Classifications

• • H—ELECTRICITY
  • H04—ELECTRIC COMMUNICATION TECHNIQUE
  • H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
  • H04L63/00—Network architectures or network communication protocols for network security
  • H04L63/08—Network architectures or network communication protocols for network security for authentication of entities
  • H04L63/0876—Network architectures or network communication protocols for network security for authentication of entities based on the identity of the terminal or configuration, e.g. MAC address, hardware or software configuration or device fingerprint
• • G—PHYSICS
  • G06—COMPUTING; CALCULATING OR COUNTING
  • G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
  • G06Q30/00—Commerce
  • G06Q30/06—Buying, selling or leasing transactions
• • G—PHYSICS
  • G06—COMPUTING; CALCULATING OR COUNTING
  • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
  • G06V40/00—Recognition of biometric, human-related or animal-related patterns in image or video data
  • G06V40/50—Maintenance of biometric data or enrolment thereof
  • G06V40/53—Measures to keep reference information secret, e.g. cancellable biometrics
• • H—ELECTRICITY
  • H04—ELECTRIC COMMUNICATION TECHNIQUE
  • H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
  • H04L63/00—Network architectures or network communication protocols for network security
  • H04L63/10—Network architectures or network communication protocols for network security for controlling access to devices or network resources
  • H04L63/102—Entity profiles
• • H—ELECTRICITY
  • H04—ELECTRIC COMMUNICATION TECHNIQUE
  • H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
  • H04L63/00—Network architectures or network communication protocols for network security
  • H04L63/30—Network architectures or network communication protocols for network security for supporting lawful interception, monitoring or retaining of communications or communication related information
  • H04L63/308—Network architectures or network communication protocols for network security for supporting lawful interception, monitoring or retaining of communications or communication related information retaining data, e.g. retaining successful, unsuccessful communication attempts, internet access, or e-mail, internet telephony, intercept related information or call content
• • H—ELECTRICITY
  • H04—ELECTRIC COMMUNICATION TECHNIQUE
  • H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
  • H04L67/00—Network arrangements or protocols for supporting network services or applications
  • H04L67/2866—Architectures; Arrangements
  • H04L67/30—Profiles

