Yes, although there's a cutting-edge form of encryption called homomorphic encryption that would allow full-text search on encrypted data if an implementation is ever perfected. IBM is working on it, among many others:
Erm, no. It's actually quite easy, if you're prepared to do a little more work.
You just create a search index before encrypting the data. Then you encrypt the index. Each time you need to search, you decrypt the index, get a reference to one or more results, fetch those, and decrypt them.
That's trivialising what can become a pretty complex scenario, but it illustrates the point.
Actually, I did an email archiver as a side project, I still have the Lucene index of my emails (including searchable attachments, of course). It's 233MB. How long do you think it would take to decrypt it with AES256 everytime I search for something? How long it takes when you search in Gmail? Can you point me to any implementation that does it like that? Or any implementation that does it at all?
I wasn't saying it's impossible, just that it's pretty hard.
I don't know of anything open source. Initialising the AES crypto provider takes a lot longer than decrypting ~250 Mb data, so keep it initialized for the duration of the session.
That said, I did it in a native client app, where state is easier to maintain.
