IoT Platforms: Powering Change

Smart Business

IoT Platforms: Powering Change

Data originates all around us and it lives in many places, too. Gathering it and linking it to analytics, storage, and a huge range of specialized apps is a role now being filled by IoT platforms. From the largest hyperscale cloud provider to narrowly focused niche players, it seems everyone has a slice of the pie.

by Gordon and Eamon McCarthy Earls

IoT platforms: Size matters – the big four

For enterprise IoT projects, cloud-based IoT platforms from the leading public cloud providers are increasingly central to new initiatives. Recent results from the 451 Research survey Voice of the Enterprise – Internet of Things, Vendor Evaluations 2018 found that 59 percent of enterprise respondents were using a commercial IoT platform. Not surprisingly, in terms of survey respondents reporting a platform in use, Microsoft Azure and IBM Cloud were neck and neck with 35 percent and 34 percent adoption respectively. Meanwhile, Google Cloud and Amazon Web Services (AWS) were similarly paired for third and fourth ranking, with 27 percent and 26 percent of respondents using their IoT platform.

Indeed, 451 Research found that customers seem to be deciding on hyperscale IoT platform providers primarily based on factors such as technical expertise, long-term viability, and total cost of ownership – but other factors like strategic vision, partner networks, or customer-service capabilities were also influential.

The Big Four are starting to differentiate by adding more AI and edge capabilities.

Frank Antonysamy, Cognizant

IoT platforms: Frank Antonysamy - Cognizant

 

“All of these platforms provide hyperscale for enterprise-grade IoT solutions and IoT/PaaS services,” says Frank Antonysamy, global markets head for connected products at consulting firm Cognizant. He adds that all of the so-called Big Four platform vendors continue to enrich their core IoT services but are gradually starting to achieve differentiation through artificial intelligence and edge capabilities. Antonysamy explains that most of the IoT offerings provided by the Big Four are essentially platform as a service (PaaS) solutions specific to data ingestion, device management, streaming analytics, and the data pipeline.


/// Read: IoT Platforms Part 2 – SMALL(ER) MAY BE BEAUTIFUL ///


They all typically function satisfactorily with the rest of each vendors’ PaaS and infrastructure as a service (IaaS) cloud platforms that provide product related compute, storage, analytics, AI, integration services, and such. “All of these services are needed to realize an enterprise IoT solution, and IoT service offerings are just another set of their PaaS services,” says Antonysamy . However, integration of IoT offerings with some vendor-specific cloud offerings vary – Azure integrated with Dynamics 365, for example, is at a relatively nascent stage.

Microsoft Azure

In contrast to some of its competitors, Azure has focused sharply on enterprise use cases for IoT, highlighting major customers such as ThyssenKrupp, Johnson Controls, and Rolls-Royce. Its IoT Central monitoring and management solution lets users, which it calls “builders,” create device templates with web-based tools to define what kind of telemetry is sent by a device and specifying its behavior and thresholds.

“Microsoft Azure has actively built partnerships with smaller platforms that they compete with – examples of this include PTC ThingWorx, and SAP Leonardo,” says Antonysamy. In addition, he notes that Azure also has the most industry-specific approach, with customizations built for different industries. “Google and AWS seem to focus more on the platform providing the scale and reliability, leaving their partners to build any industry-specific applications,” he says.

Following its acquisition of GitHub in 2018, Microsoft hosts all of its IoT software development kits on the site, with support for Java, .NET, Node.js, Python, and C (written in ANSI C for portability). If problems crop up, users are invited to create help desk tickets for Microsoft Support through the Azure portal. IBM Cloud – Of the Big Four cloud providers, IBM may have the smallest They all typically function satisfactorily with the rest of each vendors’ PaaS and infrastructure as a service (IaaS) cloud platforms that provide product related compute, storage, analytics, AI, integration services, and such. “All of these services are needed to realize an enterprise IoT solution, and IoT service offerings are just another set of their PaaS services,” says Antonysamy . However, integration of IoT offerings with some vendor-specific cloud offerings vary –Azure integrated with Dynamics 365, for example, is at a relatively nascent stage. Microsoft Azure – In contrast to some of its competitors, Azure has focused sharply on enterprise use cases for IoT, highlighting major customers such as ThyssenKrupp, Johnson Controls, and Rolls-Royce. Its IoT Central monitoring and management solution lets users, which it calls “builders,” create device templates with web-based tools to define what kind of telemetry is sent cloud market share but it is an important player in IoT and should not be underestimated. In parallel with AWS, IBM Cloud uses the lightweight message queuing telemetry transport (MQTT) protocol or HTTP and a mix of representational state transfer (Rest) APIs and real-time APIs.

