Microsoft announced general availability of Azure Stream Analytics (ASA), a fully managed cloud service for real-time processing of streaming data last week, shortly after one of its chief competitors Google launched a beta version of Dataflow, its own stream analytics service.
The big cloud providers continue to expand their offerings into the realm of big data and Internet of Things. ASA’s goal is to make it easy to set up real-time analytic computations on data coming in from sensors and other devices, collectively known as the Internet of Things, as well as websites, applications, and infrastructure systems.
Such analytics capabilities used to be the domain of big enterprises, but the big public cloud providers are beginning to offer users sophisticated analytics without the hefty price tag and complexities that often come along. IBM and HP have also been active in offering cloud analytics services, and so has Amazon.
ASA is a secure multi-tenant service, according to Microsoft. Customers can allocate and pay for the resources they use, so they can start small and scale out as needed. Resiliency and check pointing for auto recovery are built into the platform.
The cloud analytics service supports a high-level SQL-like language that simplifies the logic to visualize and act on data in real time. ASA can serve as backend for a variety of IoT applications, such as remote device management or getting insight connected cars.
“Simple configuration settings allow developers to tackle the complexities of managing network latencies from sensors sending data to the cloud, correctly ordering events across thousands of sensors to find correct patterns etc,” wrote Joseph Sirosh, corporate vice president of Information Management and Machine Learning at Microsoft. “The stream analysis logic can also be easily tested and debugged within the internet browser itself before deploying it live in the cloud.”
ASA provides a rapid development experience, removing all unnecessary overhead of traditional programming languages such as Java, according to Sirosh. An example given is computing a moving average on a temporal window. In ASA it takes five lines of code compared to hundreds in Java/Storm.
Two of the early users are NEC, which uses ASA for face detection among other things, and Fujitsu, which is using it for environmental monitoring and management in manufacturing, collecting data from a variety of end points and machines. ASA provides real-time analytics to accelerate factory-wide optimization.
“In the past, CEP (Complex Event Processing) in the cloud was not a realistic option, but Azure Stream Analytics showed great performance running on Azure,” said Hiromitsu Oikawa, a director at Fujitsu, wrote on Microsoft’s blog.