© 2005-2006 RAF Technology Inc. A portion of the disclosure of this patent document contains material that is subject to copyright protection. The copyright owner has no objection to the facsimile reproduction by anyone of the patent document or the patent disclosure, as it appears in the Patent and Trademark Office patent file or records, but otherwise reserves all copyright rights whatsoever. 37 CFR §1.71 (d).
This disclosure pertains to methods and systems for leveraging private data to authenticate the identity of a user to a target service without disclosing the private data.
Online verification of an individual comprises two parts: establishing the existence of an identity that is allowed to perform the transaction, and establishing that the present user has that identity. A common method for establishing the existence and content of the online identity claimed by the user is to ask the user for personal information items that are presumed to be known by him and by the system but by no one else. These items fall into two broad (and somewhat overlapping) categories: what we call herein “in-wallet” and “out-of-wallet” data.
In-wallet data is data you know about yourself or can find quickly and easily, for example by consulting cards carried “in your wallet”. Name, address, telephone number, Social Security number (SSN), driver's license number (DLN), and birth date are examples of in-wallet data. Mortgage holder, last transaction amount to a merchant or government agency, and car registration number are examples of out-of-wallet data. Often in- and out-of-wallet data are used sequentially to authenticate a person. In-wallet data is used to determine which identity is claimed, and then out-of-wallet data (presumably harder for a fraudster to obtain) is used to establish that the user is that person.
The difficulty with this approach is that for it to work, the “private” data of the user (e.g. out-of-wallet data) must be revealed to the authenticator so the authenticator can compare it with its file of authenticating information. This has two problems. First, the user must give the authenticator his private data and second, the authenticator had to get that data from some other source who gathered the data in the first place. Thus, as currently constructed, such a system only works if the authenticator already has a compendium of “private” data against which to compare the user's information.
Almost always that data comes from a third party who gathered the data. There are therefore at least three parties who have the “private” data—the user, the authenticator, and the third party provider. Worse yet, much of the so-called “private” data was gathered by the third party from “public” sources—phone books, credit reports, and the like. If the third party isn't the entity that gathers the data directly from the user, then they are in the same situation as the authenticator—they have to get the data from yet another party. There is no telling how many hands the “private” data passes through prior to its being used for authentication.
This presents a contradiction: to be useful for authentication, the user's data must be “private” (otherwise spoofers can get at it), but it also must be available to the authenticator—which reduces its privacy. Given the large number of authenticators currently working, and the small number of available data items on a person, it becomes clear that the more a data item is used for authentication, the less reliable it is for authentication. The increasing reluctance of users to supply social security number or mother's maiden name (two of the most common data elements previously used for authentication) indicates that users have a (correct) feeling that each use of such information makes it less private, and hence increasingly exposes the user to identity theft. Users, therefore, are increasingly reluctant to put private information in the hands of authenticators, no matter how trustworthy they believe the authenticators to be.
We have observed, however, that there are certain entities that users routinely trust with their private information: those whose business it is to use the information for its original purpose. For example, no one thinks twice about his bank possessing his bank account numbers or knowing the balance in his checking account. Banks are supposed to have that information, and require it to perform the functions users expect of them. Utilities are assumed to hold user utility bill information, such as gas and electric bills. Wireless telecommunications carriers hold a mountain of cell phone user call data.
We propose to leverage the reliability of private data held by trusted entities to provide secure, reliable authentication of a user without exposing that private data to others, or even to the authenticator. Additional aspects and advantages will be apparent from the following detailed description of preferred embodiments, which proceeds with reference to the accompanying drawings.
Embodiments described below may be useful both for initially establishing the identity of a user and for extending that initial identification to subsequent transactions. (The concept of extending identification to subsequent transactions alludes to the following. If a user provides identifying information (say social security number, driver's license number, name and address) to someone in order to get a user name and password, subsequent use of that user name and password extends the original identifying information to the transaction for which the user name and password are used. Indeed, such things as digital certificates and passwords are only as secure as the original identifying information used to obtain them).
