Leveraging the cloud—and knowing when not to

When businesses tally up their expenditures for 2017, they are likely to find they’ve been spending more on cloud computing, according to B2B ratings and reviews firm Clutch. Looking ahead, Gartner has identified the movement of computational power from the “cloud to the edge” as a key trend for 2018.

Several companies have recently introduced products that facilitate cloud computing, edge computing, or both. Moxa, which offers industrial networking products, has introduced cloud-to-edge gateway hardware and software for IIoT applications. CEVA, a licensor of signal-processing IP for smart connected devices, and Brodmann17, a developer of deep-learning technology, have announced a partnership aimed at accelerating computer-vision processing in edge devices serving consumer and other mainstream applications.

Test and measurement companies have been addressing cloud and edge technology as well. National Instruments has introduced its Data Management Software Suite, which can perform nanosecond analytics at the edge as well as across the enterprise. Addressing wireless network operators, Keysight Technologies has launched a cloud-based automated reporting service for data gathered using its Nemo measurement tools.

Businesses boost cloud spending

According to Clutch, nearly 70% of U.S. businesses surveyed early last year said they planned to increase spending on cloud computing in 2017. One in five of those businesses reported that their cloud-computing spending for 2017 would likely increase by more than 30%, according to a report on the survey results.¹

The report says companies have become less skeptical of cloud computing and more confident in its security. In fact, the survey found that the largest percentage of businesses identify security as a benefit of using the cloud, in contrast to previous years, when cloud security was often treated with skepticism and distrust.

“Cloud is the new normal,” said Jeremy Przygode, CEO of Stratalux Inc., a California-based managed service provider (MSP), as quoted in the Clutch report. “When businesses need to evaluate new solutions, or need to do a hardware refresh on existing solutions…cloud is the go-to solution to figure out how to do that.”

The survey found that most businesses are using a private cloud; however, more than 80% indicate that they are considering a hybrid cloud option, with services and infrastructure spread between a private network and off-site cloud provider.

Once the decision is made to adopt cloud computing, many businesses seek outside help for the subsequent installation. Over half of businesses surveyed (57%) say they hire an external consulting firm to help them implement their cloud strategy.

‘Cloud to the Edge’ trend

However, for some applications the cloud is decidedly inappropriate. For example, you don’t want the cloud between your autonomous vehicle’s sensors and brakes, even if the cloud is accessible via a highly reliable, low-latency 5G connection. Even when safety is not a concern, computing and data storage might best be done at or close to the edge if bandwidth constraints and connectivity challenges add to latency and reliability issues.

Such challenges might be contributing to a trend Gartner described at its Gartner Symposium/ITxpo 2017 in October in Orlando. The firm provided no figures on spending or on companies moving to or from the cloud, but it did cite “Cloud to the Edge” as one of the top 10 strategic technology trends for 2018.² The firm describes edge computing as a topology in which information collection, processing, and delivery reside near the sources of information. The firm recommends that enterprises—particularly those with significant IoT elements—should begin using edge design patterns in their infrastructure architectures.

Gartner does not see cloud and edge as competing approaches, describing cloud not as inherently centralized but rather as a style of elastically scalable computing capabilities delivered as a service.

“When used as complementary concepts, cloud can be the style of computing used to create a service-oriented model and a centralized control and coordination structure with edge being used as a delivery style allowing for disconnected or distributed process execution of aspects of the cloud service,” said David Cearley, vice president and Gartner fellow, in a press release.

As a related top 10 trend for 2018, Gartner cited “Event Driven,” which describes a business’s ability to sense and exploit digital business events—such as the completion of a purchase order or an aircraft landing. The IoT, cloud computing, blockchain, in-memory data management, and AI all can allow business events to be detected faster and analyzed in greater detail, the firm said.

Meeting OT and IT requirements

Note that Clutch surveyed IT professionals, who were also the target of Gartner’s symposium. Your interests and responsibilities are more likely to fall within the sphere of operational technology (OT), which Gartner’s IT Glossary defines as “…hardware and software that detects or causes a change through the direct monitoring and/or control of physical devices, processes, and events in the enterprise.”

Pending the completion of an anticipated IT/OT convergence, IT and OT teams may be at cross purposes, or at least fail to communicate clearly.

Moxa, which provides products for industrial networking, computing, and automation, has described the OPC Foundation’s OPC UA (Unified Architecture) M2M protocol as a method of bridging the OT and IT worlds. OPC UA is is critical for the IIoT, the company said. Originally, OPC UA worked on a client/server model, but dealing with hundreds or thousands of devices that all need to be interconnected across multiple sites required a more scalable solution. The adoption of the publisher/subscriber model allows for more streamlined communication that offers improved scalability and resilience, the company said.

To take advantage of this publisher/subscriber model, Moxa announced in September that it has partnered with Microsoft and the OPC Foundation to develop the MC-1121, an industrial-grade IoT gateway with an integrated OPC UA Publisher module. By using Windows 10 IoT and OPC UA Publisher, the gateway provides a way to get data from field-side devices securely and reliably to the cloud for analytics and monitoring through a dashboard.

