CROSS-NETWORK REPUTATION FOR ONLINE SERVICES

• US 20110271329A1
• Filed: 07/01/2011
• Published: 11/03/2011
• Est. Priority Date: 01/18/2008
• Status: Active Grant

• Current Assignee: Microsoft Technology Licensing LLC (Microsoft Corporation)
• Original Assignee: Microsoft Corporation
• Inventors: Guo, Wei-Quiang Michael, Hulten, Geoffrey John, Iverson, Kristofer Noel
• Granted Patent: US 8,484,700 B2
• Time in Patent Office: Days
• US Class Current: 726/4
• CPC Class Codes: H04L 63/102 Entity profiles
• CPC Class Codes: H04L 63/1441 Countermeasures against malicious traffic countermeasures against attacks on cryptographic mechanisms H04L9/002
• CPC Class Codes: H04L 67/306 User profiles

Abstract

A reputation server associates feedback from previous network transactions with an account of a user in a network. A reputation score for the user is calculated based on the feedback to indicate the probability the user will abuse the network. When an online service receives a request to perform a transaction from the user, the online service performs the transaction based on the user'"'"'s reputation score. Additionally, a server generates a reputation packet including the reputation score for a user for use by an online service when the user requests the online service to perform a transaction. The online service may authenticate the reputation packet with the server and, if the reputation packet is authenticated, the online service performs the transaction based on the user'"'"'s reputation score.

Claims

1. A method of authenticating a user based on a network-wide reputation for the user, said user having a user account associated therewith for accessing one or more online services of the network, said method comprising:

collecting reputation feedback associated with the user account, said reputation feedback being based on one or more past transactions between the user and at least one of the online services of the network over a predetermined period of time, said reputation feedback indicating whether or not any of the past transactions constitutes a misuse of the network;

executing a machine learning algorithm on the reputation feedback to generate a model representing a likelihood of misuse of the network by the user;

in response to receiving a request from the user to perform a present transaction with at least one of the online services of the network, determining from the model a reputation score indicating the likelihood of misuse of the network for the present transaction;

when the determined reputation score is not sufficient to grant the user'"'"'s request to perform the present transaction, requesting the user to solve a proof; and

granting the user'"'"'s request to perform the present transaction based on receiving the solved proof.

2. The method of claim 1, wherein the proof comprises either a computational proof or a human interactive proof or both.

3. The method of claim 1, further comprising receiving, from the at least one of the online services, additional reputation feedback based on the requested solved proof and adjusting the model based thereon.

4. The method of claim 1, wherein the reputation score indicates a greater likelihood of misuse of the network when at least one of the past transactions includes one or more of the following:

SPAM, spyware, a virus, copyrighted material, a phishing attack, content prohibited by a user agreement, and pornography.

5. The method of claim 1, wherein the reputation score is based on one or more of the following reputation feedback metrics:

a number of times the user has completed a transaction with the one or more online services, a number of other network users who have interacted with content generated by the user;

a number of transactions the user has performed involving other network users, a number of complaints the user has received from other network users of the one or more online services, a number of times the user'"'"'s content has been viewed by a network moderator or other highly trusted network user, a number of times the user had content removed because of an abuse report, a number of days since the user'"'"'s user account was created, a number of unique days the user has been active on the network, an amount of traffic over the past 24 hours divided by the amount of traffic generated by user in an average 24 hour period, an amount of money the user has spent in the network, an amount of proof the user has provided, and a safety status of a client machine the user uses to access the network.

6. The method of claim 1, further comprising generating a reputation packet in response to the user'"'"'s request to perform the present transaction, said reputation packet including the determined reputation score, wherein the request and the reputation packet are forwarded to the online service associated with the request, said online service granting the request of the user based on the reputation packet.

7. The method of claim 6, wherein the reputation packet includes one or more of the following:

an identity of the user, a timestamp indicating the lifetime of the reputation packet, a suggested proof needed to grant the request, and information needed for the online service to provide reputation feedback regarding the user'"'"'s present transaction.

