On November 15th, Microsoft announced that Azure Time Series Insights (TSI), which has been in preview since April, is now available to the general public. Part of Azure’s greater Internet of Things (IoT) suite of offerings, TSI gives companies the capacity to store and analyze enormous volumes of time series data in near real-time. Here we look at what time series analysis is; the problems presented by time series data; TSI’s sweeping solution; how companies are leveraging TSI; and how to get started with this powerful new platform.
What is time series analysis?
To understand the value of Azure Time Series Insights, it’s important to first understand the concept of time series analysis. This form of analysis involves gleaning statistics and insights from raw, time-ordered telemetry data, such as periodic meter readings. Time series analysis often makes substantive, cost-saving and time-saving contributions to operations troubleshooting and forecasting.
Enormous volumes of telemetry data
Because of the cumulative nature of time series data, as well as the vast, growing number of sources of new telemetry events in the IoT, a company’s time series data can quickly become voluminous and unwieldy. Before Azure’s cloud-based TSI solution, many enterprises found that storing and analyzing large volumes of time series data at scale was extremely difficult.
The inefficiency of on-premises solutions
Traditional solutions to the IoT data problem are based on-premises databases that often prove prohibitively expensive, complex and time-consuming in terms of development, data preparation, setup, and management.
As the data from IoT grows exponentially, scaling your company’s solution, no matter how robust it may be, becomes an even more burdensome task.
In the end, many companies have found that it is difficult or impossible to derive consistently valuable insight from the enormous, ever-growing volume of time series data gathered and stored with traditional on-premises solutions.
In contrast to traditional solutions, Azure Time Series Insights does not require that enterprises devote their IT resources to development or data preparation. TSI is a fully managed solution, hosted in the cloud, with all the tools in place to analyze enormous volumes of time series data, then present them to the customer in an intuitive fashion.
TSI can store terabytes of time series data for more than a year. Users can view billions of telemetry events in near real-time by location, region or device through TSI’s advanced data-visualization and analysis tools.
Compatibility and integration
TSI provides REST query APIs that make it possible to easily integrate your TSI data into your new or pre-existing apps. TSI also works with data from any device. Time series data flowing into the system may come from sensors on mobile devices, assembly-line elements, wind turbines, outdoor meters, or any number of other devices participating in the IoT.
TSI can rapidly scale to accommodate the growth of your company or your data. Elastic scaling can happen in seconds to reach multi-terabyte levels.
An example: temperature sensor data
The Internet of Things (IoT) is often pictured in terms of smart household appliances that are constantly uploading and downloading data to improve what they can do for households. However, the IoT also applies directly to the enterprise. This is especially true in industries, such as manufacturing, that involve machines with sensors that relay telemetry data to IT departments to help measure performance, quality, safety, and other key metrics over time. Depending on the size of the business, telemetry data can account for billions of events over a matter of months. The high-volume of accumulating time series data can make storage and analysis a formidable undertaking for even the most IT-savvy enterprise.
To illustrate the problem-solving power of TSI, the Azure team provides the example of a global company with stamping machines. Each machine has a heat sensor that relays temperature information throughout the day to the company’s database in the Azure cloud. At one point, a heat sensor on a stamping machine at one of the company’s many locations reports a temperature that is above a critical threshold. The event sets off a temperature alarm and automatically alerts corporate headquarters. Because the company has Azure Time Series Insights in place, they can quickly view and analyze billions of such sensor events globally to start troubleshooting the problem in seconds. The analysis quickly identifies a pattern strongly suggesting that the heat generated by adjacent stamping machines is the source of the reported high-temperature anomaly. To remedy the situation, the company sets new guidelines, presumably to cap heat emissions from adjacent stamping machines and thereby improves their global operations. Overall, the insights delivered in near real-time by TSI reduce the company’s downtime, increase their earnings and improve their global operations.
Getting started with TSI
Quality, security, and support
Before trying out a new cloud service at scale, it’s important to know that the platform is ready to meet the quality, security and support standards of the enterprise. Answering to these concerns, TSI makes it clear that their new service is ISO-certified, and comes with an SLA that helps companies extend their on-premises security and policy standards to the cloud. TSI also utilizes established features provided by Azure Active Directory including multi-factor authentication, single sign-on capabilities, role-based access restrictions, and robust identity management. Finally, TSI comes with 24/7 support.
As a first step to deploying your TSI-powered solution, you will need to choose a plan. For maximum flexibility, TSI uses as a pay-as-you-go payment structure that places no limitations on total users or queries. There are two pricing tiers: S1 and S2. For both tiers, pricing is centered around the concept an event, which Azure defines as a single datum with a timestamp. Both S1 and S2 offer up to thirteen months data retention with up to ten units. S1, at only $150 per unit per month, limits storage to “30 GB or 20 million events.” S2, on the other hand, at $1,350 per unit per month, allows for ten times the storage.
Despite the scope and power of Azure Time Series Insights, the platform is surprisingly easy to setup and use. TSI works directly with Azure IoT Hub, where your company’s time series data resides. Once IoT Hub is in place, there’s no code necessary for TSI setup, which can take place in seconds. Querying the data is also straightforward. Members of your team that are familiar with performing SQL queries will have no problem querying the TSI database, which does not require learning a new language. For those unfamiliar with SQL queries, or who simply prefer at-a-glance visualizations, TSI provides intuitive analytics and visualizations that require nothing more than a series of selections and clicks to gather and view aggregated event data in near real-time.
Leveraging TSI in the age of enterprise IoT
Microsoft’s recent announcement that TSI is now available to the general public presents an opportunity for companies struggling to store and analyze large volumes of telemetry data. TSI offers companies a near real-time, cloud-based solution that shows what the IoT can do for the cloud-connected enterprise. Above we’ve explored the problems faced by companies with high-volumes of time series data flowing in from multi-site, or even global, operations in near real-time. While scale-resistant, on-premises solutions often lead to frustration, TSI’s cloud-based, a rapid-scaling solution can store and analyze terabytes of near real-time telemetry data and can scale in minutes. As the example above of a global company using TSI to solve a temperature-related issue clearly illustrates, TSI is a highly practical solution that can consistently reduce downtime and improve company earnings. TSI’s integration with Azure IoT Hub and Active Directory, together with flexible pay-as-you-go pricing, makes TSI a remarkably affordable, secure, robust and easy-to-use new platform.