Moxa said its MC-1121 Series IIoT gateways make it easy to get your data into the Microsoft Azure IoT Hub. These IIoT gateways include a variety of interfaces to connect to Ethernet, serial, and I/O devices, and they can be used in conjunction with Microsoft Connected Device Studio.

“The Moxa gateways leveraging OPC UA technology provide the key enabling solution necessary to address information integration between control systems across all industries and the IT world,” said Thomas J Burke, OPC Foundation president, in a press release. “End users are demanding secure and reliable information integration connectivity, and Moxa’s products truly will facilitate a complete solution connecting the edge with the cloud.”

Bee Lee, Moxa’s president of global sales and marketing, added, “The Industrial Internet of Things is a paradigm shift, providing businesses with new capabilities such as dashboards that show device status and data in real-time [to enable] on-demand production of customized products. This requires network infrastructure to do much more than before in order to support IIoT applications; it must be even more reliable, more secure, and capable of supporting real-time operations. Moxa’s commitment to versatile, reliable edge-to-cloud connectivity solutions allows us to provide the connectivity necessary to bridge devices to Microsoft’s Azure IoT Hub by using OPC UA protocols, helping cloud and edge work together to help the IIoT reach its full potential.”

IIoT Gateway Starter Kit

Moxa followed up its OPC announcement with the October launch of its IIoT Gateway Starter Kit (Figure 1), which includes built-in support for Amazon Web Services (AWS). The kit contains two main components: ThingsPro Gateway, a ready-to-run data-acquisition software platform that simplifies the complex task of transferring edge data to the cloud, and the UC-8112 edge computer, an industrial-grade ready-to-deploy communication-centric computing platform that packs a 1-GHz ARM processor, 512 MB of RAM, two LAN ports, and two serial ports in a palm-sized rugged box.

Figure 1. IIoT Gateway Starter Kit with ThingsPro Gateway and UC-8112 edge computer
Courtesy of Moxa

To simplify getting your data, ThingsPro Gateway provides a Modbus framework to connect with Modbus RTU/TCP devices and SCADA systems. It also includes network support for 4G connectivity, wireless failover, firewalls, and a VPN to ensure that your data can be securely retrieved from remote field sites.

To get your data into the cloud, it has built-in client support for services such as AWS IoT and Cirrus Link Sparkplug. By integrating the AWS IoT Device SDK, ThingsPro Gateway lets you set up tags and devices on AWS IoT. You can then transfer field data to various AWS cloud services—such as Amazon Kinesis, AWS Lambda, and Amazon S3—to collect, process, and store data. The starter kit also supports running AWS Greengrass to perform computing, messaging, data caching, and sync functions locally. Furthermore, with the built-in Cirrus Link Sparkplug SDK, you can connect your IIoT Gateway to Inductive Automation’s Ignition Platform or another MQTT server.³

“The value of IIoT gateways is in connecting edge devices and taking the necessary data to the cloud,” stated Johnny Fang, a product manager in Moxa’s Embedded Computing Division, in a press release. “To enable faster integration between things in the field and services on the cloud and to help users to get their solutions to market sooner, we’re actively working to add support for more and more cloud services such as Microsoft Azure, Google IoT Core, and Schneider Wonderware in the coming releases.”

Finally, for those who need more customized solutions, Things-Pro Gateway includes C and Python APIs to accelerate your application development and create a tailored solution that meets your specific needs.

The UC-8112 edge computer provides a flexible platform suitable for a variety of applications and is available in a wide-temperature-range model that can operate in harsh environments from to -40 to 75°C. It also has a range of wireless accessories available to add LTE or Wi-Fi connectivity. If you are developing a proof of concept with a Raspberry Pi, you can migrate your IIoT design to the UC-8112 to make it suitable for industrial application and mass production.

Combining the UC-8112 edge computer hardware and Things-Pro Gateway software, the IIoT Gateway Starter Kit is suitable for remote-monitoring, data-acquisition, and data-processing applications for a variety of IIoT applications, including ones related to solar energy, wind power, electric-vehicle charging, water and wastewater monitoring, and smart manufacturing, the company said.

AI at the edge

Whereas Moxa’s recent news focuses on edge-to-cloud interfacing in IIoT applications, CEVA and Brodmann17 have announced a partnership aimed at accelerating the deployment of deep-learning computer vision in edge devices serving mainstream applications. Through the partnership, CEVA and Brodmann17 said they would bring an order-of-magnitude increase in performance and power efficiency for deep learning in edge devices compared with GPU-based implementations.

The push towards widespread adoption of AI in consumer devices continues at a relentless pace, the companies said. However, cloud-based deep learning on battery-powered devices is plagued with issues, including latency, security, and the need for a constant, reliable Internet connection, the companies said, adding that implementing the intelligence on the device itself—or on the edge—eliminates all of these issues. However, highly efficient computer-vision processors are necessary to meet the stringent power requirements, and specialized deep-learning software is crucial in delivering the accuracy and performance.