Definitions

• Patent Application Serial No. 13/612,755 is a continuation-in- part of U.S. Non-Provisional Application No. 13/229,481, filed September 9, 201 1, which claims priority to claims priority from U.S. Provisional Application No. 61/527,469, filed August 25, 201 1, and both of which applications are hereby incorporated in their entirety by reference.
• the user greatly depends on the data in those information systems.
• the transaction should ensure that the credit/debit card used for payment corresponds to the user.
• the transaction should also should ensure that the identity of the person is authenticated.
• authentication is the performing of tests to guarantee within a known degree of confidence that a user corresponds to a user identity when interacting with an information system.
• authentication is performed by several common methods. Authentication is typically performed by verifying a user's indicia for that user's identity.
• the user's indicia are called credentials.
• a user's credentials may come in the form of a user proffering a known value, such as a password or personal identification number ("PIN").
• PIN personal identification number
• a user's credentials may come in the form by a user proffering a token such as a proximity card, or a fingerprint or retina scan.
• authentication presently relies on credentials in the form of a user possessing a known value, or of a user physically holding a token.
• identity theft can occur when known values based on memorization are hacked, or tokens are stolen or otherwise misappropriated.
• many information systems only authenticate users upon logging onto a system, and subsequently limit system requests to verify identity as not to constantly interrupt the user. Accordingly, there is an opportunity to improve security and prevent identity theft via identifying additional means of authentication.
• FIG. 1 is a top level diagram illustrating an example multi-factor identity fingerprinting service collecting data relating to user historical activity for access via an example profile based authentication service.
• FIG. 2 is an example hardware platform for multi-factor identity fingerprinting.
• FIG. 3 is a flow chart of an example process for multi-factor identity fingerprinting.
• FIG. 4 is a top level diagram illustrating an example application of multi-factor identity fingerprinting in the mobile media vertical.
• FIG. 5 is a block diagram of an exemplary logical architecture for a security fingerprint platform.
• FIG. 6 is a diagram of an exemplary implementation of a logical architecture for a security fingerprint platform from an application and services perspective.
• FIG. 7 is a diagram of an exemplary implementation of a logical architecture for a security fingerprint platform from the perspective of a fingerprint.
• FIG. 8 is a diagram illustrating the relationship between fingerprint services applications, primitive fingerprint operations, and the underlying data from data assets.
• FIG. 9 is a diagram of an implementation of an exemplary application to assist a memory impaired patient based on security fingerprints.
• FIG. 10 is a diagram of an implementation of an exemplary matchmaking application based on security fingerprints.
• FIG. 1 1 is a diagram of an implementation of an exemplary application based on merging security fingerprints.
• This disclosure describes multi-factor identity fingerprinting with user behavior.
• user behavior There is presently a high frequency of user interaction with a diversity of information systems. Accordingly, each user has a critical mass of interactions that may be tracked whose factors may be associated with a user's identity.
• multiple factors relating to user behavior are stored in a data structure called a profile and aggregated in the profile as a history of the user's behavior.
• the data structure may be persisted in a computer readable memory.
• a least some subset of the user's interactions stored in the profile may be used to generate an identity fingerprint that subsequently constitute a user's credentials.
• a factor can be any pattern of observable values relating to a user interaction. These factors may then be used as input in generating an identity fingerprint.
• Example observable values may include tracking when a user accesses one of their social networking pages, tracking the web address of the page, tracking the time the page was accessed, or tracking particular action performed such as posting a new picture or entering a comment.
• these observables are stored in a user profile, they are called historical activities.
• an information system receives an event notification, that event notification may be stored as a historical activity in the user's profile.
• these values are stored in a profile and used to determine factors such as usage patterns with one or more applications and/or one or more client devices, as well as the associated user preferences.
• Usage patterns with applications and/or a client device are a factor that relates to tracking what data is accessed, and what application or client device features are typically availed to by a user.
• An example of a usage pattern is determining that www.mysocialnet.com is the most commonly accessed web site via a web browser called CoolBrowser.exe.
• usage patterns are but one consideration in generating a multi-factor identity fingerprint.
• User behavior may be another factor. User behavior relates to correlations of usage patterns with other input other that the application or client device itself. An example might be determining a user typically accesses www.mysocialnet.com around 1 1 :30AM every day, indicating that the user is updating their social network records during lunch breaks. Another example might be the user typically accessing www.fredspizza.com on rainy Sundays, indicating that the user does not typically go out for food when raining. [0023] User preferences may be yet another factor. Applications and client devices typically have user setting indicating user preferences in using those applications and client devices respectively.
• a factor can be based on any values that may be detected and stored, and subsequently may be a potential factor used in multi-factor identity fingerprinting. Factors themselves may be either stored with the profile, or otherwise dynamically derived.
• multi-factor identity fingerprinting at least a subset of a user's profile stored online becomes bound to that user. In some embodiments, the user's identity may be used as that user's credentials.
• the information system may authenticate or verify a us Q '' n ⁇ +
• the information system may have authentication capabilities able to access the user identity finger print or to query the user profile, built in-system itself, or alternatively may delegate those functions to a separate system.
• security attacks may be catalogued and aggregated. Since an information system does not rely on a password or a physical token, the information system may compare any event or notification during the user's session, compare it with the user's identity fingerprint, and determine whether the user's behavior is consistent with the identity fingerprint or alternatively consistent with a query against the user's profile. Since the identity fingerprint is readily accessible, there is no need to interrupt the user's session with requests for passwords or other tokens. Thus a larger set of security checks may be monitored. This information may be analyzed to identify patterns of security attacks/threat monitoring or for identity management. [0027] In yet another embodiment, identity fingerprints may be used to discover categories of usage among users.
• the identity fingerprint provides a snapshot of a user's history, the identity fingerprint is very difficult to diverge from a user's actual or likely behavior. Accordingly, high confidence can be ascribed in comparing and aggregating different identity fingerprints. Identified categories may subsequently be used to direct advertising or to obtain business intelligence. Overview
• Figure 1 illustrates one possible embodiment of multi-factor identity fingerprinting 100. Specifically, it illustrates how a user 102 progresses over time and develops a historical profile and an identity fingerprint that may be used subsequently for authentication.
• User 102 may have client device A 104 and use it to make an interaction 106 with an information system. Interaction 106 could possibly be user 102 using client device A 104 to access a web site called www.awebstore.com. User 102 may make some purchases during interaction 106.
• Observable values collected during interaction 106 and subsequent interactions may be stored as historical activity records in a user profile via profile collection service 108.
• the set of records of user 102's historical activities is user 102's profile.
• the information collected during interaction 106 and subsequent interactions are converted into one or more records of user 102's historical activities.
• profile collection service 108 stores records of user 102's historical activities with user 102's profile in a data store 1 10.
• interaction 1 12 may be user 102 using user client device A 104 to update the user's ⁇ -" ⁇ - : " 1 i - ⁇ * ⁇ ⁇ resort + www.mysocialnet.com.
• user 102's historical activities during interaction 1 12's are captured by the profile collection service 108 and stored in data store 1 10.
• a user 102's profile need not be specific to a particular site or to a particular type of interaction. Any definable and observable user event whose parameters may be captured is a candidate for storing as one or more historical activity records for user 102's profile. Collecting event information and collecting parameters to create historical activity records is described in further detail with respect to Figure 3.
• User 102's profile need not be specific to a particular client device. As shown via interaction 1 16, which may be after a number of other interactions, user 102 may use a different client device, here client device B 1 14 to interact with an information system.
• Interaction 1 16 could potentially be user 102 further updating user 102's social network records at www.mysocialnet.com, perhaps to upload a picture just taken with client device B 104.
• profile collection service 108 converts interaction 1 16 into one or more historical records associated with user 102's activities and stores those records as part of user 102's profile in data store 1 10.
• the user's profile may then be used to generate an identity fingerprint.
• an unknown user 120 using client device C 122 may attempt to edit user 102's social network records at www.mysocialnet.com.
• unknown user 120 may be in possession of user 102's password and thereby log into user 102's account on www.mysocialnet.com.
• unknown user 120 may attempt to make a post to user 102's social network records at www.mysocialnet.com.
• the posting attempt may trigger an event trapped by www.mysocialnet.com, which in turn may make an authentication request 124 via profile based authentication service 126.
• the profile based authentication service 126 may then convert the posting attempt into user activity indicia that is comparable to user 102's profile.
• profile based authentication service 126 may query data store 1 10 via profile collection service 108 for some subset of user 102's historical activity records. For example, authentication request 124 may limit retrieved records only to www.mysocialnet.com activity by user 102 over the past three years.