“IBM is leveraging its deep analytics capabilities around Watson to provide pre-built applications like Condition Monitoring and Predictive Maintenance in the industrial manufacturing space,” Antonysamy says. A modified version known as the Watson IoT Platform is targeted at companies looking to deploy IoT devices. In November 2016, the Watson platform was extended for IoT security goals, with the addition of Risk and Security Overview Dashboard, geared toward whitelisting and blacklisting traffic, connection security maintenance, and certificated management.

Google Cloud

When it comes to IoT and generic PaaS services, Google Cloud is gaining interest and momentum, according to Antonysamy. He says its strategies are bringing in cloud data flow innovations with AI, machine learning (ML), and cloud native platforms, and scores after telemetry ingestion.

With platforms maturing, the trend is to offer more IoT-specific services such as platform managed services, including third-party offerings. “They also offer services and tools in the adjacent areas like AI by integrating AI-based data analytics tool sets into their platform as well as IoT edge,” he says. “This is a result of providing edge runtime frameworks that even support ML and serverless computing on the edge.” At the center of Google’s approach to IoT is its Cloud IoT Core, which draws in data from devices and pushes out commands, doing management with Edge Connect, Edge ML, and TensorFlow Lite for analytics and machine learning. Google integrates MQTT notifications with its own Cloud Pub/ Sub system, which it describes as a “globally durable message ingestion service.” Pub/Sub is intended to be a “shock absorber and rate leveler” for inbound data streams as well as app architecture changes, setting topics for different streams and allowing different parts of an app to subscribe to specific streams.

From the perspective of storage and analysis, Google offers both raw and processed data storage in Cloud Datastore and Firebase Realtime Database, with the option to run “ex tract, transform, and load,” streaming computation, and batch operations with its Apache Beam-powered Cloud Dataflow. The TensorFlow opensource machine learning framework present in IoT Core can be run in a managed training or distributed approach through Cloud Machine Learning Engine.

Amazon Web Services

From the time AWS launched in 2006, Amazon entered the market as a technology upstart and quickly became a leader in the field, in part by winning over many developers and small and medium businesses to the AWS platform. While the IoT platform market has become increasingly dominated by Microsoft Azure, AWS has managed to claim the most cloud usage overall.

AWS offers a broad portfolio to organizations trying to connect devices and services. AWS IoT Core is the central offering for IoT deployments, offering a messaging and connectivity system for devices, with a device-side software developers’ kit (SDK).

Amazon FreeRTOS uses the popular real-time operating system (RTOS) for microcontrollers, common across industry, teaching, and the hobbyist community to deploy diverse sets of microcontrollers. Amazon’s derivative version is updated regularly with SDKs and connectivity libraries, in a bid to maintain compatibility for extensions and connections to AWS. “Using these SDKs in combination with our IoT Core service, you can securely send information with the managed MQTT message service. We support mutual certificate-based TLS [transport layer security] authentication. [Data is] authenticated, authorized, and secured by encryption features in the certificate-based services,” says Ian Massingham, director of AWS evangelism.

To date, AWS has been applied to wide-ranging IoT scenarios like BMW’s smart service and automotive maintenance deployment for its 7 Series cars, or Italian energy company Enel’s smart metering. Data processing and analytics is a key element of Amazon’s strategy for IoT, ranging from its S3 simple storage system to its Redshift data warehouse and Kinesis high-volume data streaming, all integrated for IoT. Because IoT devices vary in their output, AWS Lambda aims to be a nearly serverless software creator that responds programmatically to actions taken by IoT devices. US manufacturer iRobot has used Lambda with many of its house cleaning robots.

We support TLS authentication, authorized and secured by encryption in certifcatebased services.

Ian Massingham, AWS

IoT platforms: Ian Massingham - AWS

 

Massingham gave an example of smart agriculture using AWS in which a polytunnel sensor breaches a temperature threshold. “You can receive messages from IoT devices with IoT Core Service and a rules engine can evaluate the content to make a routing decision … and publish [commands] to a vent, heater, or chiller, correcting a temperature by opening a window in the polytunnel or adding humidity with a sprayer.” He adds that to make sense of the alert, data can be routed into long term storage with Redshift and S3 and used to create visualizations in a dashboard apps.

Bigger Is Better – for Many

According to Antonysamy, all four platforms have their roots in their underlying cloud/PaaS platforms and they have evolved into IoT platforms by introducing core IoT services such as device connectivity and management, and streaming analytics, along with other “product-as-a-service” offerings for storage, compute, analytics, and enterprise integration. That heady combination of size and broad capability seems destined to convey an important, long-term advantage, he thinks. In fact, according to the 451 Research survey, those companies that have adopted IoT platforms from the Big Four are farther along in their digital transformations. Potentially, this indicates that the die has been cast and that these major providers are very much baked into long-term planning.


/// Read: IoT Platforms Part 2 – SMALL(ER) MAY BE BEAUTIFUL ///


Leave a Reply

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

*