Targeting embedded devices, Brodmann17 has developed a specialized deep-learning technology for visual recognition aimed at edge-based artificial intelligence. Using patent-pending techniques, Brodmann17’s deep-learning architecture generates smaller neural networks that are faster and more accurate than other networks on the market, the company said. Through CEVA’s collaboration with Brodmann17, licensees of the CEVA-XM platforms and their customers will be able to use Brodmann17’s deep-learning object detection, which achieves state-of-the-art accuracy on the CEVA-XM at a rate of 100 frames per second.

“Our patent-pending deep-learning vision software is a perfect fit for the many CEVA customers and OEMs using CEVA-XM platforms to add intelligence to their devices,” said Adi Pinhas, CEO of Brodmann17, in a press release. “This first-of-its-kind combination of hardware and software achieves real-time performance that supports multicameras with a single DSP or higher resolutions.”

“To truly maximize the performance and capabilities of AI, in mass-market devices, it requires not just application-specific hardware like our CEVA-XM platforms, but also neural networks that are trained to be run efficiently on the edge-embedded devices,” added Ilan Yona, vice president of the vision business unit at CEVA. “Brodmann17’s deep-learning software provides the capability to create extremely light, accurate, and flexible networks, trained from the ground up with embedded in mind.”

From edge to enterprise

National Instruments is addressing edge and enterprise computing with the October release of its Data Management Software Suite, which offers a complete workflow to standardize measurement data across teams. Users can mine that data for useful information, transform the data through automated analysis, and deliver reports with valuable insight.

“Using the Data Management Software Suite, we’ve helped a major automotive manufacturer reduce the time it spends analyzing the data generated by one of its component tests from 10 hours to 7 minutes,” said Barry Hutt, CEO of Viviota, an NI Alliance Partner specializing in data management, in a press release.

Dave Wilson, vice president of platform software at NI, commented, “The amount of data being acquired to test devices, monitor physical assets, and analyze product designs continues to skyrocket. The challenge with the exponential growth in the amount of data being acquired is the establishment of a repeatable and automated process to extract valuable insights. Often, inconsistencies and errors in the data produce erroneous results.” He added that engineers must manually inspect and verify data before sending it to a manual or automated analysis process.

The Data Management Software Suite introduces new server-based software features and a new product, the Analysis Server, to simplify the workflow. The suite helps engineers and scientists automate the search, standardization, analysis, and reporting of large amounts of measurement data. The full suite is flexible enough to integrate with customers’ existing data formats and IT infrastructures, so any team with a Windows machine and a network can add data-management capabilities.

NI reports the platform provides the ability to read/write data for analysis, to standardize metadata for automated analysis, and to perform nanosecond analytics at the edge. Daniel Parrot, a subject matter expert at NI, explained by email that the solution extends from a system at the edge to the group level with multiple systems and on to the enterprise level, working across multiple groups in an organization. “The primary use case for the Data Management Software Suite is the automated post-processing of measurement data at the group and enterprise level,” he said. “This data is most typically stored on the enterprise’s network and not necessarily in the cloud.”

NI said that with the release of the Analysis Server, NI’s server-based data-management solution built on DIAdem and the DataFinder Server, the suite can be expanded to automated data processing, allowing engineers to gain initial insights without any manual interaction while preserving the original measurement data for searching and further investigation.

Cloud reporting for wireless networks

Finally, Keysight Technologies is applying cloud technology to telecommunications applications, helping customers optimize their wireless networks. In September, the company announced the launch of a cloud-based reporting service available for data measured with Keysight’s Nemo measurement solutions. The reporting service provides an automated workflow for acceptance, verification, benchmarking, in-building, and drive-test measurements. The reporting service shortens the turnaround time from data collection to reports and provides a single point of control. Figure 2 shows a snapshot of a report of in-building measurements showing coverage and wireless technology, delivered automatically as a service for data measured with Nemo measurement solutions.

Figure 2. Snapshot of a report of in-building measurements showing coverage and serving wireless technology, delivered automatically as a service available for data measured with Nemo measurement solutions.
Courtesy of Keysight Technologies

Keysight said cloud-based reporting tools have become essential for operators, network equipment providers, and service companies as they look to drive down costs and increase productivity. Keysight’s reporting service for Nemo products provides a way to generate in-depth reports from drive-test and in-building data with insight into the customer experience and network performance. The log-file transfers to a back office, followed by log-file processing and report execution, are fully automated. Users can receive and share in-depth reports throughout the organization swiftly after finalizing a measurement campaign.

“The reporting service is ideal when there is little variation in the reports used or you have a measurement project where seamless and quick access to reporting is one of the key requirements,” said Mikko Hyvärinen, director, CEM and platform products at Keysight, in a press release. “Our end-to-end measurement solution, from Nemo-branded data-collection tools to our remote monitoring and post-processing tools, ensures a smooth integrated information flow speeding up the whole measurement analysis process and enabling more efficient operations, ultimately, giving customers the possibility to save time and reduce costs.”

References