Heretofore, it has been presumed that for an authenticator to use private data for authentication, he had to possess that information. One advantage of several embodiments described herein is to make that possession unnecessary. The present patent application proposes systems and methods of using private information without revealing that private information to anyone beyond those who already had that information because it is material to an existing business relationship with the user. The authenticator at no time has direct access to the user's private information, i.e., he cannot read it, and hence that information isn't “cheapened” through use.
The proposed methods and systems generally function as follows: A user wishes to be authenticated for some electronic transaction. The authenticator wants to establish (1) that there exists an identity capable of performing the transaction (that is, both able to and, if appropriate, authorized to perform it) and (2) that the user is that person. There exists a third party who, as part of his normal business relationship with the user, gathers private information on the user. An example would be a utility company who knows the amount of a user's most recent electric bill because they sent it to him. I use the utility bill as an example, but this invention does not depend on the precise nature of the private data—only that it is highly unlikely to be known by anyone other than the user and an entity that gathers it directly from the user (or issues it to him).
In this illustration, the authenticator asks the user for the amount of his most recent electric bill. Instead of providing that information “in the clear”, the user encrypts (or otherwise obfuscates) the data before providing it to the authenticator.
The authenticator then sends the encrypted information on to the utility company or other trusted entity. The utility company can then do one of two things, either of which fulfills the requirements of an embodiment of the present invention: They can decrypt the information and compare it with their records, or they can directly compare the encrypted submission with a corresponding record encrypted in the same manner. It does not matter for the purpose of this invention which method is used.
In one alternative embodiment, the utility can provide files of personal data in encrypted form to the authenticator, who then compares its contents to the data provided by the user. The user data must have been encrypted using the same algorithm as that used by the utility (otherwise comparison of the two bits of encrypted data won't work). The user data must have been encrypted by the user without the data having been first supplied in clear form to the authenticator. The important consideration is that the private data in a readable form remain only in the hands of the user and the data possessor (here, the utility). This prevents the data from being cheapened with use, because unlike with current methods, no other party gets access to the data in readable form.
It is not necessary that the data used for authentication be merchant data, or indeed any particular kind of data. For example, in an alternative embodiment, a user provides text strings of identifying information (either in person or online) and, once identified to the service (the exact means for this is not relevant to the invention), provides a data set to the service that can later be used in the manner of this invention. In this case, the system is used to extend the identification provided at registration to all transactions performed by the user, but making it much easier on the user to do so and not requiring that the initial identifying information be presented over and over to multiple authenticators in a readable, unencrypted form (which would cheapen the identifying information).
The data set can contain almost anything—children's names, nonsense words, dates. The data can be provided to the service in clear-text form or, for additional security, provided only in encrypted form. In the latter case, only the user ever knows what the information strings contain.
What particular type of data is supplied in this manner is not critical. Later, when the user seeks to be authenticated by a site using the service, he provides one of the secret strings (in encrypted form) to the site. The site passes the encrypted string on to the service, which authenticates it as having come from the user. Alternatively, the service could provide a list of strings (in encrypted form) and the authenticator could do its own comparison. Because the strings are encrypted, they are not revealed to the authenticator. The encryption method should be sufficiently thorough so as to not be subject to a black box attack.
The attached drawing FIGURE depicts one possible embodiment, in which the numbered steps represent the sequence of events occurring during authentication. With reference to the enclosed drawing, a transactor (user) sends an access request to a target service (step 1). The target service then sends to a verification service a request for authentication of the transactor (step 2). (In this example, the verification service is labeled “RAF Verification Service” after the name of the assignee of this invention, RAF Technology Inc.) The verification service then issues a knowledge-based authentication challenge (“KBA challenge”) to the transactor (step 3), which may include an embedded encryption agent, such as a Java object, Active-X control, etc. The KBA challenge requests that the transactor provide certain specified private data to authenticate the transactor. The transactor's response to the KBA challenge is encrypted at the transactor's computer, by the encryption agent or otherwise, and submitted to the verification service (step 4).