• Profile based authentication service 126 may generate a summary file of the retrieved records into an identity fingerprint for the user.
• the identity fingerprint comprises a summary of the user's history and may take many potential forms.
• the identity fingerprint may identify several different activities, and store the frequency the user performs those activities.
• the identity fingerprint may store other users that the user's account may send information to.
• the identity fingerprint may be cached, such that in lieu of the profile based authentication service 126 generating the identity fingerprint dynamically, it may be served directly.
• Profile based authentication service 126 may then correlate unknown user 120's activity against the identity fingerprint.
• the profile based authentication service 126 may report a low correlation with respect to the identity fingerprint. If the correlation is sufficiently low, the profile based authentication service 126 may send an error message indicating that authentication failed. Alternatively, if the correlation is sufficiently high, the profile based authentication service 126 may send an authentication message indicating successful authentication. If there is insufficient information to provide a statistically significant conclusion, the profile based authentication service 126 may simply send a message indicating no conclusion. In this way, the profile based authentication service 126 may lower false positives during authentication.
• unknown user 120 did not have to use the same client device as previously used by user 102. Rather than having physical possession of credentials, authenticating unknown user 120 was based on the user's profile, specifically as an identity fingerprint used as a credential and readily retrievable from data store 1 10. Furthermore, note that authentication using the identity fingerprint may operate independently or alternatively in conjunction with the www.mysocialnet.com' s login authentication. Even though unknown user 120 had user 102's password credentials, those credentials were independently verified against the user's identity fingerprint credential via the profile based authentication service 126. Moreover, this authentication process was transparent to unknown user 120.
• unknown user 120 cannot obtain the information from the user 102, since the behavioral aspects of user 102 is cannot be obtained through recollection and/or coercion. Accordingly, because of a lack of access to the profile based authentication process, unknown user 120 may have been able to hack or spoof www.mysocialnet.com' s login, but unknown user 120 was not able to spoof the profile based authentication process as it uses historical behavioral attributes. Unknown user 120 simply could not have changed the user 102's history over the past three years of never posting profanity. In this way, profile based authentication provides a more secure authentication, and provides continuous authentication separate from login's and other means where a user must explicitly enter credentials.
• the profile based authentication service 126 may be configured to simply block unknown user 120 from interacting with the information system. For less sensitive scenarios, the profile based authentication service 126 may be configured to require the unknown user 120 to proffer alternative credentials. For even less sensitive scenarios, the profile based authentication service 126 may be configured to simply send a notification in the form of electronic mail, text message, or other messaging services to user 102 that an unusual event occurred.
• the profile based authentication service 126 may be configured to have multiple of correlation models.
• Each correlation model is a statistical model which specifies how to calculate a similarity score of the user event and historical event data in the user profile and/or the user identity fingerprint.
• the correlation model may be very simple where the presence of certain terms is sufficient to return a result of zero correlation. Alternatively, the correlation model may be very complex and may comprise learning algorithms with a varying degree of confidence.
• the profile authentication service 126 may combine different correlation models to derive additional confidence in authentication results. Confidence models are discussed in further detail with respect to Figure 3.
• profile based authentication may be configured to meet the different authentication needs for different information systems.
• the profile based authentication service 126 may expose an application programming interface ("API") to be programmatically accessible to an arbitrary information system.
• API application programming interface
• the profile based authentication service 126 may be used in conjunction with credit card companies to provide additional indicia as to the identity of an arbitrary user.
• the user need not be in possession of a client device.
• the client device itself may be subject to authentication.
• the cellular service may make an authentication request 124 against the profile based authentication service 126 and may require the user provide additional credentials.
• the profile based authentication services can be configured to provide just the identity a specific verification answer, such as yes/no/ inconclusive, thereby protecting the subscriber's privacy.
• the profile based authentication service 126 may be used for non-authentication applications. For example, the profile based authentication service 126 may be queried by other services 128 for user identity fingerprints for analysis, and categories of user behavior may thereby be identified. These categories in conjunction with the histories of user behavior may be used for directed advertising or to generate general business intelligence.
• a service 128 desires to have access to more extensive data beyond the identity fingerprints, the service 128 can access the profile collection service 108 directly, which has a critical mass of user historical activities stored in data store 1 10.
• the services 128, such as business intelligence or advertising targeting services may access the user historical activity records in data store 1 10 via profile collection service 108 to perform queries unrelated to authentication.
• Other services 128 may include business intelligence and advertising applications as discussed above. However, they may also include servicing law enforcement data subpoenas, identity management, and threat management request.
• the profile collection service 108 and profile based authentication service 126 may incorporate a billing system to monetize authentication and data requests.
• the billing system may be a separate module, or alternatively incorporated into both the profile collection service 108 and profile based authentication service 126.
• the profile collection service 108 and profile based authentication service 126 may store records of each data and authentication request in data store 1 10 or other data store, which may then be queried to generate a bill.
• the profile collection service 108 and profile based authentication service 126 may store request counts by particular parties, and may generate a bill per alternative billing arrangements such as flat fees or service subscription models.
• Figure 2 illustrates one possible embodiment of a hardware environment 200 for multi-factor identity fingerprinting.
• a client device 202 configured collect user historical activity data either on the client device 202 itself or alternatively hosted on servers 204 and accessed via network connection 206. Examples of historical activity data collected on the client device 202 itself include trapping keystrokes, accessing local data such as photos, or monitoring local application usage such as entering web addresses into internet browsers.
• Figure 2 also illustrates the client device 202 configured to connect to the profile collection service 108 and/or profile based authentication service 126 as hosted on application server 208 via network connection 210.
• Network connection 206 relates to client device 202 accessing information systems as part of user activity and network connection 210 relates to accessing the profile collection system 108 and/or profile based authentication system 126.
• both network connection 206 and network connection 210 may be any method or system to connect to remote computing device. This may be in the form of both wired and wireless communications.
• the client device 202 may be personal computer on a wired Ethernet local area network or a wired point of sale system in a store.
• the network connections 206 and/or 210 may be wireless connections either via Wi-Fi for packet data or via cellular phone protocols which may include CDMA 2000, WCDMA, HSPA, LTE or successor cellular protocols. Accordingly, the preceding specification of network connections 206 and 210 is not intended to be limited by selection of network protocol.
• client device 202 might store user historical activity data or authentication requests locally. Interfacing with information system servers 204 or with profile based authentication application server 208 need not be via network collection. For example, locally stored user historical activity data or authentication requests may be stored on a portable memory stick and then used to manually access information servers 204 or profiled based authentication application server 208.
• Client device 202 is any computing device with a processor 212 and a memory 214.
• Client device 202 may optionally include a network interface 216.
• Client device 202 may be a cellular phone including a smart phone, a netbook, a laptop computer, a personal computer, or a dedicated computing terminal such as a point of sale system terminal.
• Client device 202 would also include distributed systems such as a terminal accessing a centralized server as with web top computing.
• Client device 202's memory 214 is any computer-readable media which may store include several programs 218 and alternatively nonexecutable data such as documents and pictures.
• Computer-readable media includes, at least, two types of computer-readable media, namely computer storage media and communications media.
• Computer storage media includes volatile and non-volatile, removable and non- removable media implemented in any method or technology for storage of information such as computer readable instructions, data structures, program modules, or other data.
• Computer storage media includes, but is not limited to, RAM, ROM, EEPROM, flash memory or other memory technology, CD-ROM, digital versatile disks (DVD) or other optical storage, magnetic cassettes, magnetic tape, magnetic disk storage or other magnetic storage devices, or any other non-transmission medium that can be used to store information for access by a computing device.
• communication media may embody computer readable instructions, data structures, program modules, or other data in a modulated data signal, such as a carrier wave, or other transmission mechanism. As defined herein, computer storage media does not include communication media.
• Programs 218 comprise computer-readable instructions including operating system and other system functionality as well as user applications.
