Meeting the challenges of data security

Our research has significantly improved the security of printing and network environments.

The security of data in printing and network environments is an area of increasing concern to individuals, businesses and government organisations throughout the world.

The threat to printing and imaging devices has increased over the last decade. This is as a result of more sophisticated tactics used by hackers, increasingly mobile workforces and changes in industry regulations. The threats posed include unauthorised changes to settings that would allow someone to reroute print jobs or access network and password information.

Data theft can result in the loss of millions of pounds, and huge damage to reputation. As a result companies worldwide have sought to combat unauthorised and illegal practices, whilst not impacting normal usage.

Mathematical algorithms developed by the School of Mathematics represent a significant step-change in existing data security techniques.

The algorithms developed by Cardiff University are novel to data mining and are extremely valuable to our organisation. They enable us to successfully compete within the industry and drive technology forward to meet the evolving needs of our clients. Without the research produced by Alex Balinsky our achievements in this area would not have been possible.

Dr Steven Simske Director and Chief Technologist for Security Printing and Imaging Engineering, Hewlett Packard

Rapid change detection

The algorithms developed enable greater security in automatic document classification and summarisation, recognising normal and abnormal patterns of data and distinguishing between confidential and non-confidential documents.

Hewlett-Packard (HP), the world’s leading PC vendor, funded the research underpinning this development and patented the resulting software, with the aim of strengthening its position as the market leader in the global information technology industry.

The algorithms have been incorporated in a schedule of upgrades in over ten million of HP’s electronic devices.

Research team

Professor Balinsky was assisted by two PhD students: N. Mohammad (2007-2011, EPSRC CASE award with Hewlett-Packard) and B. Dadachev (2011- , funded jointly by Cardiff University and Hewlett-Packard). Dr Mohammad was immediately employed by HP on completion of his PhD.

Providing security benefits

The new approach to rapid change detection developed at Cardiff identifies documents that are confidential and prevents them from being printed by unauthorised users.

The algorithms are essentially valid safeguards for all data transmitted via printing and network applications.

The feature extractor developed by Cardiff University detects unusual behaviour, and this ability means that unauthorised access can be prevented and therefore it provides enormous security benefits for HP’s extensive client base.

When evaluating this feature extractor, HP found that it considerably outperformed other feature extractors currently in use.