Systems and methods to securely generate shared keys

CROSS-REFERENCE TO RELATED APPLICATIONS

0001This application claims the benefit of and is a continuation-in-part of U.S. patent application Ser. No. 11/093,954 filed on Mar. 30, 2005 (“Deploying and Provisioning Wireless Handheld Devices” by Herb Little and Michael K. Brown) which claims the benefit of U.S. Provisional Application No. 60/559,092 filed Apr. 2, 2004 and U.S. Provisional Application No. 60/559,646 filed Apr. 5, 2004. This application also claims the benefit of U.S. Provisional Application No. 60/566,790 filed on Apr. 30, 2004. The contents of each of these applications are incorporated herein by reference.

BACKGROUND

00021. Technical Field

0003This application relates to an apparatus and method of establishing an authentic and secure relationship between two messaging systems to exchange data. More specifically this application describes an apparatus and method of establishing an authentic relationship between a wireless handheld device (“mobile device”) and a message center or host system using password-based authentication methods. The apparatus and method described herein is applicable to land-line environments as well as wireless environments.

00042. Background of the Art

0005There are several strong password-based encryption mechanisms currently known in the computer industry. Some of these implementations include Encrypted Key Exchange (EKE), Password Derived Moduli (PDM), and Simple Password-authenticated Exponential Key Exchange (SPEKE). These mechanisms are limited in their implementations and have not addressed the need of mobile devices. Additionally these mechanisms do not address the need of implementing perfect forward secrecy, such that if a security breach does take place all previous messages exchanged remain secure. A protocol provides perfect forward secrecy if past session keys are not compromised even if the long-term keys are compromised. (See, e.g., Menezes et al., Handbook of Applied Cryptography, 1996, p. 496). Perfect forward secrecy, also known as break-backward protection, means that all previous securely exchanged messages should remain secure despite anything that happens in the future.

BRIEF DESCRIPTION OF THE DRAWINGS

0006<figref idref="DRAWINGS">FIG. 1</figref> shows a block diagram of a first exemplary communication system, between a fixed and a wireless system.

0007<figref idref="DRAWINGS">FIG. 2</figref> shows a block diagram of a second exemplary communication system, between two wireless systems.

0008<figref idref="DRAWINGS">FIG. 3</figref> shows a block diagram of a third exemplary communication system, between two fixed systems.

0009<figref idref="DRAWINGS">FIG. 4</figref> shows a message exchange diagram of an exemplary set of data exchanges for implementing the communication system of <figref idref="DRAWINGS">FIG. 1</figref> where a user is the initiator of the data exchange.

0010<figref idref="DRAWINGS">FIG. 5</figref> shows a message exchange diagram of an exemplary set of data exchanges for implementing the communication system of <figref idref="DRAWINGS">FIG. 1</figref> where a service provider is the initiator of the data exchange.

0011<figref idref="DRAWINGS">FIG. 6</figref> shows a data flow diagram of the steps within the user software for carrying out the steps in <figref idref="DRAWINGS">FIG. 4</figref> where the user is the initiator of the key exchange.

0012<figref idref="DRAWINGS">FIG. 7</figref> shows a data flow diagram of the steps within the service software for carrying out the steps in <figref idref="DRAWINGS">FIG. 4</figref> where the user is the initiator of the key exchange.

0013<figref idref="DRAWINGS">FIG. 8</figref> shows a data flow diagram of the steps within the service user for a re-key sequence when regenerating another key in the environment illustrated in <figref idref="DRAWINGS">FIGS. 1, 2</figref> and <b>3</b>.

0014<figref idref="DRAWINGS">FIG. 9</figref> shows a data flow diagram of the steps needed within the service provider for a re-key sequence when regenerating another key in the environment illustrated in <figref idref="DRAWINGS">FIGS. 1, 2</figref> and <b>3</b>.

DETAILED DESCRIPTION

0015Referring to <figref idref="DRAWINGS">FIG. 1</figref>, there is shown a block diagram of a first exemplary communication system, between a fixed and a wireless system. This overview diagram shows a network environment where the invention is used. The diagram shows an exemplary embodiment of the invention and focuses on a network topology that includes a mobile device that is wireless. In this Figure there are systems offering services <b>20</b> and <b>22</b>, and systems using the services <b>30</b> and <b>32</b>. Between the service offering (also referred to herein as a service provider) and the service user are one or more networks and one or more connections to enable the flow of data between the two systems.

