Robust Hashing for Models

Abstract

The increased adoption of model-driven engineering (MDE) in complex industrial environments highlights the value of a company's modeling artefacts. As such, any MDE ecosystem must provide mechanisms to both, protect, and take full advantage of these valuable assets. In this sense, we explore the adaptation of the Robust Hashing technique to the MDE domain. Indeed, robust hashing algorithms (i.e. hashing algorithms that generate similar outputs from similar input data), have been proved useful as a key building block in intellectual property protection, authenticity assessment and fast comparison and retrieval solutions for different application domains. We present a novel robust hashing mechanism for models based on the use of model fragmentation and locality sensitive hashing. We discuss the usefulness of this technique on a number of scenarios and its feasibility by providing a prototype implementation and corresponding experimental evaluation. CCS CONCEPTS Software and its engineering Model-driven software engineering; System modeling languages; Information systems Similarity measures; Near-duplicate and plagiarism detection;

Efficient plagiarism detection for software modeling assignments

“…Robust hashing functions, as described in Section 2, meet this requirement (among others such as being resistant to certain manipulations). Thus, we adopt here the robust hashing approach for models (Martínez, Gérard, & Cabot, 2018) were the authors adapted the minhash technique (Broder, 1997) in order to transform models into vectors of n symbols.…”

“…An introduction to LSH for different metric spaces is given in Leskovec, Rajaraman, and Ullman (2014) from where we borrow notation and definitions. This transformation towards the vector metric space works because, as shown by Martínez et al (2018), robust hashing vectors estimate well the similarity between models.…”