Akamai Technologies, Inc.

Terence Lobo

James V Luciani

Prasandh Balakrishnan

Richard G D'Addio

Lawrence Gensch

Ravi S Aysola

Ittehad Shaikh

A content delivery network (CDN) is enhanced to enable mobile network operators (MNOs) to provide their mobile device users with a content prediction and pre-fetching service. Preferably, the CDN enables the service by providing infrastructure support comprising a client application, and a distributed predictive pre-fetching function. The client application executes in the user's mobile device and enables the device user to subscribe to content (e.g., video) from different websites, and to input viewing preferences for such content (e.g.: 'Sports: MLB: Boston Red Sox'). This user subscription and preference information is sent to the predictive pre-fetching support function that is preferably implemented within or across CDN server clusters. A preferred implementation uses a centralized back-end infrastructure, together with front-end servers positioned in association with the edge server regions located nearby the mobile core network. The predictive pre-fetch service operates on the user's behalf in accordance with the user preference information.

Managing mobile device user subscription and service preferences to predictively pre-fetch content

Disclosed herein are methods and systems for detecting congestion in a mobile network and for determining those end-user mobile devices that are affected. In one embodiment, a server communicates with a set of mobile devices, on which have been installed a suitable client application. At certain times, the server initiates a congestion detection routine. The server may request the mobile devices to report on their current wireless attachment point to the mobile network. The server can then test for congestion by performing a data transfer between itself and the mobile clients, which may be in either direction. The server can use the results to determine whether a given attachment point is congested. In one embodiment, a dynamically selected, random subset of mobile devices for a current attachment point are tested, and the result is imputed to all mobile devices similarly situated.

Congestion detection in mobile networks and delivery of content in non-congested conditions

Growing community of inventors

Acton, MA, United States of America

Terence Lobo