I've been rolling the Elastic license stuff over in my head this weekend. Here's a brain dump.

1/7

An original sin of most recent open source infrastructure is that it's corporate driven. But it's hard to be successful running an open source project while at a corporation; there are competing demands and differing requirements.

2/7

So most recent open source spins out from the company into its own NEW company. Now they have to pitch to VCs and they are faced with three strategies: support, proprietary add-ons, and cloud hosting.

3/7

Support is a non-starter; it scales linearly and VCs don't like it. It's a good supplement, though.

4/7

Proprietary works OK, not great. This was the Cloudera/Horton strategy. You can build a multi-billion dollar business, but it has limits. Some enterprises are stand-offish of proprietary solutions. Open source clones come out. It also discourages open source improvements.

5/7

Cloud is the gravy boat. But now you're competing against AWS, GCP, and Azure. And you're competing on THEIR turf. Their hosting cost will ALWAYS be lower than yours. You have to figure out which levers to pull.

6/7

Some differentiators:

* Integration network effects
* Development velocity
* Hybrid/multi-cloud
* Support/proprietary add-ons
* Better operating the software
* Better roadmap; deeper understanding of the space
* Community influence

7/7

Or just be Snowflake.

8/7