• the operating system may support the ability to trap application events. Trapping application events enables a program to capture observable data that may subsequently stored as a user historical activity record. Examples include, but are not limited to journaling hooks and trampoline functions.
• a trapped application event may be associated with a programmatic handler which in turn stores input and/or output parameter data associated with the operation of the event. In this way, an arbitrary user event and interaction with application, may be monitored, associated data stored, and then processed for conversion into one or more user historical activity records.
• User applications may include applications designed for local use such as word processors or spreadsheets. Local applications may include utilities such as programs to monitor local usage. Applications in this class may include, but are not limited to keystroke monitors and near field communication monitors. Alternatively, user applications may include applications such as web browsers or cloud clients designed to interact with a remote systems.
• Application server 208 is any computing device capable of hosting profile collection system 108 and/or profile based authentication server 126.
• Application server 208 comprises processor 220, memory 222 and network interface 224.
• memory 222 is any computer-readable media including both computer storage media and communication media.
• programs 226 which may include an operating system and computer-readable instructions for profile collection system 108 and/or profile based authentication server 126.
• Memory 222 may also store programs 226 that may include a database management system if data store 228 is configured as a database.
• Data store 228 may be configured as a relational database, an object-oriented database, a columnar database, or any configuration supporting queries of user profiles and user historical activity data.
• FIG. 3 illustrates one possible embodiment of a multi-factor identity fingerprinting process 300.
• There are at least three different actors for multi-factor identity fingerprinting process 300 including: (1) the profile based authentication system, (2) a user being tracked and authenticated by the profile based authentication system and (3) a vendor or information system seeking to use the multi-factor identity fingerprinting system.
• Different actors will perceive different subsets of multi-factor identity fingerprinting process 300.
• the vendor or information system's perspective will vary depending on the application.
• Some systems will simply use the multi-factor identity fingerprinting system for authentication. Others will use the system to aggregate users and to identity usage patterns by a set of users.
• the multi-factor identity fingerprinting process 300 as a whole may be subdivided into the following broad sub-processes: 1. Data Collection/User Identity Fingerprint Generation 302,
• a user profile is bound to a particular user.
• the user profile will contain the user's historical activity records, and will be used as to generate the user's identity fingerprint. Since the user's identity fingerprint will be used the user's credentials and accordingly, the binding must be accurate.
• the user profile need not be bound to a particular client device. However, the user profile may contain a record that the user always uses particular client devices.
• Binding may be either static or dynamic. With static binding, a user may affirmatively create a user profile record with the profile based authentication system. In the record, the user may indicate client devices or applications typically accessed. From this information, the multi-factor identity fingerprinting system may more easily determine whether an incoming user historical activity record relates to a particular user profile. However, binding need not be static. Since the profile based authentication system's client devices may track indicia of user identity such as user logon information, the multi-factor identity fingerprinting system may aggregate records from similar logons independent of any static input from a user.
• One advantage of dynamically binding user historical activity records to a particular user is to distinguish different users who happen to use the same user accounts. For example, a single family account may be used by the owner of the account, the owner's spouse and the owner's child. In this case, the profile based authentication system may correctly generate three profiles (and subsequently user identity fingerprints corresponding to each of the three profiles) rather than just one. Thus the multi-factor identity fingerprinting system not only is not tied to a client device, it is also not tied to a particular user login or account for an information system.
• a client device or information system the client device is interacting with collects user information.
• a client device or information system enlists in a correlation model.
• the correlation model may specify particular user events, and for each user event may further specify data to be captured.
• the user event typically is an interaction with an application that may be captured by an operating systems eventing or notifications system. For example, if a user clicks on a button, the operating system may capture the button click, and as user information may capture the active application, the button identity along with the user identity.
• client device or information system may have an event handler that performs additional information lookup not specific to the captured event. For example, in addition to capturing the button click, the event handler may run a program to capture what other applications were open, or if there were any active network sessions.
• the client device may capture a very wide range of user information. It is precisely because it is possible to capture a wide range of possible user information that user information captured may be limited to events specified by a correlation model and the specific data used by the correlation model for each event.
• user information is imported into the associated correlation model.
• the user information is converted into user historical activity records. Specifically, the user information is parsed, and then mapped to a format that may be imported by the profile collection service 108 into the data store 1 10, for subsequent retrieval by the profile based authentication service 126 or other services 128.
• the raw data for a button click in an application called MyApp may come in the form of ("OKButton", UserBob, 12: 12:00PM, MyApp). This raw data may be converted into the following record (Profilel l l, MyApp: OKButton) through the following transformations:
• the account name UserBob may be mapped to a user profile with an identifier of Profile 1 1 1.
• the correlation model may have a format where the application and user interface element are concatenated together into a single field.
• OKButton and MyApp are converted + ⁇ ⁇ ⁇ ⁇ . / ⁇ > bark++ ⁇ (3)
• Some data may be eliminated as not relevant to a particular correlation model. In this example, the 12: 12:00PM time was simply dropped.
• Any number of transformations data actions may be performed against the raw user information prior to conversion into a user historical activity record.
• Third party data may be accessed for inclusion in the user historical activity record. For example, credit card identification or phone number identification information may be looked up and included in the user historical activity record.
• data validation may be performed. For example, prior to loading a record via the profile collection service 108 into the data store 1 10, the client can perform record format validation and value validation checks.
• event user information trapped need not be specific to a particular correlation model.
• a client device or information system may enlist in events rather than correlation models.
• Data store 1 10 may have a single database or multiple databases. Notwithstanding the number of databases used, data from multiple users from multiple client devices for multiple events may all be stored in data store 1 10.
• the multi-factor identity fingerprinting system generates a user identity fingerprint.
• the user ide" +;+T ' ⁇ ⁇ + — ⁇ - w ⁇ generated on demand or alternatively be proactive ly refreshed in a background process.
• At least a subset user historical records stored in a user's profile are used as the raw data to generate a user identity fingerprint.
• the user identity fingerprint is a summary of the user's history.
• the user identity fingerprint may be as simple as generating a single number used as a straightforward numerical score such as generating a credit rating or a grade for a class.
• the user identity fingerprint may provide a parcel of data summarizing relevant user activity.
• the fingerprint might report the number of bounced checks, the number of credit card rejections, and the number of returns a user performed at a store.
• Data in the identity fingerprint need not be numerical.
• the identity fingerprint may simply store a Boolean value.
• Data in the identity fingerprint need not be limited to data collected by a single system, but may be combined with external data.
• an identity fingerprint may combine a number of bounced checks with a record of times a user was arrested for credit card fraud.
• Authentication scenario 304 is from the perspective of the multi- factor identity fingerprinting system servicing a vendor's information system request to authenticate a user.
• an information system will trap an event that the information system is programmed to perform a profile based authentication request.
• the information system will trap the event and associated user data, convert the data into one or more user historical activity record as described with respect to block 312. These user historical activity records will be used as indicia of user activity and submitted as part of an authentication request 124 to the profile based authentication service 126.
• Indicia of user activity may include a broad range of potential values. Table 1 enumerates some potential indicia values:
• Client Device History Determining if the client device used is one of client devices commonly used by the user
• Table 1 Exemplary User Indicia [0071] Table 1 is not intended to be an exhaustive list of user indicia.
• User indicia may come from third parties, such as credit checks. User indicia may be provided via interfaces to other information systems.
• the profile based authentication service 126 receives the authentication request 124, and proceeds to analyze the authentication request 124. Analysis may comprise identifying a correlation model corresponding to the authentication request 124. The identified correlation model will then specify user historical activity records to retrieve from data store 1 10. The correlation model will then determine if the user indicia in the authentication request 124 is similar to the retrieved user historical activity records. In some embodiments, a correlation model will identify content patterns, for example comparing the degree of profanity in the user indicia in the authentication request 124 to historical patterns.
• a correlation model will identify usage patterns, for example determining if a credit card payment is made immediately after browsing a web site when in contrast the user historically views the same web site at least a dozen times prior to committing to a purchase.
