Kafka Summit 2017 – Key Takeaways and Commentary
Last Monday on August 28th, a number of the VoltDB staff attended the Kafka Summit in San Francisco. This one-day event brought in over 700 attendees from the developer audience to attend 30+ valuable technical talks and breakout sessions, and meet directly with exhibiting sponsors like VoltDB to learn more about their technology – and most importantly, score some free booth swag.
In case you missed the event or were unable to catch all of the highlights from throughout the day, here were some of our key takeaways from the 2017 Kafka Summit.
The Summit as a whole is focused on developers and (unsurprisingly) Kafka streaming. Thus, most of the talks were about Kafka and streams, which were somewhat lackluster. This included a keynote and breakout session regarding KSQL. KSQL is still in its early stage of being and has a lot of weird limitations. The KSQL breakout session talk immediately turned into how good is the SQL support. When someone asked “can I do ‘INSERT’,” the answer was “not yet,” — it would require the user to add syntax to time-bound streams on both sides of the join. For example “join the last 5 minutes of this stream with the last 5 minutes of that stream”. However, it appears that people seem to really like SQL, as many vendors are providing SQL interfaces to their platforms.
Throughout the conference, the idea of a “continuous query” was being pushed more. This query produces output based on incoming events and never terminates, instead it just keeps producing rows based on the most recent input. We aren’t surprised that this feature is becoming more popular; VoltDB’s materialized view gives you exactly what a continuous query does.
Highlights from Booth #113
The conference attracted a good crop of developer audience. At the VoltDB booth (booth #113), we fielded a lot of discussion on what is transaction processing and why it is important. Our demo — which put the spotlight on our just-released User Defined Function (UDF) support feature — showcased a real-time fraud prevention use case, and drew in many folks perusing the various sponsor booths. Others who visited us were eager to share the details of the problems that they’re trying to solve using VoltDB and the Kafka importer. We heard from a few developers of streaming systems who are trying to present an interface that looks more like a database versus an immutable stream of messages.
Overall, the 2017 Kafka Summit was a great event and provided an excellent opportunity to speak directly with some really bright minds from the developer community. If you have any questions you wanted to follow-up with us on about the event – or you want to see our fraud demo leveraging the new VoltDB 7.6 UDF feature – just shoot us a note and we’d be happy to set up a meeting to go over it with you.