Mobile device reputation system

In a centralized reputation computer, a method of managing a centralized reputation service, the method comprising:

• receiving ratings of source identifications of communications from a plurality of end user mobile cellular devices;
• receiving a query from a first end user device for information regarding the reputability of a first source identification;
• using a collaborative filtering algorithm to calculate a prediction for the first source identification from the ratings of source identifications corresponding to the first source identification, wherein the prediction is an indication of the likelihood that the source identification is malicious or otherwise untrustworthy, and wherein said prediction is calculated on-the-fly after said receiving said query and is not calculated before receiving said query;
• sending said information including said prediction regarding the reputability of the first source to the first end user device;
• and
• comparing said prediction to a predefined threshold by a software agent on said first end user device.

A mobile cellular device reputation system comprising:

• a plurality of end-user mobile cellular devices, wherein each of the plurality of end-user mobile cellular devices includes a software agent configured to interface with the control module of the centralized reputation server computer;
• a rating information database including ratings associated with a communication source, said ratings originating from said mobile cellular devices;
• a prediction information database including a prediction of reputability for said communication source based on said ratings, wherein said prediction is calculated on-the-fly after said receiving a query and is not calculated before receiving said query;
• and
• a centralized reputation server computer having a control module and in communication with said end-user mobile cellular devices, wherein said control module facilitates receiving said ratings from said mobile cellular devices and delivering said prediction of reputability to one of said mobile cellular devices, and wherein said software agent of said one mobile cellular device being further configured to automatically, without user intervention, reject a communication from said communication source when said delivered prediction does not meet or exceed a predefined threshold.

A method of handling a communication received on a first mobile cellular device, the method comprising:

• determining a source identification of the communication;
• scanning a locally stored banned list for the source identification of the communication;
• rejecting the communication if the source identification is on the locally stored banned list;
• and
• when it is determined that the source identification is not on the locally stored banned list:
• querying a centralized reputation service with the source identification;
• receiving a prediction of the source identification from the centralized reputation service, wherein the prediction is an indication of the likelihood that the source identification of the communication is malicious or otherwise untrustworthy, as calculated by a collaborative filtering algorithm from a plurality of ratings of the source identification by users on a plurality of mobile cellular devices other than the first mobile cellular device, and wherein said prediction is calculated on-the-fly after said querying and is not calculated before said querying;
• comparing the received prediction to a predefined threshold by a software agent on said first mobile cellular device;
• and
• automatically, and without user intervention, rejecting the communication when the received prediction does not meet or exceed the predefined threshold.