• the correlation model could track behavioral patterns where the user updates a social network record only during lunch time.
• Analysis may work with an arbitrary subset of user historical activity records as stored. Accordingly, the analysis may compares results from different correlation models before making a final determination of correlation.
• the correlation model may identify the degree of correlation, for example in the form of a similarity score, and will determine whether the similarity score exceeds a particular threshold.
• the correlation model may indicate that confidence in a particular determination is insufficient and will make no determination. For example, analysis may determine that the correlation model has insufficient user historical activity records to make a determination.
• Thresholds for whether correlation is sufficiently high to warrant authentication may differ based on the information system making the authentication request. Financial transactions and personal information may require high thresholds. Alternatively, general web sites may require relatively low thresholds. Thresholds may vary according to the scope of interaction of the user. For example, a per transaction authentication may have a lower threshold than a per session authentication. Similarly a per session authentication may have a lower threshold than an interaction that spans multiple sessions. Different vertical applications may have different thresholds. For example, a medical information system may have a higher threshold than an entertainment application.
• Analysis results may be shared in many different ways.
• a common scenario may be to send a message indicating either authentication, or an error message indicating either insufficient data or rejecting authentication.
• the analysis results may be accessed directly through an exposed application programming interface ("API").
• API application programming interface
• the analysis results may be aggregated into a single similarity score and exported for use by other applications or scenarios. For example, a contest web site may determine that it is 70% confident that a user is who the user claims to be. Based on the 70% confidence value, it may limit contest prizes to lower values than if it had 100% confidence.
• adverse events may be determined either substantively in real time, for example if the information system making an authentication request is an interactive system. Alternatively, adverse events may be determined in batch, for example in collecting disputed charge records which are to be presented to the user in a monthly bill. [0078] Once an adverse event is identified, the adverse event is handled in block 322. Just as adverse events may be determined in real time or alternatively in batch, the adverse event may be handled in real time or alternatively in batch as well. [0079] Real time handling of adverse events may include shutting the user out of the system, or providing a modal dialog box requiring the user to proffer alternative user credentials. For less critical scenarios, adverse events may simply be captured, and notification sent to the user via electronic mail, text messaging, or other forms of asynchronous communications.
• the profile based authentication service 126 may receive a request to correct a correlation model. Correlation models may be refined, or may be replaced. For example, if the profile based authentication service 126 determines that there is a high degree of false positives where authentication is rejected, but the unknown user is able to proffer correct alternative credentials, the correlation model may be marked as flawed or subject to correction, refinement or replacement in block 326.
• a service 128 may perform data mining on the user profiles, the user identity fingerprints, or both.
• a service 128 determines the desired data and performs a data query against the user profiles, the user identity fingerprints, or both.
• the data query may be in the context of some external correlation model.
• the query may retrieve pre-generated identity fingerprints corresponding to a time period. Alternatively, the query may request new user identity fingerprints to be generated dynamically with the most recent data.
• the service 128 applies an external correlation model to determine patterns of users corresponding t ⁇ ⁇ ° + ⁇ - ⁇ - 1 ⁇ ,.
• patterns may relate to the users themselves, such as in identifying popular products purchased.
• the patterns may relate to the historical user activity such as identifying the most common scenarios that authentication requests failed (e.g. in threat assessment).
• the external correlation model results may be analyzed to detect errors in the multi-factor identity fingerprinting system, thereby providing a sort of debug facility.
• Figure 4 illustrates an exemplary application of multi-factor identity fingerprinting 400. Specifically, Figure 4 illustrates loading existing user profile information and applying multi-factor identity fingerprinting for mobile device multimedia content requests on mobile devices 400 in a Service Delivery Gateway ("SDG”) 402 and Third Party Billing Gateway (“3PG") 404 infrastructure.
• SDG Service Delivery Gateway
• 3PG Third Party Billing Gateway
• WCDMA (“WCDMA") cellular provider.
• Third party content providers 406 wish to serve paid content 408 to users using mobile devices 410 over the WCDMA provider.
• One possible configuration for a WCDMA network to support third party data services is to use an SDG 402 to interface with Data Services 412. When content is served by the SDG 402, billing is handled by the 3PG 404.
• Sources may include pre-existing business intelligence sources 416 such as credit scores and default rates, billing information 418 for cellular subscriptions, and prepay information 420 for prepay cellular customers. Information from these user information sources 414 may be loaded into the data services layer 412 which is optionally filtered via a privacy engine 422.
• the information from the user information sources 414 is loaded via an extract transform and loading routine ("ETL") 424 as informed by a ETL Model 426 and then converted into profiles for storage into data store 428.
• ETL extract transform and loading routine
• the ETL Model 426 may be comprised of a data model and several rules and constraints.
• the SDG 402 may perform authentications via profiling service 430. Specifically, an unknown user 410 makes a content request of a third party content provider 406.
• the SDG 402 may have a local profiling client or may directly perform authentication by accessing the profiling service 430.
• the profiling service will access records via data store 428. According to one or more correlation models 432, the profiling service 430 will return a message indicating whether to authenticate, to reject, or whether there is insufficient information to make a determination.
• the SDG 402 and the third party content provider 406 will serve the requested content 408 to user 410, and third party content provider 406 will have the 3PG 404 bill the user 410 as authenticated by SDG 402.
• the third party content provider 406 will reject the request.
• the third party content provider 406 may generate a report or send a notification to the account owner of the failed authentication.
• Security fingerprints may be generated for customer users either on the basis of a profile or independent of a profile. They are served to fingerprint requestors based on authentication protocols.
• Figure 5 illustrates an exemplary security fingerprint logical architecture 500, to support security fingerprint functions.
• the fingerprint database 502 is the main database of the Security
• the Fingerprinting platform contains configuration and metadata for Security Fingerprint applications, monitoring data of processes, input data acquired from enterprise customer data sources and the data mining result data. All historical data is kept until a configurable time period ends, for example 10 years.
• the fingerprint database 502 may be a relational database, or may be a non-SQL or columnar database.
• the fingerprint database 502 receives data from enterprise data assets via the information integration layer 504.
• the information integration layer 504 performs data acquisition, federation and integration tasks. It supports multi-format, multi-protocol, real time, near real time, or batch modes for data integration. In general, the information integration layer 504 performs any tasks to manage the extraction, transformation and loading, and the time of loading of fingerprint relevant data from enterprise data assets.
• the fingerprint analytical application 506 is the main application which performs the processing logic and/or rules for fingerprints and the fingerprint generation algorithms. It also contains the controller module for process management and scheduling. Fingerprint logic and/or rules are further discussed with respect to Figure 6 and 7.
• the fingerprint services application 508 performs the serving of fingerprints based on a request and may also perform the serving of data from fingerprints based on a query. Note that alternatively, that a fingerprint may be queried independent of the fingerprint services application 508.
• the fingerprint services application 508 is the services layer to expose the security fingerprint data. While there are a set of business services based on the data, there are also multiple customer users and/or requestors facing proxy services created to satisfy business purposes, policy, contract terms and conditions.
• a fingerprint may be based on a profile, may support different fingerprint modes, and may utilize various security and/or encryption protocols.
• the customer profile console 510 is the customer facing user interface for managing profile data and/or privacy information use preferences.
• Fingerprint requestors maintain accounts to allow them to receive fingerprints according to subscription and security constraints.
• the fingerprint requestor console 512 is the self-help interface for fingerprint requestors to manage accounts and/or subscriptions, and to order services and request support.
• the quality of service (“QoS”) components 514 and security component 516 provide services orthogonal to other functionalities in the logical platform. Specifically, the quality of service (“QoS”) components 514 provide system administration tools to administer the fingerprint platform and applications to ensure the Security Fingerprint platform performs reliably within the specified Service Level Agreements ("SLAs").
• SLAs Service Level Agreements
• the security component 516 implements the security protocols used for authentication, authorization and other security functions. Specifically, the security component 516 ensures valid access to the fingerprint applications, services and the fingerprints database. It verifies users, fingerprint modes, roles, security credentials, authentication tokens, digital signatures, and the like to comply with information security standards and regulations.
• FIG. 6 illustrates one potential implementation 600 of the Security Fingerprint Logical Architecture 500 set forth in Figure 5.
• Data is first collected from a number of enterprise data assets 602.
• Enterprise data assets 602 include any data that may be collected by an enterprise, or acquired via third parties.
• enterprise data for a cellular operator might include the following: - Customer account information;