0016Turning now to <figref idref="DRAWINGS">FIG. 1</figref> the service offering <b>20</b> or <b>22</b> can be many possible computers offering services to users. For one skilled in the art some well known service providers could be computers on the Internet within an Internet Service Provider (ISP) or Application Service Provider (ASP) office. The service offering <b>20</b> and <b>22</b> can also be one or more computers running within a private or public company, like a bank, stock broker, insurance broker or some other service-oriented company. The service offering <b>20</b> or <b>22</b> may also be run as part of a cluster of computers operating world-wide, making up a Universal Description, Discovery and Integration Cluster (UDDI cluster). The common element in all these service offerings <b>20</b> and <b>22</b> is that these service offerings <b>20</b> and <b>22</b> need to establish a secure data channel with a user. In the case of UDDI the secure relationship might be needed to exchange private service listings, or even to allow UDDI to proxy a service offering.

0017The mobile devices and the service hosts may be addressed in a variety of different ways. In some embodiments, they may be addressed with IP (internet protocol) addresses. In other embodiments, the host system may be addressed by an e-mail address. In yet another embodiment, the destination address may be an e-mail address of a user of the mobile device within the host system.

0018One skilled in the art will appreciate that the user of the service <b>30</b> and <b>32</b> might be a mobile hyper-text transfer protocol (HTTP) browser, a mobile wireless application protocol (WAP) browser, a proprietary transmission control protocol/internet protocol (TCP/IP) based application or some proprietary corporate solution. In this field there are new methods being developed quickly, including for example the new Java 2 Micro Edition (J2ME) solution for small wireless mobile devices, like cell phones and personal digital assistants (PDAs). For devices that use J2ME the option of attaching and downloading software through a service offering is becoming commonplace. Similarly the service offering <b>20</b> and <b>22</b> can be based on an HTTP web server solution, a Java Enterprise solution, a wireless markup language (WML) based service offering or some proprietary service solution created for a specific purpose.

0019It will be appreciated that mobile systems and host systems referred to herein can each comprise one or more respective memories (e.g., containing processing instructions) and one or more respective processing units, such as those conventionally known, e.g., general purpose processing units and/or special purpose processing units such as application specific integrated circuits (ASICs) and field programmable gate arrays (FPGAs), wherein the processing units can be configured (e.g., programmed with suitable software and/or firmware instructions, and/or produced with specialized hardware circuits) to carry out the approaches described herein. Each of such systems can also include any suitable interface(s), such as those conventionally known, which can operate in conjunction with a respective processing unit(s) to facilitate communication with other systems.

0020The end-points in the communication path are coupled through one or more data networks that allow the exchange of data, voice, video, music, photographs or any other digital media that can be exchanged through a data communications channel. The two main networks included in this illustration are a Wide Area Network (WAN) <b>26</b>, the most common one being the Internet, and a wireless network <b>28</b>. The wireless network 28 could be a GSM/GPRS network, a CDMA/1XRTT network, a CDMA2000 network, a 3^rdGeneration network like EDGE or UMTS or many other public wireless networks soon to be available. In an exemplary system these networks are coupled using links <b>24</b> like ISDN, T1, Ethernet (land-line and 802.11), Frame Relay, ATM, ADSL or some other high speed Internet connection to the host service <b>10</b><i>b</i>. As greater amounts of data are being exchanged it is clear that security needs to be improved and made more foolproof to hackers and eavesdroppers. The invention works with these existing data communication paths to provide advanced password-based authentication. This level of security provides greater confidence that the recipient of any communicated data is exactly the entity you expect. One embodiment for a data communication path <b>36</b> is illustrated between a Host System service offering <b>22</b> and a user of the service on a mobile device <b>32</b>. Another embodiment for a data communication path <b>40</b> is illustrated between a UDDI service offering <b>20</b> and a user of the service on a mobile device <b>30</b>.

