Microsoft makes Trill engine open-source for high-speed analytics

Microsoft has today created Trill open-source. That’s a data processing engine that can analyze billions of events per second. The project is the latest in a growing list of internal technologies that the company shares with the world. Many of these techniques are aimed at artificial intelligence.

That’s what the Silicon Angle site reports today. The Trill engine is an important addition to that list of open-source techniques. The engine is versatile enough to handle both real-time and historical data – a combination that offers few open-source tools. Vibration links this support of multimodal analytics with a set of other functions that Microsoft believes are not covered by other systems.

Billions of operations

The Trill engine delivers strong performance. For example, certain events, such as temperature fluctuations measured by soil sensors, can be processed billions of times per second. In the case of advanced operations, such as extracting the highest temperature from multiple sensors, the engine can process 10 million to 100 million events per second.

Vibration works just as well on real-time as offline datasets and on the whole delivers the best performance, says James Terwilliger, chief researcher at the Microsoft Azure Group in a blog post. This makes it the most interesting engine for users who want to perform all their analyses with one tool.

Another important function of Vibration lies in its adaptability. The engine allows developers to change the interval at which incoming data is processed. This allows users to control the delay.

Microsoft uses Vibration for a range of different applications. For example, the engine forms the basis of the Azure Stream Analytics service, supports the Bing Ads and helps the Microsoft video game unit analyze data generated by players. The fact that Trill is widely used within Microsoft makes it an interesting engine for developers.