• Products and/or devices used by the customer e.g., cell phones, smart phone applications
• the data collection against the enterprise data assets 602 may be performed by a fingerprint data acquisition application 604.
• the fingerprint data acquisition application 604 may be implemented by invoking services of the information integration layer 504.
• the fingerprint data acquisition application 604 may receive data from the enterprise data assets 602 in a number of ways, including receiving periodic messages or real time data streams from various enterprise data assets 602, performing real time data lookups against the enterprise data assets 602, or performing periodic extract/transform/load ("ETL") functions against one or more enterprise data assets 602 potentially as a batch process.
• ETL extract/transform/load
• the fingerprint data acquisition application 604 may perform various analytics and/or transformations on the received data. Examples include, performing semantic analysis, executing a rules engine on the received data, queuing the messages in message queues, performing stream post-processing of received data streams, and federating data. Since there is likely to be a disparate number of enterprise data assets 602, the fingerprint data acquisition application 604 may maintain a corresponding number of data adapters to support data in various formats and via various protocols. In this way, the fingerprint data acquisition application 604 has an extensibility model to support an arbitrary data format and/or data protocol.
• the fingerprint data acquisition application 604 loads the security fingerprint database 606.
• the fingerprint database 606 may support various sub databases including input data which are the finalized loaded data from the fingerprint data acquisition application 604, configuration data which are profile and/or preference data from consumer users, and result data which are the results of analytic performed on data in the fingerprint database 606.
• Configuration data is entered via the customer profile console 608 as accessed by customer users 610. Customer users 610 enter profile and/or preference data via web or standalone application. Profile and preference data is discussed in more detail with respect to Figure 7.
• a fingerprint requestor 612 can either request a fingerprint via the fingerprint requestor console 614 or via the web service gateway 616.
• the fingerprint requestor console 614 is typically used to maintain a fingerprint requestors account and/or subscription, preference setting, service issue management (e.g., help desk) and web service catalog.
• the fingerprint requestor console 614 may also include utilities to request, receive and query fingerprints. If a fingerprint requestor is an automated process, a fingerprint may also be directly accessed via web service gateway 616, which provides a gateway to the fingerprint platform 600, specifically the fingerprint services application 618.
• the web service gateway 616 makes use of QoS 514 and security components 516 of the fingerprint logical architecture 500. In particular, the web service gateway 616 may provide services including, authorization, authentication, policy enforcement and routing and mediation.
• the fingerprint services application 618 corresponds to the fingerprint services application 508 in Figure 5, and provides the key interface for serving fingerprints, usually via a set of business services. To track and provide access to the web service gateway 616, the fingerprint services application 618 may support a web service registry and a set of proxy web services.
• fingerprint primitive operations embodied in the fingerprint analytic application 620.
• the fingerprint primitive operations provided by the fingerprint analytic application 620 may include operations to transform, analyze, version, maintain and query fingerprints. Fingerprint primitive operations are discussed in more detail with respect to Figures 7 and 8.
• system administrators 622 access the fingerprint security platform 600 via a system administration console 624.
• the system administration console 624 has access to all components enumerated in the fingerprint logical architecture 500 of Figure 5. In particular, it can access the fingerprint requestor console 614, the web service gateway 616, fingerprint services application 618 and fingerprint analytical application 620, as well as fingerprint database 606. Accordingly, the system administration console 624 may track all aspects of a fingerprint from creation through service to utilization of the served fingerprint for analysis.
• Figure 7 illustrates one potential security fingerprint platform implementation 700 of how a profile may be populated, and the internals of a fingerprint.
• a customer user 702 accesses the fingerprint platform 700 via a customer profile console 704.
• the customer profile console 704 may comprise a privacy editor and a profile editor.
• the privacy editor may be used to set various privacy settings in profile 706 and may be used to create, update and delete modes for the customer user's fingerprint or fingerprints.
• the profile editor may be used to edit non-privacy data fields, such as customer user identity fields (e.g., name, address, and identification numbers).
• the profile editor also specifies which enterprise data assets are to be encapsulated in a customer user's fingerprint or fingerprints. In this way, the customer user need not rely on security functions alone to prevent access to data; rather the customer may prevent the data from being included in a fingerprint in the first place.
• the customer profile console 704 also supports an audit trail function in order to track change histories in profiles. This audit trail may be used, for example, to detect unauthorized profile changes. ⁇ """ 00 + n ⁇ " the customer profile console 704 are persisted to security fingerprint database 708.
• Fingerprints are generated from the security fingerprint database 708 via the fingerprint services application 710.
• the fingerprint services application generates fingerprints in response to a fingerprint request via its business services components.
• the business services components may implement functionalities in terms of primitive fingerprint operations encapsulated in the fingerprint analytic application 712. Since fingerprints may be served over the web, the fingerprint services application 710 may maintain a web service registry and a proxy web services component.
• a fingerprint 714 once generated, will encapsulate some subset of data corresponding to the customer user 702.
• the encapsulated data may be encrypted and/or compressed.
• a fingerprint 714 may support zero, one or many fingerprint modes.
• a fingerprint mode specifies which portion of the data, which fingerprint requestors and under what circumstances the encapsulated data may be accessed via a query.
• a mode selector of the fingerprint 714 detects the identity of the fingerprint requestor, and the context, and selects a corresponding fingerprint mode.
• a default fingerprint mode may also be invoked. Accordingly, only a subset of the encapsulated data is exposed to the query 716.
• the query is then executed on the exposed data, and data returned to a fingerprint requestor console 718 or querying application, and thereby to the fingerprint requestor 720.
• a fingerprint requestor may simply retrieve a fingerprint 714 for querying independent of the fingerprint implementation 700.
• FIG. 8 illustrates the interaction 800 of example fingerprint services applications, primitive fingerprint operations, and the underlying data from the enterprise data assets.
• Primitive fingerprint operations may operate on fingerprints, on fingerprint data in the fingerprint database 802, and/or on other data such as data directly accessed from enterprise data assets 804.
• the direct access of enterprise data assets 804 may be effected via the information integration layer 504 services as described with respect to Figure 5.
• Primitive fingerprint operations may support the following classifications of operations:
• a fingerprint can be uniquely identified.
• a fingerprint may be cloned such that its encapsulated data and modes are replicated, except for a unique fingerprint identifier. Copy/replicate primitive operations may perform this replication.
• a compare function may determine whether two fingerprints are identical except for the unique fingerprint identifier.
• a similarity score function may determine the degree that two fingerprints are similar. Since the functions do not expose data to a fingerprint requestor, the encapsulated data remains secure, while allowing a fingerprint requestor to perform a comparison.
• Fingerprints may be merged or split.
• a set of customer users may decide to be treated as one entity, for example four cellular users in the same family may simply wish to have a single fingerprint for the entire family.
• the four fingerprints corresponding to the four family individuals may be merged into a single fingerprint.
• the family members may opt to merge replicated copies of their fingerprints, thereby preserving an option to maintain a personal fingerprint.
• the opposite of merging is splitting. For example, consider a fingerprint that captures data for a husband and wife. The husband and wife may seek to have individual fingerprints if they are receiving product recommendations for each other. Accordingly, a split operation will review the underlying encapsulated data, identify compatible clusters of data, and separate the fingerprints into two fingerprints, with a best guess of which data belongs to which spouse.
• ⁇ receive new underlying data. They may be merged or split. If a fingerprint requestor desires to know that a fingerprint has current data, a version function may be invoked to report a version number and/or a creation date. An out of date fingerprint may be refreshed by invoking an update function. In this way, a large fingerprint need not be regenerated from scratch. Rather a delta of underlying data need only be added. A fingerprint may be locked from update as well. For example, if a fingerprint requestor is only authorized to receive fingerprints for a period of time, a lock would help ensure that over time, the fingerprint's data was less useful due to data staleness.
• Fingerprints may be queried.
• a query primitive operation executes a requery against a fingerprint's encapsulated data subject to the presently set mode and any additional filters.
• the set mode primitive operation is used by the mode selector to set the mode to be used for a particular query.
• the set filter primitive operation may be used to add additional query conditions independent of a mode. Additional primitive operations to enumerate modes and filters may be specified as well.
• Applications 814, 816 and 818 are exemplary fingerprint services applications.
• An Alzheimer's application 814 is described with respect to Figure 9.
• a matchmaking application 816 is described with respect to Figure 10.
• a fingerprint merging application is described with respect to Figure 1 1.
• the security fingerprint architecture supports a wide range of use cases.
• FIG 9 illustrates an application 900 of fingerprint services for tracking user behaviors.
• One example application is tracking user behaviors for future retrieval with a long term memory loss or memory impaired patient.
• a patient such as an Alzheimer's patient who cannot remember much of their past.