0021In one embodiment the host system service offering <b>22</b> has an out-of-band communication <b>34</b> (i.e., a communication over any suitable secure channel) with a user of a mobile device <b>32</b>. The out-of-band communication path <b>34</b> is used for exchanging a shared secret, avoiding the insecure path that is desired to be made secure. Since the UDDI service cloud provides some level of security, a UDDI service cloud might be used to locate the service and receive the out-of-band shared secret with the final destination service. The following are a few examples of out-of-band communication paths <b>34</b> and <b>38</b>: <ul id="ul0001" list-style="none"><li id="ul0001-0001" num="0000"><ul id="ul0002" list-style="none"><li id="ul0002-0001" num="0022"> (a) The mobile device user <b>30</b> or <b>32</b> and an operator at the host system <b>20</b> or <b>22</b>, establish a phone call with each other to exchange the shared secret. The secret is then entered into each system and used in the process of creating an encryption key. </li><li id="ul0002-0002" num="0023"> (b) The mobile device user <b>30</b> or <b>32</b> connects to a secure web site <b>20</b> or <b>22</b>, either wirelessly or over a wired network and requests a key. The key is received and manually entered into the mobile device <b>30</b> or <b>32</b>. The host system <b>20</b> or <b>22</b> could receive the key automatically from the web server, or it could also be manually entered. In some embodiments, a record is automatically generated after a shared secret was requested. </li><li id="ul0002-0003" num="0024"> (c) The user of the mobile device <b>30</b> or <b>32</b> makes the request for the service and the shared secret is e-mailed by the host system <b>20</b> or <b>22</b> to their corporate mailbox that is known to be in a secure area. The user retrieves the shared secret from their electronic mailbox and manually enters it into the mobile device <b>30</b> or <b>32</b>. </li><li id="ul0002-0004" num="0025"> (d) The user of the mobile device <b>30</b> or <b>32</b> makes the request for the service and an operator at the service <b>20</b> or <b>22</b> generates a shared secret and it is given to a specified person who is known to be trusted and secure. This person could be a secretary or administrator of a given group; ideally it is someone that can confirm the identity of the user making the request. This trusted person then gives the shared secret to the final user of the mobile device <b>30</b> or <b>32</b> and it is manually entered into the mobile device <b>30</b> or <b>32</b>. </li></ul></li></ul>

0026This short list shows that there are many ways to authentically give a shared secret to a mobile device <b>20</b><i>b </i>user. The common property of these exemplary out-of-band communications <b>34</b> and <b>38</b> is that some level of authentication should be built in or assumed in the choice made. This authenticated communication path should preferably be different than the non-authenticated data communication path.

0027Once the shared secret is exchanged the next step in creating a secure communication path can take place <b>36</b> and <b>40</b>. One of the better-known methods for creating a secure and authenticated link is using a strong password-based encryption method like SPEKE. SPEKE is a cryptographic method for knowledge-based authentication that leverages and protects easy-to-remember passwords—i.e. shared secrets. SPEKE is the simplest of the known strong password methods. It is a password-authenticated Diffie-Hellman exchange, where the password forms the base or “generator” of the exchange. (In standard Diffie-Hellman, the base is usually a fixed public number.) Once the communication path through the WAN <b>26</b> and wireless network <b>28</b> has been made secure, the re-key sequence can be initiated. The re-key sequence allows for the generation of a new set of keys after a predetermined number of weeks or months. During this re-key sequence the advanced use of long-term encryption keys allows for the implementation of perfect forward secrecy. Once the authentication secret (shared secret) is used to create a secure path, it can be reused to create new keys at later dates. By using this invention the re-keying operation does not compromise previous keys and all previous conversations remain secret into the future.

0028Turning to <figref idref="DRAWINGS">FIG. 2</figref> there is shown a block diagram of an exemplary communication system, between two wireless systems, according to an embodiment of the present invention. In this embodiment, a secure path can be created between two mobile devices. In this embodiment mobile device <b>1</b><b>46</b> and mobile device <b>2</b><b>48</b> exchange a secret and are able to establish a common key using that shared secret. The out-of-band conversation <b>50</b> could take place via a phone call between the two parties, or a face-to-face meeting, or using one of the other methods already outlined or any other suitable method. Once the secret is shared, it can be manually typed into the mobile devices <b>46</b> and <b>48</b>, and one station can initiate the exchange of messages to create a common master security key. This type of embodiment might be commonly used for private point-to-point e-mail conversations. It could also be used for point-to-point secure instant messaging data exchanges. In advanced usage mobile device <b>1</b><b>46</b>, who is providing the service, might be running a web server on the mobile device <b>46</b> and offering some form of secure service offering that is also mobile.

