A secure comparison technique for tree structured data

 Comparing different versions of large tree structured data is a CPU and memory intensive task. State of the art techniques require the complete XML trees and their internal representations to be loaded into memory before any comparison may start. Furthermore, comparing sanitized XML trees is not addressed by these techniques. We propose a comparison technique for sanitized XML documents which ultimately results into a minimum cost edit script transforming the initial tree into the target tree. This method uses encrypted integer labels to encode the original XML structure and content, making the encrypted XML readable only by a legitimate party. Encoded tree nodes can be compared by a third party with a limited intermediate representation.