yet to figure out their story for edge computing.

    Amazon’s edge platform is delivered through AWS Greengrass – a service that was announced at re:Invent event in 2016 and became generally available in June 2017. AWS recently added the ability to perform inferencing of machine learning models. It also started bundling AWS Greengrass in devices such as AWS DeepLens, a smart camera that can run neural nets at the edge.

    Microsoft shipped Azure IoT Edge almost after a year of AWS Greengrass’ general availability. However, the wait has been absolutely worthwhile. Firstly, the market dynamics have evolved in the last year giving the team an opportunity to align with customer scenarios. Secondly, Microsoft got a chance to improvise its platform to make it better than the only other offering – AWS Greengrass.

    1. Open sourcing the platform
    2. Containers at the core
    3. Ecosystem engagement
    4. Security
    5. AI @ Edge

    Azure IoT Edge plays a crucial role in Microsoft’s vision of delivering Intelligent Cloud and Intelligent Edge. Some of the design decisions such as containerized modules, tight integration with HSM, plugins for Visual Studio turn Azure IoT Edge into one of the most comprehensive edge computing platforms in the industry.

  7. Tomi Engdahl says:

    Japan Sniffing Out Its AI Niches

    As with any trade show featuring “embedded technology” anywhere in the world, the Embedded Technology 2018 Exhibition in Yokohama earlier this month got hijacked by today’s two hot topics: AI and IoT.

    On one hand, Japanese electronics heavyweights — mostly Fujitsu, NEC and Toshiba — showcased new materials and wireless technologies they deem critical to the spread of IoT applications.

    On the other hand, this year’s Embedded Technology/IoT show trotted out a host of Japanese startups, including Ascent Robotics, LeapMind, Robit and others with an intense business and technology focus on AI.

    Japanese startups tend to differ from startups elsewhere in their commitment to leverage Japan’s decades of experience in building robots and automobiles. They want to use their proximity to automated manufacturing sites and to experienced factory managers as a head start toward developing AI algorithms for industrial applications.

    While Google, Facebook, Amazon and others in the United States may have already established a stronghold in areas like big data, data centers and deep learning, Japan’s hopes focus on making edge devices smarter, more connected and autonomous.

  8. Tomi Engdahl says:

    Imec, CEA-Leti Form AI and Quantum Computing Hub

    Two of Europe’s key electronics and nanotechnologies research institutes — imec in Belgium and CEA-Leti in France — will collaborate to develop a European hub for artificial intelligence and quantum computing.

    As security and privacy issues rise up the agenda in almost every organization, the race is on to process more at the edge and put more intelligence at endpoints. For electronics systems design, most of the major chip companies now offer or are developing deep learning and edge AI devices or intellectual property. The edge AI devices are often complete computer sub-systems displaying intelligent behavior locally on the hardware devices (chips), analyzing their environment and taking required actions to achieve specific goals.

    Edge AI is considered now to hold the promise of solving many societal challenges — from treating diseases that cannot yet be cured today, to minimizing the environmental impact of farming. Decentralization from the cloud to the edge is a key challenge of AI technologies applied to large heterogeneous systems. This requires innovation in the components industry with powerful, energy-guzzling processors.

  9. Tomi Engdahl says:

    Advanced algorithms for analytics on the edge

    Substantial computing power in modern industrial PCs and cloud bandwidth considerations make the case to analyze machine performance directly on controllers, before the cloud.

    The debate between cloud and edge computing strategies remains a point of contention for many control engineers. However, most agree smart factories in an Industrie 4.0 context must efficiently collect, visualize, and analyze data from machines and production lines to enhance equipment performance and production processes. Advanced analytics algorithms allow companies to sift through this mass of information, or Big Data, to identify areas for improvement.

    To some, edge computing devices seems to create an unnecessary step when all data can be managed in the cloud with limitless space. Messaging queuing telemetry transport (MQTT) encryption and data security built into the OPC Unified Architecture (OPC UA) specification ensures all data will remain secure while it’s being transferred. When it comes to analytics and data management, however, edge computing presents important advantages to monitor equipment health and maximize production uptime.

    Because of the massive amount of data that modern machines can produce, bandwidth can limit cloud computing or push costs outside of a set budget. New analytics software strategies for PC-based controllers allow controls engineers to leverage advanced algorithms locally in addition to data pre-processing and compression. As a result, a key advance in analytical information is the concept to process data on the edge first, which enables individual machines and lines to identify inefficiencies on their own, and make improvements before using the cloud for further analysis across the enterprise.

    asset condition monitoring in intralogistics, used 20 sensors at a 1,000 hertz sampling rate and required 11.4 Mbps JSON. This a relevant test since JSON is a common format to send data to the cloud or across the web.

