Open source solutions will precipitate proprietary opportunity

Open source has changed the game from a cost of software point of view. The traditional model with a split of support over cost/license of 70/30 is changing to a full service model, the 30% has disappeared. This "cost free" element of Free Open Source has launched a bunch of new commodities, for example, LAMP is a serious contender to .NET or J2EE. The fact that it is freely available has been hugely influential here.

But the software bits are not the whole picture, it is the solutions that are build on them that are important. I think we will find that the Open Source commodities will precipitate the commoditisation of traditional closed source or proprietary components through complete integrated solutions. The proprietary modules that are really good at what they do and that play well with Open Source will be successful. That is, they are easily integrated with existing open source components and platforms, are well supported, reasonably priced and reasonable open (from a standards and extensibility point of view).
The proprietary components that work well with open source will become commodities on the back of the open source wave, the customers/market will decide what works well and what does not; essentially the overall integrated solution will be open, the proprietary bits that integrated easily will form an integral part of that solution.
Other uses with the same itch will try the same solution, to facilitate the proven complete solution they will pay for a suitable proprietary component.

One problem with this model is that a real open source alternative commodity will eventually get created by the community, unless of course there are large barriers to entry like huge complexity or the need for a mainframe to test it or something. It may simply mean that traditional proprietary source has a very short life span, build it, charge for it, then give it away once the 'critical mass' of use can provide a support revenue stream.

Hybrid open solutions that combine mostly FOSS with proprietary modules are inevitable, I think a software company can find a niche with a traditional business model applied in 'internet time'.

Services companies will find a niche in customization, and the commonality their-in will be the genesis for a traditional product.