• Such a patient might record events of their life as it happens. Much of this data is personal, and should be restricted with privacy settings. But the patient may not be capable of maintaining such settings. Accordingly, a security fingerprint may provide a solution.
• a security fingerprint is generated from the fingerprint database. Specifically, the fingerprint encapsulates data from a query on the fingerprint database. Note that the amount of data may be large, and it may be overly time consuming to regenerate the fingerprint. Accordingly, the memory augmentation application may invoke the version function of the versioning primitive operations 810 to find out the time of the last update and then perform an update function also of the versioning primitive operations 810. In this way, the delta between the last update need only be queried.
• the mode selector determines the identity of the querying fingerprint requestor and selects a mode.
• V ⁇ a — r, ⁇ requestor is the patient him or herself, all encapsulated data is made available via the set mode function of the querying operations 812. If the querying fingerprint requestor is a spouse, then a smaller subset of the encapsulated data based on the spouse's interest may made available by setting the fingerprint mode to a mode specific to the spouse. If the querying fingerprint requestor is a caregiver, then a mode specific to providing medical data may be set via the set mode function as well.
• FIG 10 illustrates a matchmaking application using security fingerprints.
• Matchmaking is the introduction of two or more people who are compatible. Contexts of compatibility may include determining whether two individuals are complementary enough for a potential date, professional conferences where individuals with similar professional interests are matched, or service providers whose skills are matched to customer's needs.
• a matchmaking application may automatically detect whether the two customer users are compatible with each other to the degree that they may want to introduce each other.
• Fingerprint A is resident on cell phone A of customer user A.
• Fingerprint B is resident on cell phone B of customer user B.
• Cell phone A may have an application that automatically checks for compatible proximate customer users. Specifically, when cell phone A is proximate to cell phone B, it can query customer user B's fingerprint. Upon receiving the customer user B's fingerprint, cell phone A may run a similarity score function from the replication/identity primitive operations 806. If the similarity score function exceeds a predetermined threshold, the application may query customer user B's fingerprint for data in common with customer user A via a query operation using query primitive operations 812. The application may then make an alert populated with the data returned from the query. Since the similarity score is against an amount of data larger than what customer user B would have made public, the comparison is more likely to be more accurate. Furthermore, the alert can be customized to specify what features customer user A and customer user B have most in common.
• a security fingerprint can represent an individual or a set of individuals.
• Figure 1 1 illustrates a scenario where two spouses seek to merge their fingerprints together.
• a husband seeks to identify vacation spots for himself and his spouse.
• his personal fingerprint A only reflects his personal preferences.
• the husband copies his personal fingerprint using the copy/replication primitive operations 806.
• the user retrieves his spouse's fingerprint B and merges it with the copy of his fingerprint A using the merge function of aggregation primitive operations 808.
• the husband uses the merged fingerprint to retrieve alerts from an advertising engine specific to vacation spots.
• the advertising server is able to surface a set of vacations that the husband is more likely to purchase - even though the vacations may be less compatible with this personal profile.
• the fingerprint services application is able to account for the desires of other customer users that impact the buying decision.
• Authentication is the verification that a user matches a claimed identity. In the case of profile based authentication, as described above, verification occurs when a user's behavior matches a security fingerprint to a predetermined confidence threshold.
• authentication need not be performed solely by consulting a security fingerprint.
• third party data sources that may supplement a security fingerprint, but need not be stored in a security fingerprint.
• a user's social network site such as a FacebookTM site may contain information about a user that is too bulky or redundant to a security fingerprint.
• a security fingerprint need not store this information.
• third party data sources may supplement an authentication process.
• the calculation of a confidence score may include the following steps: (1) determine the factors used by the confidence score, (2) query the security fingerprint for factors available from the security fingerprint, (3) identify third party data sources corresponding to the remaining factors, (4) query the third party data sources for those remaining factors, (5) report an error if not all factors from third party data sources are not available, and (6) calculate and return a confidence score if all factors from third party data sources are available.
• Some confidence scores may conditionally use third party data sources. For example, if a first confidence score is calculated from a security fingerprint, a second confidence score that makes use of third party data sources may be calculated, only if the first confidence score fails to meet a predetermined threshold. In this way, third party data sources may be used to provide an alternative authentication method.
• a first confidence score may specify that some subset out of a set of specified factors be available in order to calculate the confidence score. In this way, authentication may proceed if one or more third party data sources are unavailable.
• Third party data sources may also be used as input for a profile for a security fingerprint.
• a user context may be determined, for example location from a user's cell phone geolocation facility and date time reading from the cell phone clock.
• a user's geolocation at a particular data time stamp may be stored in a profile, as well as used a present context to be compared to a security fingerprint.
• User input may be from other input devices, connected via BluetoothTM or other connectivity methods.
• a pacemaker or other medical device reading may be interfaced to a computing device and the data used for storage in a profile for a security fingerprint.
• One example of how third party data sources give rise to a business specific use case is the collection of medical data.
• an insurance company may verify that a claimant was in fact at the scene of an accident that purported to have caused an injury.
• an insurance company or a treating doctor may verify that a person is in fact compliant with a therapy or specified medical course of action, such as not smoking, taking insulin, and dietary modifications as part of a diabetes treatment plan.
• security fingerprints may be aggregated according to a target profile, the target profile being a set of range values for behavioral factors of interest as tracked by profiles. In this way, individuals may be grouped together and targeted, for example for advertising.
• aggregation may be triggered on a condition. Specifically, the following may be performed: (1) receive a target profile, (2) specify a condition when aggregation is to be performed, (3) upon receiving notification that the condition has been satisfied (a) query a security profile data store for security fingerprint satisfying the target profile, and (b) aggregate the retrieved security fingerprint into a single security fingerprint.
• conditional aggregation may be targeted users who are attending a soccer game at a stadium.
• the target profile may be for home team soccer fans, but only fans who are present at the stadium are aggregated. In this way, the aggregated security fingerprint may be targeted for advertisements of soccer merchandise only available to those physically at the stadium, for example in the stadium vendor booths.
• Another application may be the conditional aggregation of the security fingerprints of a family.
• a target profile may be for a husband and wife and their children. When members of a family are proximate to each other, perhaps as determined by geolocation information from their cell phones, the respective security fingerprints of those proximate family members are aggregated. When those family members are no longer proximate, the security fingerprint is disaggregated. In this way, the aggregated security fingerprint may be targeted for advertising or recommendations for family members who are proximate to each other, such as restaurant recommendations targeting only the interests of the family members who are together at that point in time.
• Security fingerprints may be easily aggregated, since aggregation is simply the combination of stored factors, perhaps with the removal of redundant data. Security fingerprints may also be easily disaggregated, provided that some indicator, such as a flag, is associated with a factor indicating the original security fingerprint it was associated with.
• a security fingerprint needs to be disaggregated without any indicia of its constituent security fingerprints. For example, consider a husband and wife who have a single security fingerprint for the both of them. However, the two get divorced, and desire their own respective security fingerprint. If the single security fingerprint was originally generated for both husband and wife, then there is no original husband security fingerprint and no original wife security fingerprint. Accordingly, two security fingerprints will have to be generated by heuristics identifying one or the other.
• Aggregation of a security fingerprint may also be based on a heuristic. For example, an individual person may first have a pay as you go cellular plan. Later that individual becomes a cellular subscriber. There may be a first security fingerprint as a pay as you go customer and a second security fingerprint as a subscriber. The cellular company may wish to extract a single fingerprint based on the two original fingerprints. However, the decision to combine the fingerprints is based on a heuristic.
• Aggregation and disaggregation heuristics are associated with confidence levels of accuracy. Accordingly, use of a security fingerprint may be conditioned on the confidence level of the heuristic. Specifically, (1) a security fingerprint is generated by a heuristic, (2) a confidence level corresponding to the heuristic is associated with the generated security fingerprint, (3) a query is made against the generated security fingerprint, the query associated with a confidence level, and (4) if the query confidence level is satisfied by the security fingerprint, the queried data is retrieved, but if the query confidence level is not satisfied by the security fingerprint, and error is reported. In this way, false confidence in security fingerprints generated by heuristics may be guarded against. For example, a security fingerprint generated by a low confidence heuristic might be used for advertising, but not for authentication.