0029Turning to <figref idref="DRAWINGS">FIG. 3</figref> there is shown a block diagram of an exemplary communication system, between two fixed systems, according to an embodiment of the present invention. In this embodiment the communication takes place between two Host Systems <b>60</b> and <b>62</b>. In this illustration the service offering <b>60</b> and the service consumer <b>62</b> have an out-of-band conversation <b>66</b> and exchange a secret key. As described already this out-of-band communication could be a phone call, a communication via a browser with a secure SSL connection to generate and retrieve the key, or some other suitable communication such as provided earlier. Once the secret is exchanged an encryption key can be generated using strong password-based key generation methods like SPEKE. The communication path to exchange the key in this illustration could be over a WAN network like the Internet 26, or through an internal Intranet 64, or other suitable communication path such as or similar to an 802.11 or Bluetooth link. In these latter examples the service consumer <b>62</b> might be running a laptop or palmtop and already have a limited access to the Intranet, but greater security is desired or required. It is well known in the art that 802.11b lacks the robust security requirements requested by most large computer departments inside companies. This embodiment illustrates that the invention can be used to provide the option of perfect forward secrecy when using a password-based authentication mechanism. Once suitable messages are exchanged to create the master key, the data communication path <b>68</b> can be used to exchange all forms of data secretly with high security.

0030Turning to <figref idref="DRAWINGS">FIG. 4</figref> there is shown a message exchange diagram showing an exemplary set of data exchanges for generating and verifying a master key, where the user is the initiator of the data exchange. This illustration shows exemplary steps and message exchanges between a service consumer <b>100</b> (user) and a service provider <b>102</b>. In this illustration one end of the connection is considered a service consumer or user <b>100</b>, and has been given the label system A. The other end of the connection is considered the service provider (also referred to as a service offering) or host system <b>102</b>, and has been given the label system B. In this example the user <b>100</b> initiates the exchange of data to create a secure connection. Between System A and System B is a message exchange over one or more data communication networks such as illustrated in <figref idref="DRAWINGS">FIG. 1</figref>. Similarly as shown in <figref idref="DRAWINGS">FIGS. 1, 2</figref> and <b>3</b>, the user could be a mobile device <b>30</b>, <b>32</b> or <b>48</b>, or a Host System <b>62</b>. Likewise the service provider could be a mobile device <b>46</b> or a Host System <b>20</b>, <b>22</b> or <b>60</b>.

0031As shown at step <b>104</b>, the user <b>100</b> contacts a known service provider <b>102</b> through one of the methods already described for out-of-band communication or through another suitable method to exchange a shared secret. This service provider <b>102</b> wants to facilitate this exchange and issues a secret password or simple, easy to remember password strings (step <b>106</b>). Through this mechanism a shared secret is generated and exchanged between the two parties. The user <b>100</b> receives and saves the secret to assist in encryption key generation. Alternatively, the service provider <b>102</b> can receive a secret password (shared secret) from the user <b>100</b>. In either case, the service provider saves the shared secret in relation to this user.

0032After exchange of the shared secret, the user <b>100</b> then initiates (in this example) steps of generating key pairs (step <b>108</b>) and transferring key information to the service provider (step <b>110</b>). In particular, the user <b>100</b> generates a long-term encryption key pair at step <b>108</b>, i.e., the public and private parts of an encryption key. A short-term authentication key pair is also generated at step <b>108</b> by the user <b>100</b>. This short-term key pair is referred to as an authentication key pair in this example because it is generated using the shared secret as discussed further below.

0033Once the user's short-term and long-term key pairs are generated , the public keys thereof are transmitted at step <b>110</b> to the service provider <b>102</b> to further generate the final master key (also referred to as a master secret). This transfer can take place over an insecure link, as only the host system <b>102</b> that issued the shared secret can understand and use the short-term authentication key to generate the master key. Once the user's public keys are received by the service provider (step <b>112</b>), the user is verified, and the shared secret for that user is recalled <b>112</b>. Once the user is verified and the shared secret for the user is recalled, the service provider <b>102</b> proceeds to generate its own short-term authentication key pair using the shared secret (step <b>114</b>). The service provider <b>102</b> also generates its own long-term encryption key pair (step <b>114</b>). Using the public keys generated by the user <b>100</b> and using the shared secret, the service provider <b>102</b> generates a master encryption key (or master secret) as shown at step <b>116</b>. The shared secret provides the authentication necessary to trust the information exchanged. The service provider's short-term public authentication key, the service provider's long-term public encryption key, and a key confirmation value that has been calculated by the service provider using the newly generated master encryption key, and some known string, are sent to the user (step <b>116</b>).

