19 May 2016

An algorithm for juicier gossip

Today, more and more of us get our news and information from social networking sites, which include the likes of Twitter, Facebook, Google+, LinkedIn and countless others. This runs in obvious parallel with the slow decline of printed magazines and newspapers as well as television being usurped by YouTube and on-demand video streaming services such as Netflix. Our lives are changing as a result.

A new study from the Nanjing Research Institute of Electronics Engineering (NRIEE), in China, suggests that there is now a need to improve exactly how information is spread through online social networking. A “gossip algorithm” could improve the reach and rate at which important, useful or simply entertaining information spreads, Bo Yang and colleagues suggest in the International Journal of Ad Hoc and Ubiquitous Computing. Their findings are particularly pertinent to the ease with which information can be divested across a mobile social network.

The team has demonstrated how a gossip algorithm can be used to spread information with simplicity and robustness through the 4G-enabled world of smart phone ad hoc social networks that utilise D2D (device to device) technology including LTE Direct and Wi-Fi Direct. The team has successfully tested their protocol under in two typical application scenarios where smart phone users took a random walk. They determined the theoretical upper bounds of how long a piece of information would take to spread to all users in the ad hoc D2D mobile network. “Simulation results indicate that the growth of communication cost is almost linear [as opposed to geometric or exponential growth] when the scale of mobile social networks becomes larger, which is a quite encouraging result for application,” the team reports.

Given just how reliant we are all becoming on continuous and rapid access to information even when we are on the move, the advent of mobile social networks that allow clusters of smart phone users to spread that information efficiently without recourse to all devices interminably sending and receiving data to and from the cellular network and thence the internet will become increasingly important. Yang and his colleagues offer a novel solution in their gossip algorithm for reducing the cellular network overheads.

Yang, B., Liu, D. and Zhang, W. (2016) ‘Information dissemination in mobile social networks with gossip algorithms‘, Int. J. Ad Hoc and Ubiquitous Computing, Vol. 21, No. 4, pp.259-265
via Inderscience – Science Spot http://ift.tt/252gd8w

No comments: