Great read about how OSS companies can defend ourselves against cloud providers: https://www.forbes.com/sites/glennsolomon/2020/10/13/finding-trillions-in-the-clouds-how-open-source-companies-can-win-in-a-world-dominated-by-amazon-microsoft-and-google/

There’s a lot of backlash against Elasticsearch for the way they changed their licensing in response to AWS doing what they do. I think that’s backfired from their intention of stopping AWS from eating their lunch. From a business perspective (where the core value proposition of a business is not software) the idea of using an AWS managed version (knackered as they are) of some open source solution is very appealing. I don’t want to be a sysadmin and worry about observability or site reliability or managing backups, HA etc Especially for data pipelines where egress costs or sovereignty is a concern, there’s also resistance to hooking yet another SaaS in via ssh tunnels or VPC peering, but I think those things are getting easier and businesses are more willing to accept multi cloud configurations like that. One way to compete is just be a lot better at it e.g. Confluent as a managed Kafka, or Starburst as a managed Presto (vs Athena) or I would have said until recently Stichdata as a managed singer-tap. I’m fairly disappointed with the way singer has ended up. Such a great idea and huge amount of effort into the project only for it all to be abandoned to community contributions. It’s great that it supports so many taps but then so many of them are broken and there’s no developer advocacy or effort to bring the community taps up to the quality of core taps. And now of course Talend has bought them and there’s even more commercial focus and lack of understanding about FOSS and effective business models. I hope that AirByte doesn’t go that way. One great example where I think they are doing it right is Fishtown / dbt. Their community is amazing, they welcome blogs and tools that help their users to host and run the open source tools themselves. The core of their SaaS offering is the OSS code. Their business model is a very reasonably priced cloud orchestration, monitoring, resilient service and an IDE for “analytics engineers” to ease them into the world of git flow and SDLC. If AWS tried a managed version of dbt cloud I can’t see how they could make it nicer than Fishtown’s offering. It would end up being some weird extension to cloud9 or code deploy. There’s another good example, GitHub, Gitlab, bitbucket and others are very popular as are CI/CD services that integrate with them like CircleCI for example. AWS has offerings in that space but those businesses that focus on CI/CD just do it so much better. I’ve ranted ... but also, free user feedback!

Thanks very much, @Dave! Very insightful 🙂. And completely agree with you!!