Smart recommendation of mobile services to consumers

This paper describes the design and implementation of a smart recommendation system used to discover and suggest to consumers the 'best' mobile services in the emerging ubiquitous consumer wireless world (UCWW). The cloud-based system builds up and dynamically manages personal profiles of consumers with big data processing techniques. A consumer identity module (CIM), based on the Java Card technology, supports a third-party authentication, authorization and accounting procedure (3P-AAA), and provides a trusted execution environment for corresponding mobile applications running on the consumer's mobile device. A distributed data management platform (DMP), built with the Hadoop, Storm, and Kafka technologies, provides an efficient computing environment for real-time data scheduling and message processing in the cloud. With this distributed architecture, the system is able to turn mobile services' activities of consumers into actionable analytic datasets and recommend the 'best' mobile services applicable to each particular consumer under the 'always best connected and best served' (ABC&S) paradigm. A number of novel software solutions, both on the consumer and cloud side, are presented. An approach towards running this service recommendation system in a distributed way and high throughput capacity is elaborated.