    Without compressing or pre-processing mechanisms, an average 7.2 Mbps internet connection cannot stream data from three or more large machines or from a full logistics operation

    Edge devices and advanced algorithms

    In the past, most programmable logic controllers (PLCs) were capable of controlling repetitive tasks in machines, but possessed the computing prowess of a smart toaster. Industrial PCs (IPCs) feature ample storage and powerhouse processors, with four, or as many as 36 cores. The automation software packages for these IPCs run alongside Windows, and can support third-party applications and can be accessed remotely. PC-based control software can provide advanced algorithms to manage data, such as pre-processing, compression, measurement, and condition monitoring. This does not require a separate, standalone software platform.

    More extensive machine vibration evaluations are possible using DIN ISO 10816-3: Mechanical Vibration-Evaluation of machine vibration by measurements on non-rotating parts. To monitor bearing life and other specific components, algorithms are available to add to a PLC program to calculate the envelope spectrum first and then the power spectrum.

    Implementing a cloud and edge strategy

    Running advanced algorithms on a local edge device reduces cloud bandwidth requirements and offers an efficient strategy for process optimization. However, that does not mean an operation can or should disconnect from the cloud.

    To decide what needs to be sent to the cloud and what can be processed or pre-processed locally, make sure to ask the following questions:

    What are the goals your operation wants to achieve through data acquisition in this instance?
    Which data sets from which machines need to be analyzed in order to achieve these goals?
    What types of data insights does the operation need to improve efficiency and profitability?

    Local monitoring with edge computing often works most efficiently to improve the operation of individual machines. However, the cloud provides the best platform to compare separate machines, production lines or manufacturing sites against each other. Implementing both allows an operation to maximize its capabilities.

  10. Tomi Engdahl says:

    AI Chip Architectures Race To The Edge

    Companies battle it out to get artificial intelligence to the edge using various chip architectures as their weapons of choice.

  11. Tomi Engdahl says:

    The Importance of Edge Computing in Industrial Transformation

    Based on a recent LNS survey, over 60% of participating companies have now instituted an Industrial Transformation initiative. To support these initiatives, many industrial companies are rethinking operational architectures and taking a hybrid approach to compute infrastructure; with a combination of traditional on-premise, cloud, and edge.

    However, many companies aren’t necessarily realizing the anticipated benefits from Industrial Transformation and there is still confusion regarding how to prioritize these investments.

  12. Tomi Engdahl says:

    Living On The Wireless Edge With AI And 5G

    We are at the cusp of something truly transformational, driven by two much-hyped yet groundbreaking technology mega trends: artificial intelligence (AI) and 5G. Together they make possible things that never existed and seemed utopian not too long ago. A vivid example of this is the rise of self-driving vehicles. Notwithstanding the recent setbacks in the first iteration of self-driving, a fully autonomous self-driving vehicle will be the epitome of AI and 5G technology.

    If you apply the traditional AI approach to self-driving, this intelligence will reside in a centralized cloud. The data collected from the vehicle will be hauled to the cloud for processing, and instructions will be sent back to the vehicle. However, when you consider a moving vehicle in which decisions have to be made in split seconds, this approach simply won’t work.

  13. Tomi Engdahl says:

    Intelligent Connectivity: the Fusion of 5G, AI and IoT

    Intelligent connectivity is the combination of high-speed, low-latency 5G networks, cutting-edge artificial intelligence (AI) and the linking of billions of devices through the Internet of Things (IoT). As these three revolutionary technologies combine they will enable transformational new capabilities in transport, entertainment, industry and public services, and much more beyond.

  14. Tomi Engdahl says:

    Vision AI developer kit combines AI and ML to push deep neural network models out to the intelligent edge

    Companies are in the process of digitally transforming their business by using artificial intelligence (AI) and machine learning (ML). Currently this task is only possible once data is collected from internet-connected devices and stored in the cloud.

    The technology challenge with this approach is the strong dependency on a consistent connection to the cloud for sending and collecting data. As data volumes approach larger scales and deep learning requires increasingly more complex algorithms, the inevitable bottleneck will limit the quick adoption of AI and ML technologies.

    Currently, AI computations are feasible when vast, potentially bordering on infinite, amounts of computing resources are available from cloud resources. This requires an investment in expensive and powerful computational machines running at the edge. This requires continuous power supplies and direct connectivity to all sensor devices.

  15. Tomi Engdahl says:

    Giving a flexible edge to the IoT

    How flexible sensors will revolutionize electronics as we know them.

    As the Internet of Things (IoT) continues to revolutionize our daily lives, the demand for smaller, smarter, and more diverse flexible technology has never been greater. Increasingly complex demands have driven the development of smart sensors to monitor everything from velocity and proximity to pressure, humidity, and more. Future devices will need to interact with the ambient environment by performing intelligent activities such as fingerprint, vein, and odor recognition, with sensors so small and flexible that they can be integrated into almost anything.

    Giving a flexible edge to the IoT

    How can you make ML work on plastic?

    The consortium believes that customized processing engines such as neural networks (NNs) are the key to accelerating development of low-cost and customized flexible, integrated smart systems. Customized for a specific application and capable of operating in extremely parallel fashion to achieve high performance, and consume low power, this will be the first time that a flexible smart device has been created to take advantage of machine learning algorithms in hardware.


Leave a Comment

Your email address will not be published. Required fields are marked *