Certainly, given that each ebook could be a static hash. But you'd need to revise index pages periodically. And I'm not sure how search could be implemented. Except through Google, which does index pages from https://ipfs.io/ipfs/ . However, I gather that IPFS respects take-down requests, so you might need to use another gateway.