Distributed search, similar to bittorrent, DHT-based designs and the like are notoriously difficult.
I've participated in such efforts, like the Seeks Project [1], Yacy [2], and related initiatives like Unhosted [3], and it takes a certain amount of dedication (and suffering ;) ).
However, I believe it is not entirely impossible that we see a true alternative sometimes. From what experience, what is needed is a slightly better set of distributed algorithms, a business model with the ability to sustain such a technical effort, and a range of features that no search engine can yet offer (because centralized).
[1] http://www.seeks-project.info/ [2] http://www.yacy.net/ [3] https://www.unhosted.org/