Clayton M. Christensen in his book "Harvard Business Review on Innovation" (2001) the author explains why it is usually not large companies with resources, capabilities and staff that initiate the disruptive innovations but rather support sustaining (incremental~) innovations.
Sustaining innovations are nearly always developed and introduced by established industry leaders. But the same companies never introduce - or cope well with - disruptive innovations. Why? Our resource-process framework holds the answer. Industry leaders are organized to develop and introduce sustaining technologies.
He goes on by telling:
Disruptive innovations occur so intermittently that no company has the routine process for handling them.
The issue is that smaller companies may lack the resources , "but that does not matter" because they have qualities that allow them to respond to emerging growth markets: their processes for R&D and market research and the fact that not all decisions need to be backed up by management, give them some unique capabilities.
Interestingly enough, larger companies can create for themselves these capabilities in various ways. Though he does not elaborate on this very much (he mentions the 'superteams'), communities of practice provide exactly the advantageous attributes that smaller companies have (the leadership and role of management, decision making by the collective). "Organizational boundaries are an impediment" Clayton says; communities of practice in their very nature cross these formal organizational boundaries and support the emergence and creation of new processes to support radical innovation in a unparalleled way.
Fiona Lettice wrote a nice article with a more general scrope of disruptive innovation (not in relationship to communities) in the Knowledgeboard "Q&A on Disruptive Innovation: The Challenges for Managing Knowledge"