In talking with a large enterprise buyer today, I was struck by an argument he used against subscription models, open source and otherwise. Granted, he was negotiating and I've heard the inverse argument from others, but he had a good point.
The point? That an upfront proprietary license might actually work better for some IT buyers.
Subscription value is clear: The vendor is tasked with delivering constant software and support to earn the customer's business on a daily basis. No sell-them-and-run deals. It completely changes the way vendors engage with their customers.
But on the IT buyer's side, a subscription's price is likely to be higher on an annual basis than the maintenance on a proprietary license. In the first year, a subscription is dramatically cheaper. But over five years...? Or how about just in the second year, or third? It's no longer so clear-cut.
What is clear is when the vendor takes their profit: Upfront in the case of license-based businesses, and in the future in the case of subscription-based businesses. It's that niggling question of the future that may be problematic for open-source businesses.
Consider your "typical" open-source business, giving the software away and charging for support. The most burdensome year for support is the first year, when a customer is going into production. Arguably, support businesses don't see much profit from a customer until the second and third years, when the customer has become proficient with the software and needs less support.
Perversely, this is also when they have less need to pay for software support (though they may still need to pay for maintenance/upgrades), as Jon Williams noted in his OSBC keynote. Could it be that support-based businesses end up shooting themselves in the foot in their very attempt to help customers?
Back to license-based businesses. I don't like this model because it's the exact opposite problem: Vendors shooting their customers in the foot, with customers seeing little benefit beyond "Well, at least I won't have to write that million-dollar check again!" The benefit mostly flows in the vendor's direction, even if short-sighted IT buyers quickly forget about the big upfront license fee to focus on the low ongoing maintenance cost.
Perhaps MySQL and others are wise to blend the best of open source and license-based business models to ensure a more equal footing between their current and future ability to service customer needs? The object isn't lock-in but rather buy-in, buy-in that supports ongoing product development, now and in the future.