I'm about to open a big can of worms here...right ...
# general
c
I'm about to open a big can of worms here...right now https://littlehorse.dev is closed-source (we plan to make it available as a SaaS/PaaS). The spiritual goal of the project is similar to Temporal (plenty of differentiation, but that's a different story). The clients are Apache 2.0 OSS, and there is a public sandbox to "try it out". We are considering going with the Server-Side Public License for the core runtime. A few questions: • How much of a difference would that make in terms of getting adoption (versus keeping the core code proprietary)? • Is the SSPL safe in terms of preventing AWS from hosting LittleHorse-as-a-Service? • Once something is OSS, you can't put the cat back in the bag. Is there a risk of losing significant revenue this way?
l
You ask a lot of good questions. I think OSS can fundamentally make a difference in adoption because people will feel that they have an ejection point, even if they never plan on using it - giving you a good answer to vendor lock-in. This is especially helpful for situations where you are pitching to technical folks. However - you need to ensure that your platform provides significantly more value than the OSS offering. Initially this will be marketed as "not having to do the work to setup, manage, & scale" your PaaS. But over the long-run, infrastructure becomes a commodity - managing it will not be enough of a value-add to justify adoption. You need to have an excellent feature-set around your PaaS that is closed source - I'm sure Gwen would point to Confluent as a good example - Kafka is open-source, but Confluent gives you an entire ecosystem on top of it and lots of cool things like automatic partition rebalancing/scaling. If they layer between your OSS & your managed PaaS is thin, it won't work out in the long-run. As far as the license prevent AWS from copying you - I'm not sure about that. What I do know is that AWS fundamentally wants to be the place that hosts SaaS/PaaS businesses; they want companies like your to build products like yours. Unless you reach a giant critical mass on the scale of something like Postgres or Kafka, they're probably not going to copy you. And even if they did, it's highly likely they would do so in a proprietary manner similar to many of their offerings. I would view AWS as more of a partner than potential competitor - in particular the AWS Marketplace may be a great way to deliver your platform to potential customers'
šŸ™Œ 1
Also - OSS isn't end-all be all; it works for a lot of businesses, but there are closed-source SaaS companies that are incredibly successful (looking at you DataDog)
c
Got it, thanks Lucas!
here are closed-source SaaS companies that are incredibly successful (looking at you DataDog)
I would add to that list: • Salesforce • Snowflake • Splunk One observation though: while those companies are related to software development, they're not core to software development. DataDog and Splunk are monitoring, Snowflake is analytics, and Salesforce is for salespeople. In contrast, Temporal, Confluent, and Mongo are core to their customers' tech stacks. I'm not sure if I can find an example of another closed-source company that is core to tech stacks, but I'm also young šŸ™‚
šŸ‘ 1
l
I don't think that's by accident - the closer you are to the core stack, the more marketing you need to do to technical folks who are going to be nervous about vendor lock in. That being said, the cloud providers themselves are generally closed-source (even though many of their offerings are managed open-source). Vercel, Fly.io, and Railway are some examples of startups whose core runtimes are closed-source. Ultimately it's up to you to decide which is right for your company! Given that Temporal seems to be a direct competitor for you, it's not unwise to assume that you'll also need to be OSS - but they are also still pretty early and will need to address the challenges I listed above
c
That all makes sense. Thank you for the insight!
šŸ‘ 1