A practical key management and distribution system for IPTV conditional access

Harding, Gregory Leighton (2013)

Thesis

Conditional Access (CA) is widely used by pay-television operators to restrict access to content to authorised subscribers. Commercial CA solutions are available for structured broadcast and Internet Protocol Television (IPTV) environments, as well as Internet-based video-on-demand services, however these solutions are mostly proprietary, often inefficient for use on IP networks, and frequently depend on smartcards for maintaining security. An efficient, exible, and open conditional access system that can be implemented practically by operators with large numbers of subscribers would be beneficial to those operators and Set-Top-Box manufacturers in terms of cost savings for royalties and production costs. Furthermore, organisations such as the South African Broadcasting Corporation that are transitioning to Digital-Terrestrial-Television could use an open Conditional Access System (CAS) to restrict content to viewing within national borders and to ensure that only valid TV licence holders are able to access content. To this end, a system was developed that draws from the area of group key management. Users are grouped according to their subscription selections and these groups are authorised for each selection's constituent services. Group keys are updated with a key-tree based approach that includes a novel method for growing full trees that outperforms the standard method. The relations that are created between key trees are used to establish a hierarchy of keys which allows exible selection of services whilst maintaining their cryptographic protection. Conditions for security without dependence on smartcards are defined, and the system is expandable to multi-home viewing scenarios. A prototype implementation was used to assess the proposed system. Total memory consumption of the key-server, bandwidth usage for transmission of key updates, and client processing and storage of keys were all demonstrated to be highly scalable with number of subscribers and number of services.

Collections: