Today we’re pleased to announce two key capabilities that Azure Time Series Insights will be delivering later this year:
- A cost-effective long-term storage that enables a cloud-based solution to trend years’ worth of time series data pivoted on devices/tags.
- A device-based (also known industry-wide as “tag-based”) user experience backed by a time series model to contextualize raw time series data with device metadata and domain hierarchies.
Additionally, Time Series Insights will be integrating with advanced machine learning and analytics tools like Spark and Jupyter notebooks to help customers tackle time series data challenges in new ways. Data scientists and process engineers in industries like oil & gas, power & utility, manufacturing, and building management rely on time series data solutions for critical tasks like storage, data analysis, and KPI tracking and they’ll be able to do this using Time Series Insights .
Time series model and tag-centric experience
Time Series Insights’ current user interface is great for data scientists and analysts. However, process engineers and asset operators may not always find this experience natural to use. To address this, we are adding a device-based user experience to the Time Series Insights explorer. This new interface and the underlying time series model that backs the experience will enable OT workers to intuitively find devices related to the assets they care about. By enabling hierarchy and device-based semantics that contextualize the raw time series data, we enable richer and deeper analytics. This means that finding and comparing observation targets (i.e., devices/tags) to trend and explore in the Time Series Insights user experience or with our REST APIs will be seamless.
The below diagram is an example of the rich tag-based experience that we are developing in the Time Series Insights explorer.
To help organizations effortlessly scale their time series solutions, Time Series Insights will offer seamless integration with massively-scalable, cost-effective storage, archival and queryability in Azure Storage. This additional layer of storage creates a powerful duality for customers to engage their data while maximizing costs. Our warm layer, what Time Series Insights customers know and use today, continues to support interactive analytics. With the cold layer, these same customers will now have a single-source of truth for their time series data in the cloud, at a price that works. We expect most customers to store 30–120 days of data in the warm layer, and 1–20 years in the cold layer, thereby blending the best of both worlds. Cold storage, coupled with device/tag-centric querying across all data, means that customers maximize the cost-benefits of the cloud while still realizing performant querying and trending of historical time series data.
Below is a diagram that describes Time Series Insights’ high level architecture and scenarios:
Connectors to enable rich e2e solutions
Time Series Insights will store data in Apache Parquet files based on device/tag and timestamp properties, thus optimizing integration with powerful tools like Azure Databricks. Integration with machine learning tools like Azure Machine Learning Studio and Jupyter Notebooks is simplified, so organizations can build models to predict future device states and avoid wasteful maintenance. Time Series Insights makes it easy to collaborate and share insights in seconds through integration with Power BI, Microsoft Excel, and other business intelligence reporting software.
We’re collaborating closely with customers like TransAlta, to define and build this new infrastructure and solutions using Time Series Insights. Below are quotes from TransAlta’s CTO and Enterprise Architect attesting to the business advantages they see with adopting Azure IoT and Time Series Insights solutions.
“Time Series Insights is the cornerstone of our IoT platform and will be pivotal in enabling our larger AI and Machine Learning strategies going forward. We have partnered closely with the Microsoft TSI team on the two features being announced today. We believe that the tag-centric user experience will enable all our users, from Engineering to Operations, to get more value from our data, due to an ease of access that we have not had as an organization with previous solutions. Time Series Insights’ new long-term storage will empower our data science and engineering teams, in collaboration with our operations teams, to discover and quickly take action on insights previously hidden in our data. As we migrate from an on-premises to a cloud-based solution, building our platform on Time Series Insights has allowed us to focus our energy on enabling our business partners to meet the growing demands of a changing Power industry, rather than becoming experts on cloud technologies.”
– Jason Killeleagh, Enterprise Architect, TransAlta
“Energy demand is growing every year, and so are the pressures to deliver energy efficiently and sustainably. TransAlta is right in the middle of a major transformation that will ensure our global generation operations have the appropriate tools to enable timely decision-making in a rapidly changing energy market. Time Series Insights and other Azure IoT services are supporting our ability to tackle these challenges in a more dynamic, flexible and cost-effective manner than the traditional on-premises solutions we have reviewed. We see Microsoft and the Azure IoT team as a strategic partner helping us drive digital transformation in the energy sector.”
– Nipa Chakravarti, CTO, TransAlta
Time Series Insights provides a global view of an organization's data – enabling customers to collect and generate insights from highly distributed IoT data. Time Series Insights’ REST APIs can query across devices/tags, so customers can build domain-specific solutions on top of Time Series Insights to view data streaming from multiple sites in seconds and query data as fast as they can today with a server stored down the hallway. A great example of a customer building on top of Time Series Insights’ APIs is the industrial automation leader, ABB. Recently, ABB has adopted Time Series Insights as a focal point for their mining, ES, and robotics platforms that will use Time Series Insights’ long-term storage to empower their customers with rich monitoring and analytics solutions. Below is a quote from ABB Ability’s Group Vice President of Product on their use of Time Series Insights.
“ABB is leveraging Azure Time Series Insights inside our ABB Ability™ Platform to help us build innovative solutions for our customers such as ABB Ability™ Connected Services for robotics and ABB Ability™ Remote Diagnostic Services for mining. We chose to build on Time Series Insights because we needed a scalable and performant, fully-managed platform. Time Series Insights enables us to deliver rich, interactive analytics in our solutions that make it easy for customers to solve problems and keep their assets running at peak performance.”
– Sean Parham, Group Vice President of Product Management, ABB Ability™
With the capabilities we are announcing today, Time Series Insights is evolving from a short-term asset monitoring and diagnostics service to a modern cloud IoT platform for customers and partners to build highly-capable and scalable IoT solutions.
We will be at Hannover Messe this week. Microsoft will exhibit in the Digital Factory at HMI (booth #C40), focusing on the benefits of intelligent manufacturing. We’ll feature the world’s leading innovators through solution showcases and allow you to engage with our newest technologies enabling our customers to build a IoT solutions. Don’t hesitate to stop by and learn more.
Source: Azure Blog Feed