Back in early 2017, we posed the question of whether the world needed another streaming engine. In an announcement, timed for the Asian business day, Alibaba gave its answer to that question. It bought Berlin-based data Artisans for $103 million, which was a huge premium for a company that previously had drawn modest financing ($6.5 million Series A plus an undisclosed Series B). It was announced by respective blogs from both companies.
In the hours since the announcement, there has been plenty of coverage.
Yes, there are lots of streaming engines out there. The prominent comparison for Apache Flink is Apache Spark: both are about unified, high-performance computing, not just streaming. The differences are that Flink was designed around streaming whereas Spark's origins were microbatch – but both have since moved to bridge those differences.
What distinguished Flink is its approach to stateful streaming, it wasn't alone here – that was also the aspiration of the dormant Apache Apex project (where the prime backer, Data Torrent, went out of business last year). Data Artisans has taken that a step further with patent-pending technology for serializable transactions, meaning it can handle the more complex ACID stuff involving multiple rows and tables. data Artisans Streaming Ledger is the first engine to bring distributed ACID transactions to stream processing.
But it's no secret that data Artisans and Apache Flink have operated in the shadows of Databricks and Spark. Until now, data Artisans has drawn only a small fraction of the backing of Databricks, which in turn recently established an OEM deal with Microsoft for OEM'ing its service as a native part of the Azure analytics portfolio.
But Alibaba, better known as the Amazon.com of China, has been stepping up its technology investments of late, for instance, funding much of its latest $27 million venture round, and a supported service for MariaDB in its cloud.
Investment in Flink shows that Alibaba is charting its own course – as part of an increasingly assertive China, Alibaba clearly aspires to be leader, not follower. While Spark had the earlier start in garnering commercial support, Alibaba's acquisition of the company provides the second chance to build a critical mass skills and application base. And that leaves open the possibilities for synergies with MariaDB, which could integrate or OEM the technology as an ACID-compliant streaming database.