Maintaining docker builds isn’t that huge of a burden (and likely very useful for them too), and they’re delegating hosting to a third party… I don’t get what they’re trying to achieve here.
I am really interested in this case, as I cannot imagine any commercial benefit from this decision.
Free users will not pay tens/hundreds of thousands for just binary files.
Obviously this will slow down the adoption of the AGPL version, against which the growth of the paid version potentially will look better in Excel reports for VCs, but something tells me that this is unlikely to be the real reason.
> Free users will not pay tens/hundreds of thousands for just binary files.
Sad to say but this isn't true. This is a failure to understand the pricing model of this kind of enterprise software.
What happens is the free version is used in some product somewhere. Then product's company gets acquired by HugeBigCo. Product company brain drain happens and HugeBigCo looks at poorly understood free software dependency as a liability. It's cheaper/better-on-the-balance-sheet to pay for a license and a support contract than to move off of that dependency or hire competent people to look after it...For a few years, anyway -- until that business unit is worth investing resources in.
That's how a company like Neo4j can charge a half-million bucks a year for one production cluster and get HugeBigCo as label that they can use to try to convince other companies that this pricing is even remotely reasonable.
Anything enterprise data-storage SaaS related is looking to charge at least a quarter-million a year.
It's not just binaries. The goal is probably to get freeloaders into the sales funnel where salespeople can work on them about the value of the entire enterprise package, especially security and support. Even if sales doesn't work on you it does work on a lot of people.