The verification service then retrieves from its database encrypted identification elements stored in its database and corresponding to the purported identity of the transactor (step 5) and compares them against the encrypted challenge response submitted by the transactor (step 6). In the embodiment shown, the identification elements are provided to the authentication service in encrypted form by original data processors (in the drawing there are three, all labeled “data provider”) running a compatible encryption process, and the identification elements are then stored by the authentication service in encrypted form. In another embodiment, the comparison is done against a hash table rather than actual data.
The verification service then returns the results of the comparison to the target vendor, e.g., an indication of the correlation (or lack thereof) between the transactor's response to the KBA challenge and the encrypted personal information accessible to the verification service (step 7). The indication may comprise a correlation score calculated by the verification service. Alternatively, it may comprise a pass/fail indication based on predetermined minimum correlation criteria of the authenticator or minimum correlation criteria that are preselected by the target service either at the time of requesting authentication or beforehand. Upon receiving an indication of a positive verification or authentication from the verification service, the target service provider then grants to the transactor access to the requested service (step 8).
Skilled persons will appreciate that the embodiment depicted in
In much of the discussion above, the identifying data has been used to establish who the user is. It is also possible to use the data to establish that the user is capable and/or authorized to perform the requested transaction. For example, the user could supply in encrypted form a request to his bank to certify him for a particular dollar amount. The bank could then release to the authenticator a dollar amount authorized for the transaction. No information beyond that on the user's creditworthiness need be released. This particular example is similar to the existing protocol when a merchant sends in an authorization request to a credit card company—he specifies the amount of the bill and the card company authorizes the transaction (or not) for that amount. Thus, in the prior art, there is no veiling of the token (the merchant has the credit card, with its number, in his hand). In accordance with the present invention, by contrast, the merchant would send an encrypted token whose contents were unknown to him, along with the dollar amount.
Authorization also can be established by the user sending an encrypted message via the authenticator to, say, his employer, and the employer could then send back an authorization for the user to perform the desired transaction.
None of the above need necessarily take place over a computer network. All the authorization methods described above would also work were the user present in person before the authenticator. As a practical matter, however, the data storage and retrieval steps performed by the authenticator are facilitated by the use of digital computers and high-speed data storage and retrieval systems.
PKI (because it uses a private key registered to the user) authenticates that a particular message comes from the user. This is useful for people who want to go to the trouble of getting and registering public/private keys, but takes place after the initial authentication of the user. It is not useful for initially establishing user identity, while the proposed system is useful for both—establishing initial identity and establishing that the user is the same person who was previously authorized.
Many current systems encrypt a password supplied by a user and compare that encrypted password with a file of encrypted passwords. This is used to allow a transaction or other access to the system. This use somewhat resembles the invention, but the differences are several: The use of a password (encrypted or otherwise) does not in itself provide any identity information on the user. That must be provided at the time the password is issued—the best the password can do is extend that identifying information to the current transaction. The embodiments described herein, on the other hand, establish that the user holds a particular identity, and they do so without any tokens (like passwords) previously established between the user and the system he wishes to access.
Another advantage of the present invention is to enable extension of knowledge of a previously-established identity to subsequent transactions without decreasing the value of personal information for subsequent authentication. This is true whether or not the previous identity was established using the methods of the present invention.
It will be obvious to those having skill in the art that many changes may be made to the details of the above-described embodiments without departing from the underlying principles of the invention. The scope of the present invention should, therefore, be determined only by the following claims.
This application claims priority from U.S. Provisional Application No. 60/665,273 filed Mar. 24, 2005 and incorporated herein by this reference.