8. The method of claim 1, wherein the reputation score that indicates an acceptable likelihood for granting the user'"'"'s request to perform the present transaction differs as a function of one or more of the following:

a type of transaction, a type of online service, a time of the request, and a number of past transactions.

9. The method of claim 1, wherein the online services include one or more of the following:

a social networking service, an email service, a chat service, a photo sharing service, a file storage and sharing service, a shopping service, a search service, a map viewing service, and an instant message service.

10. The method of claim 1, wherein the past and present transactions include one or more of the following:

sending an email, sending an instant message, sending message in a chat room, saving a file, accessing a file, purchasing a product or service from a shopping service, searching the Internet, searching a web space, searching for directions, searching for an address, posting a message to a blog, posting audio to a blog, posting video to a blog, and posting a comment on a blog.

11. A method for authenticating a user based on a network-wide reputation for the user comprising:

sending a login request to a reputation server of the network, said reputation server receiving reputation feedback regarding the user'"'"'s past interactions with the network over a predetermined period of time and applying a machine learning algorithm to the received reputation feedback to generate a statistical model for the user, said model representing a probability that the user will misuse the network, said reputation server further generating a reputation packet in response to the received login request, said reputation packet including a reputation score of the user based on the model;

receiving the reputation packet from the reputation server;

sending a request to an online service of the network to perform a transaction, said request including the reputation packet, said reputation packet being authenticated via the reputation server before the online service grants the request;

receiving a request from the online service to solve a proof when the user'"'"'s reputation score is not sufficient to grant the request to perform the transaction; and

sending the requested proof to the online service, said online service granting the request to perform the transaction based on the reputation score in the reputation packet and in response to the user providing the solved proof.

12. The method of claim 11, wherein the proof comprises either a computational proof or a human interactive proof or both.

13. The method of claim 11, wherein the online service provides additional reputation feedback to the reputation server based on the proof received from the user and the reputation server adjusts the model based thereon.

14. The method of 11, wherein the reputation packet includes one or more of the following:

an identity of the user, a timestamp indicating the lifetime of the reputation packet, a suggested proof needed to grant the request, and information needed for the online service to provide feedback regarding the user'"'"'s requested transaction.

15. A system for calculating a reputation score of a user of a network, said network including one or more online services, said system having a memory area storing a plurality of computer-executable components, said components comprising:

a feedback component for receiving reputation feedback regarding one or more transactions by the user with at least one of the online services, said feedback indicating whether any of the user'"'"'s transactions with the online services constitutes a misuse of the network;

a machine learning model applied to the reputation feedback for determining the reputation score of the user based on the reputation feedback, said reputation score indicating a probability that a subsequent transaction by the user is a misuse of the network;

a reputation interface for receiving requests from one or more of the online services for the reputation score of the user and for sending the reputation score of the user to the requesting online service; and

a proof interface for requesting the user to perform a proof when the user'"'"'s reputation score is not sufficient to grant a request from the user and for receiving the requested proof from the user, said requested proof being sent to the online service, and the online service granting the user'"'"'s request in response to the proof.

16. The system of claim 15, wherein the online service provides reputation feedback to the feedback component based on the received proof and wherein the machine learning model adjusts the user'"'"'s reputation score based on the provided reputation feedback.

17. The system of claim 15, wherein the proof comprises either a computational proof or a human interactive proof or both.

18. The system of claim 15, wherein the reputation interface generates a reputation packet in response to the request from the user, said reputation packet including the determined reputation score, wherein the request from the user and the reputation packet are forwarded to the online service associated with the user'"'"'s request, said online service granting the request of the user based on the reputation packet.

19. The system of claim 18, wherein the reputation packet includes one or more of the following:

an identity of the user, a timestamp indicating the lifetime of the reputation packet, a suggested proof needed to grant the request, and information needed for the online service to provide reputation feedback regarding the user'"'"'s present transaction.

20. The system of claim 15, wherein the online services include one or more of the following:

a social networking service, an email service, a chat service, a photo sharing service, a file storage and sharing service, a shopping service, a search service, a map viewing service, and an instant message service.