

Berlin-based big-data company data Artisans GmbH today announced that its dA Platform based on the Apache Flink stream processing framework, is generally available almost a year and a half after it was launched in preview.
Apache Flink first came to the attention of enterprises back in 2015 when it was promoted to “Top Level Project” status by the Apache Software Foundation. Flink was originally designed to replace the MapReduce component of the Apache Hadoop big data software framework. It boasts a much faster engine that can handle large flows of real-time data, including activity monitoring, machine learning and business intelligence.
Flink is often compared to Apache Spark, another big data processing framework that can tackle stream processing thanks to its Spark Streaming extension. However, most experts agree that Spark is not a genuine streaming engine, because it processes streaming data in small chunks, in a process referred to as “microbatching.” This necessitates the accumulation of data in batches for a short time before processing, so it can’t claim to stream data in truly real time.
Data Artisans defines genuine stream processing as “the processing of data in motion, or in other words, computing on data directly as it is produced or received.” The company describes Flink is a “stateful stream processing engine” that processes data in true real time without any latency.
This gives Flink a significant advantage over Spark because data that’s processed via microbatching cannot be computed at the same time. With Flink’s genuine real-time streaming, however, the ability to compute on data as it’s received provides genuine real-time analytics capabilities that are unavailable on Spark.
The dA Platform has been adopted by a number of well-known enterprises to power their live streaming data-based applications. One early adopter is the ride-hailing firm Uber Technologies Inc., which finally settled on Flink after experimenting with alternatives including Spark. Chinmay Soman, a staff software engineer at the company, said in an interview on SiliconANGLE’s mobile TV studio theCUBE last year that Uber decided to use Flink because Spark and others it tried were too “resource intensive.”
The latest version of the dA Platform comes with a new feature called the Application Manager, which provides tools that streamline and simplify the deployment of real-time data streaming apps in production. With Application Manager, it’s also possible to integrate the Kubernetes container orchestration platform, and logging and metric systems such as Grafana and Lostash, with the dA Platform.
Data Artisans has also built a set of application programming interfaces that can be used to integrate other platforms with the dA Platform.
“dA Platform makes it easier than ever for businesses to deploy and manage streaming applications, allowing them focus on building strategic new real-time products and services for customers rather than the supporting infrastructure,” said Kostas Tzoumas (pictured), co-founder and chief executive officer of data Artisans.
A trial version of the dA Platform is available to download now. Pricing is based on the number of cores used via a minimum one-year subscription.
THANK YOU