Number | Name | Date | Kind |
---|---|---|---|
4890323 | Beker et al. | Dec 1989 | A |
5259025 | Monroe et al. | Nov 1993 | A |
5265159 | Kung | Nov 1993 | A |
5375244 | McNair | Dec 1994 | A |
5436970 | Ray et al. | Jul 1995 | A |
5613004 | Cooperman et al. | Mar 1997 | A |
5687236 | Moskowitz et al. | Nov 1997 | A |
5742685 | Berson et al. | Apr 1998 | A |
5774551 | Wu et al. | Jun 1998 | A |
5832497 | Taylor | Nov 1998 | A |
5838814 | Moore | Nov 1998 | A |
5841886 | Rhoads | Nov 1998 | A |
5841888 | Setlak | Nov 1998 | A |
5844817 | Lobley et al. | Dec 1998 | A |
5850442 | Muftic | Dec 1998 | A |
5870499 | Bender et al. | Feb 1999 | A |
5870723 | Pare, Jr. et al. | Feb 1999 | A |
5872848 | Romney et al. | Feb 1999 | A |
5883810 | Franklin et al. | Mar 1999 | A |
5884270 | Walker et al. | Mar 1999 | A |
5893090 | Friedman et al. | Apr 1999 | A |
5910988 | Ballard | Jun 1999 | A |
5940507 | Cane et al. | Aug 1999 | A |
5973731 | Schwab | Oct 1999 | A |
5987440 | O'Neil et al. | Nov 1999 | A |
5991408 | Pearson et al. | Nov 1999 | A |
6014650 | Zampese | Jan 2000 | A |
6026166 | LeBourgeois | Feb 2000 | A |
6026491 | Hiles | Feb 2000 | A |
6039248 | Park et al. | Mar 2000 | A |
6073242 | Hardy et al. | Jun 2000 | A |
6085322 | Romney et al. | Jul 2000 | A |
6111954 | Rhoads | Aug 2000 | A |
6118872 | Kashima et al. | Sep 2000 | A |
6122403 | Rhoads | Sep 2000 | A |
6125349 | Maher | Sep 2000 | A |
6175923 | Bailey | Jan 2001 | B1 |
6202151 | Musgrave et al. | Mar 2001 | B1 |
6230148 | Pare, Jr. et al. | May 2001 | B1 |
6263447 | French et al. | Jul 2001 | B1 |
6282658 | French et al. | Aug 2001 | B2 |
6363376 | Wiens et al. | Mar 2002 | B1 |
6363485 | Adams et al. | Mar 2002 | B1 |
6385620 | Kurzius et al. | May 2002 | B1 |
6496936 | French et al. | Dec 2002 | B1 |
6532459 | Berson | Mar 2003 | B1 |
6618806 | Brown et al. | Sep 2003 | B1 |
6643648 | Ross et al. | Nov 2003 | B1 |
6714944 | Shapiro et al. | Mar 2004 | B1 |
6836845 | Lennie et al. | Dec 2004 | B1 |
6845453 | Scheidt et al. | Jan 2005 | B2 |
6904407 | Ritzel | Jun 2005 | B2 |
6959394 | Brickell et al. | Oct 2005 | B1 |
7039951 | Chaudhari et al. | May 2006 | B1 |
7086085 | Brown et al. | Aug 2006 | B1 |
7131009 | Scheidt et al. | Oct 2006 | B2 |
7318048 | King | Jan 2008 | B1 |
7343623 | Ross | Mar 2008 | B2 |
20020091945 | Ross | Jul 2002 | A1 |
20030037248 | Launchbury et al. | Feb 2003 | A1 |
20030105762 | McCaffrey et al. | Jun 2003 | A1 |
20040039917 | Ross | Feb 2004 | A1 |
20040064454 | Ross et al. | Apr 2004 | A1 |
20040139098 | Margolus et al. | Jul 2004 | A1 |
20050187873 | Labrou et al. | Aug 2005 | A1 |
20050238207 | Tavares | Oct 2005 | A1 |
20080127305 | Ross | May 2008 | A1 |
Number | Date | Country |
---|---|---|
WO9922330 | May 1999 | WO |
WO03103215 | Dec 2003 | WO |
Number | Date | Country | |
---|---|---|---|
60665273 | Mar 2005 | US |