I don't believe that this is a problem. Actually I think this is a huge step forward, at least from my experience as a computer science researcher. What do the journals really provide? They assemble an editorial board, which decides who is going to review what. The reviews get written and the papers get published. Then universities get charged heaven and earth (those are your tax dollars, or my tuition dollars) for these things. The journals don't make information more accessible, which is necessary for science to advance, they restrict it by pricing it outrageously. But as we all know by now, scarcity of information products is a myth, and those who control choke points of information are loathe to give it up. Researchers need to publish or perish, so they have a vested interest in doing peer review -- plus it's how people find out about interesting results before everybody else. I don't think that peer review will collapse without Elsevier, Kluwer, or the other journal publishers.

I find it ironic that open source (also known as commons based peer production) is often described as the application of the scientific method to computer software. Yet, the scientific method is now constrained by artificial scarcity of the information product that it produces. Perhaps the open source and scientific communities could work together to solve this problem.