Announcing key industrial IoT capabilities in Azure Time Series Insights

Posted on 4 December, 2018

Principal PM Manager, Azure IoT

Earlier this year, we announced new features to be added to Azure Time Series Insights by the end of the calendar year. Today, we proudly deliver on that promise and announce the public preview of features that will continue to empower our customers to achieve more with their IoT data. Specifically, today we launch the following capabilities, which are described in further detail below:

  • A scalable, performance- and cost-optimized, multi-layered time series data storage that enables a cloud-based IoT solution to trend years’ worth of time series data in seconds.
  • Semantic model support to describe the domain and metadata associated with the derived and non-derived signals from assets and devices.
  • Enhanced analytics user experience, Time Series Explorer, that combines asset-based data insights with rich, ad-hoc data analytics for business and operational intelligence.
  • Seamless integration with advanced machine learning and analytics tools like Databricks, Apache Spark, Jupyter notebooks, and Power BI to help customers tackle time series data challenges in new ways.

We also took critical strides in updating the pricing model for Time Series Insights and are excited to announce a new pay-as-you-go pricing model that provides customers a lower entry price point as well as separate levers for data processing, storage and query; offering the flexibility and scalability that IoT business demands.

Embracing the IoT journey with our customers

Since the general availability last November, customers have been using Time Series Insights to effectively address their IoT insights needs. Along the way, we have listened to our customers and learned from their IoT journey with our product.

Our customers span all major industrial IoT segments including manufacturing, automotive, oil & gas, power & utility, smart buildings, and consulting. Their scenarios involve data exploration for use cases where the shape of the data is not known, as well as operational analysis over schematized (explicitly modeled) data to drive operational efficiency. Platform capabilities like multi-layered storage (warm and cold) with the ability to store decades worth of time series data, and the ability to explicitly model and optimize queries for asset-based operational intelligence are becoming key to the success of large industrial IoT enterprises and their digital revolution.

To help maximize the value of time series data and thereby drive operational intelligence, Microsoft is updating Time Series Insights offering to support a broad range of industrial IoT data analytics scenarios by combing the current in-market interactive ad-hoc analytics capabilities with asset-based operational insights to enable customers to derive the highest value out of data collected from IoT assets.

Details of the new features in public preview

Scalable, performance- and cost-optimized time series data storage

Time Series Insights provides scalable multi-layered, warm (in-market) and cold time series data storage. Time Series Insights cold storage, now in public preview, is built on top of Azure Storage, which is a customer-owned storage account. Data is stored in open sourced Apache Parquet file format for efficient data compression, space, and query efficiency. This has an added advantage of allowing seamless connection to other data solutions such as Databricks, Azure Machine Learning, PowerBI, or other third-party services for advanced analytics and business scenarios. Data is uniquely identified with time series ID and time stamp properties. While customers own the data in their storage account, data partitioning is controlled by Time Series Insights platform to help drive efficient storage and queries as data is ingested. Data is ingested through Azure IoT Hub or Azure Event Hub, as is the case in the in-market solution. Additional ingestion sources will be supported over time.

Time series model for contextualizing raw telemetry and deriving asset-based insights

IoT data is highly unstructured, and only a negligible fraction of this data gets used for operational and business purposes to provide consistent, comprehensive, current, and correct information for business reporting and analysis. Turning IoT data into actionable insights requires among other things, a structure to navigate and understand the data. In this public preview, Time Series Insights provides support for time series model that helps contextualize raw telemetry data and makes it easy to find, curate, maintain, and enrich time series data. Semantically rich data is easy to find, curate, maintain, and enrich time series data. Semantically rich data is easy to query and navigate, thereby making computation and analysis over asset-centric data simple and highly valuable for operational analysis.

Time series model allows customers to model types (e.g., temperature sensor), hierarchies (e.g., property names and relationships) and instances (e.g., time series such as deviceID or assetID). Time series modeling enables customers to:

  • Author and manage computations, transform data leveraging scalar functions, and aggregate operations.
  • Define parent/child relationships to enable navigation and reference to provide context to time series telemetry.
  • Define properties associated with the instances and use these to create hierarchies.

Rich analytics user experience (Time Series Insights Explorer) for asset-based data insights and ad-hoc data analytics

Time Series Insights Explorer is now significantly enhanced to support authoring and management of time series models as well as perform rich asset-based queries over highly contextualized data. The user experience also folds in the current in-market ad-hoc interactive data exploration over raw data. In the public preview timeframe, the data exploration and asset-based insights are achieved via two separate time series environments; they will be seamlessly integrated into a single, unified environment in a soon-to-release public preview update. Customers using the in-market offering should expect to see their ad-hoc analytics in the enhanced experience while leveraging the public preview capabilities for asset-based insights.

Time Series Insight evironment creation

The enhanced user experience provides modern user experience enhancements for analytics including navigation to fast search and query of time series data, charting controls for easy and interactive visualization, and rich analytics experiences including markers to perform time-based correlations and analysis.

Data exploration example in Time Series Insights

Azure Time Series Insights is committed to our customers’ success

We look forward to continuing to deliver on our commitment of simplifying IoT for our customers and empowering them to achieve more with their IoT data and solutions. For more information, please visit the Time Series Insights product page and documentation. Also, try out the quickstart to begin using Time Series Insights today.

In addition, please provide feedback and suggestions on how we can improve in the product and the documentation. You can do this by scrolling down to the bottom of each documentation page where you can find a button for “product feedback” or sign in to your GitHub account and provide documentation feedback. We value your input and would love to hear from you.