Landscapes

• Engineering & Computer Science (AREA)
• Business, Economics & Management (AREA)
• Accounting & Taxation (AREA)
• Finance (AREA)
• Marketing (AREA)
• Computer Hardware Design (AREA)
• Strategic Management (AREA)
• Physics & Mathematics (AREA)
• General Business, Economics & Management (AREA)
• General Physics & Mathematics (AREA)
• Development Economics (AREA)
• Theoretical Computer Science (AREA)
• Power Engineering (AREA)
• Economics (AREA)
• Computer Security & Cryptography (AREA)
• Computing Systems (AREA)
• General Engineering & Computer Science (AREA)
• Computer Networks & Wireless Communication (AREA)
• Signal Processing (AREA)
• Telephonic Communication Services (AREA)
• Collating Specific Patterns (AREA)
• Management, Administration, Business Operations System, And Electronic Commerce (AREA)

Applications Claiming Priority (2)

Publications (2)

Family

ID=50278673

Family Applications (1)

Country Status (3)

Families Citing this family (6)

Family Cites Families (12)

• 2013
  • 2013-09-12 CN CN201380047486.8A patent/CN104704521B/zh active Active
  • 2013-09-12 EP EP13837114.1A patent/EP2896005A4/de not_active Withdrawn
  • 2013-09-12 WO PCT/US2013/059479 patent/WO2014043360A1/en unknown

Also Published As

Similar Documents

Legal Events