I learned a lot at the inaugural Kafka Summit in San Francisco last year, so I was excited when its organizers recently announced that there would be two summits this year: one in San Francisco again, and one in my backyard in New York City. Not only that, but a few different Confluent employees — some very smart folks — suggested that I submit a talk proposal. This was gratifying but also a bit enervating — I wasn’t sure I’d be able to find the time to come up with a thoughtful talk idea. Somehow though, I managed.

First I just came up with a title: My First Kafka Streams Apps: a case study, a comedy of errors, a happy ending and ran that by a few Confluent folks. The responses were positive, so I went ahead and wrote up an abstract and submitted it. I thought I’d share it here for posterity.

And so here it is:


Please provide a short, up to 400 characters, description of your proposal. If your submission is accepted, this is what will be published on the Kafka Summit website.

I’d been excited about KS for years, and I’d built stream processing apps before — what could go wrong? Plenty. Maybe I shouldn’t have tried to learn KS, Avro, and Ruby all at the same time! Follow along as I share the joys and sorrows of being an early adopter; figure out transactional stream processing with KS and JRuby; launch a new system; screw up; and ultimately succeed!


Please provide a brief overview of why your submission should be selected for Kafka Summit.

There are lots of “here’s what I did” talks/posts, and lots of “here’s how this works, here’s what to do” explainer-type talks/posts, but not so many “here’s what not to do” talks/posts. Sometimes negative examples can be equally as helpful as positive examples, because you can learn about the sharp edges of a technology or technique; you just get a different perspective. This talk won’t be only about negative examples, but they’ll probably be the core of it.


What is required of the participants?

Basic conceptual familiarity with stream processing.

Please provide at least 1 link to a video of yourself presenting to an audience. If a video is not currently available online, please create a very short clip (2-3 minutes) proposing the session. The quality of the video is not important, we care about the quality of the speakers. If your video isn’t already online, post it to a third-party site (like YouTube), and then share the link with us.


Of course, I have no idea whether the talk will be accepted or not… I’ll find out on 17 February!