jonb-streamlio
commented
6 years ago @RobIsHere appreciate the question. Although the software has been in production use at Yahoo for over 3 years, it was only open sourced as Apache Pulsar more recently and so is still earlier in its lifecycle than some other projects.
That said, there is a rapidly growing number of Pulsar users. A number of them are big companies who like its performance and scalability (but who don't like to talk publicly about newer parts of their technology stack), and there are also a growing number of mid-size and even startup companies using it for its flexibility and features. For example there's a large industrial company using it, a well-known consumer electronics company, a large media company, a transportation services startup, and a tax planning company, just to name a few. (For what it's worth, StackOverflow isn't always an accurate indicator of adoption--I've worked for companies who had several hundred companies using the technology but very little activity on StackOverflow. When there are responsive, knowledgeable alternatives such as a vendor supporting the technology, questions tend to go directly there instead.)
It's true that the connector ecosystem is smaller than that of longer-established projects like Kafka, but with the recent release of Pulsar IO (see https://streaml.io/blog/introducing-pulsar-io/), it's become a lot easier for new connectors to be created. At Streamlio we've been working with a number of customers and solutions partners on connectors using that framework, which will also help expand the connector universe. In addition the client options continue to grow--in addition to C++, Java, and Python, there have been recent new additions from the Pulsar community including Go (see https://pulsar.apache.org/docs/en/client-libraries/#go-client and https://github.com/Comcast/pulsar-client-go) and Scala (https://github.com/sksamuel/pulsar4s).
To the broader question of whether marketing is the deciding factor in success of new projects, in general marketing isn't the biggest factor in the early days of a project, it's the technology-savvy people who look beyond "what is a safe choice because lots of other people are doing it" to adopt new technology based on understanding the differentiation in the technology and connecting that to what matters for them. With that as the foundation, marketing ramps up and helps to scale sharing their stories to the broader set of adopters who are less driven by an understanding of the technology differentiation and more by seeing that other people like them are using it.
The great thing about newer projects is how quickly they evolve--without the legacy baggage of design decisions made years ago, they generally outpace older technologies in innovation, which adds fuel to the early adoption fires.
joefk
pckeyan
RobIsHere
ivankelly
I'm looking into Pulsar and from an architectural standpoint it's made like I would have designed the system. So I'd prefer it over Kafka from this perspective. But there is a really small user base, e.g. only some messages about Pulsar on Stackoverflow.
So my quesition is: as there are many technical better projects loosing against others that are better in marketing - will this project last and what do you do to get users onboard?
I think this is a quesition that concerns a lot of people who look on pulsar. Because even when pulsar is open source, my Investment in my software that works with pulsar is huge. Addressing this on the web or in docs could make sense.