The goals of this paper are to analyze the ongoing trends in innovation related to security in cloud computing, and to identify areas for improvement. To this end, we collect all cloud security patent applications (about 512 patent applications) from the United States Patent Trademark Office (USPTO) and apply data mining techniques to their titles. This allows us to identify unique keywords that these patent applications share, and to apply social network analysis to study the relationships between them. We then use cluster analysis to uncover specific focal innovation areas that correspond to the well-known confidentiality, integrity and availability (CIA) triad, thus complementing the USPTO classification of cloud security innovations. Our results reveal in particular that the area of cloud availability is under-investigated, suggesting that more innovation is needed on this front.