At a time when businesses expect greater value - more quickly - from IoT, you can deliver rapid results using the new features and enhancements in Cumulocity IoT 10.7.
Register for our webcast (18 Mar 2021) to see the demo and hear details from product managers on the new release. Until you can tune in, here’s a quick look at what you can do better and faster with Cumulocity IoT 10.7.
Machine learning enables you to use self-learning computer algorithms for automatic improvements based on actual experiences. In our latest release, you can build machine learning models directly within Cumulocity IoT for greater efficiency. Our machine learning workbench is designed so non-data scientists (i.e., anyone) can put machine learning models to work in one quick click.
With the new machine learning workbench, you can:
- Deliver use cases faster using out-of-the-box models
- Create machine learning models by seamlessly accessing, ingesting and transforming IoT data in operational stores and data lakes
- Automate your model selection
- Build models using graphical tools, such as Neural Network Editor and an embedded Jupyter Notebook capability
- Deploy models in one click into production in the cloud, at the edge or on-premises
Use machine learning to improve how your business operates—for example, to automate defects tracking on the assembly line, track the activity of assets in the field, forecast consumption and demand patterns, and prevent equipment downtime by predicting when equipment is likely to fail.
Machine learning enhancements in 10.7 include:
- A workbench for model training and one-click model deployment to the inference engine
- Updated UI and API to improve the representation of input/output schema for ONNX models
- NYOKA microservice, which adds a clustering API to support unsupervised learning on time-series data
- Predictive analytics (Zementis) support for ONNX standards-based model deployment and management, including pre/post-processing resources and pipelines
- Support for multivariate state space models representation and computation of confidence intervals, multivalued output for multistep forecasts, and confidence intervals
New Machine Learning Workbench
The number of devices you need to manage is likely to grow. That’s why Cumulocity IoT 10.7 improves the efficiency of IoT device management for growing numbers of devices. Count on Cumulocity 10.7 to help you:
- Configure groups of devices with one click
- Apply changes for software, firmware and configuration simultaneously using device profiles
- Roll out the same change to many devices with the easy-to-use bulk operations wizard
- Authenticate IoT devices and manage trusted root certificates from within the Device Management UI using device certificates
- Use best-practice device certificates to meet security requirements
Device management & connectivity enhancements include:
- Enhanced performance, specifically in MQTT connectivity and measurement queries
- Bulk operations so it’s easier to manage a large portfolio of devices—the new device profile lets you apply a target state to a group of devices
- Improved documentation for LWM2M and OPC UA protocol support
- Support for device certificates
Bulk device operations
With Analytics Builder, you can create models, code-free, using the drag-and-drop interface. You can now reuse those same models with parameter adjustments. That means you won’t have to create a new model each time you want to analyze different measurements—such as temperature or pressure—on different devices. This time-saver enables you to deploy analytics much faster and makes it easier to manage and maintain similar models.
New streaming analytics enhancements include:
- Template models in Analytics Builder
- Improved scalability in Analytics Builder for models with multiple input devices
- Branding support in Analytics Builder
Values in Analytics Builder blocks can be fixed or taken from template instances
Cumulocity IoT DataHub is now available in the public Cumulocity IoT cloud—beyond private cloud, on-premises and edge environments. That means more people, in fact, any Cumulocity IoT customer, can offload operational data to a cost-efficient data lake (including Azure® Data Lake Storage Gen 2 – ADLS2 and Amazon® Web Services Simple Storage Service – AWS S3), and then use their favorite BI and machine learning tools to analyze the data to see trends and gain business insights. In addition, self-hosted businesses using Cumulocity IoT DataHub can now use Cloudera® or Hortonworks® Hadoop® Distributed File System (HDFS) as a data lake.
New DataHub enhancements include:
- Availability on the cumulocity.com public cloud
- New UI that shows usage statistics on the amount of data being processed
- Support for Cloudera® and Hortonworks® HDFS (Hadoop Distributed File System) as another data lake option for storing offloaded data for self-hosted customers
- Import/export offloading configurations
Cumulocity IoT 10.7 makes it easier to add IoT integration to your solution so you can integrate IoT data with business systems and applications. New webMethods.io templates for Cumulocity IoT speed up setting up alert notifications with some of the most popular communication applications, such as Vonage®, Twilio, Zendesk®, ServiceNow®, Twitter®, Slack®, and Trello®. You can use these templates, for example, to quickly receive a text or email through your usual communications methods when a device triggers an alarm.
New integration recipes available from webMethods.io
With IoT edge computing, you can analyze and aggregate data close to edge devices and minimize data sent to the cloud. You can also act on local events even when cloud connectivity is down. To help you reduce interruption of edge devices, Cumulocity IoT 10.7 enhancements enable you to:
- Ensure enough disk space is available on edge devices remotely using a Cumulocity IoT cloud tenant—create alarms and notifications, so you know when disk space is close to full
- Modify and reconfigure edge networking and operating system-level properties, such as the Docker subnet, the edge hostname and default configuration of the data and binary disk—ideal for power users
Platform enhancements make Cumulocity IoT even more performant and secure:
- Look for improved data retention in high-throughput scenarios on the Data Broker—our internal architecture tolerates a larger backlog of data to reduce the risk of data loss when a tenant receiving data can’t keep up with the one sending it
- Microservice network isolation ensures all interactions between microservices use the authentication and authorization that has been configured for that tenant
- New Security-Enhanced Linux makes it more difficult for those with malicious intent to hack into the core platform
It’s now even easier to create IoT solutions and develop applications using Cumulocity IoT 10.7. Everyone, including business users and OT teams, can quickly get up to speed, thanks to improved performance in the dashboard UI and more helpful guidance. Make selections quickly using updated UI widgets, alarms, and events lists. The web SDK simplifies customizing dashboards and the Cockpit app with your own components and applications.
The demand for building automation via IoT is growing fast, especially among businesses interested in reducing energy consumption. With the Cumulocity IoT 10.7 release comes a new IoT Solution Accelerator for Building Management. This out-of-the-box Cumulocity IoT application provides pre-packaged building management widgets, dashboards, and modules to support heating and air conditioning (HVAC) systems and energy monitoring.
Connect devices, configure the application through our no-code UI process and begin using the solution for insights, such as average energy expenditure and mapping device locations within buildings. With the accelerator, you can:
- Stay ahead of energy regulations and green initiatives
- Measure and generate insights into energy and heating expenditure across multiple buildings
- Customize the solution for your own customers
To see when the new release will be available across the regions visit http://status.cumulocity.com/