0034The user receives the information (step <b>118</b>) sent from the service provider <b>102</b> including the service provider's short-term and long-term public keys and generates the user's own master key (step <b>120</b>). With this master key the user verifies the key confirmation value (step <b>120</b>). In this example, the key confirmation value could be the hash of the master key and the name of the service or some other known string, agreed upon by the user and the service provider. If the key confirmation value does not verify, the master key created by the user <b>100</b> is not trusted, and it is assumed that someone is trying to compromise the connection. If the master encryption key generated by the user <b>100</b> seems valid the user then sends a final key confirmation value back to the service provider (step <b>122</b>). The service provider receives the message, verifies the key confirmation value and marks the user as ready to go (step <b>124</b>). This allows full data exchange to take place from the service provider's point of view (step <b>128</b>). On the user side, once the verification message is sent there would be a slight pause in transmission but then full data exchange can begin (step <b>126</b>).

0035Transmissions may comprise e-mail messages, HTTP (hyptertext transfer protocol)-based traffic, such as XML (extensible markup language), WML (wireless markup language), etc., or other types of traffic.

0036In some embodiments, the host system is capable of sending a data payload in a message sent to the mobile device before the final confirmation value is sent to it from the mobile device. The payload in this message may be a service book entry that defines the host service at the host system. In some embodiments the service book entry may be a UDDI service entry that defines attributes of a host service at the host system being accessed.

0037It will be appreciated that the long-term encryption key pair generated by a first party (e.g., a user) as described herein is an example of, more generally, a first key pair, wherein the public key portion and the private key portion thereof can be referred to as a first public key and a first private key. Similarly, the short-term authentication key pair (also referred to as a short-term encryption key pair) generated by the first party (e.g., the user) as described herein is an example of, more generally, a second key pair, wherein the public key portion and the private key portion thereof can be referred to as a second public key and a second private key. Also, the long-term encryption key pair generated by a second party (e.g., a service provider) as described herein is an example of, more generally, a third key pair, wherein the public key portion and the private key portion thereof can be referred to as a third public key and a third private key. Similarly, the short-term authentication (or encryption) key pair generated by the second party (e.g., the service provider) as described herein is an example of, more generally, a fourth key pair, wherein the public key portion and the private key portion thereof can be referred to as a fourth public key and a fourth private key. The first party that generates the first and second key pairs could be a user, such as described in the example above, or a service provider, such as described in the example below.

0038Turning to <figref idref="DRAWINGS">FIG. 5</figref> there is shown a message exchange diagram showing an exemplary set of data exchanges for generating and verifying a master key, where the service provider is the initiator of the data exchange. The steps within <figref idref="DRAWINGS">FIG. 5</figref> substantially correspond to the steps within <figref idref="DRAWINGS">FIG. 4</figref>, except the service provider takes the first step. This example highlights that either the user or the service provider can be the initiator of the data exchange. In this illustration one end of the connection is considered the user <b>100</b>, and is labeled system A—service consumer. The other end of the connection is considered the service <b>102</b>, and is labeled system B—Service Provider. Between System A <b>100</b> and System B <b>102</b> is a message exchange over one or more data communication networks <b>26</b>, <b>28</b> and <b>64</b> such as illustrated in <figref idref="DRAWINGS">FIGS. 1, 2</figref> and <b>3</b>. Similarly as shown in <figref idref="DRAWINGS">FIGS. 1, 2</figref> and <b>3</b>, the user could be a mobile device <b>30</b>, <b>32</b> or <b>48</b>, or a Host System <b>20</b>, <b>22</b>, <b>46</b> or <b>60</b>.

0039As shown at steps <b>200</b>/<b>202</b>, the service provider <b>102</b> contacts the user <b>100</b> (in this example) to exchange a shared secret. Alternatively, the user could initiate this communication. It is contemplated that an administrator within a host company <b>102</b> might contact the user <b>100</b> and inform the user that the user has to perform some action with the shared secret being provided. Using any suitable method selected from the extensive list of out-of-band communications already provided, or some other suitable method, the shared secret is generated and exchanged (steps <b>200</b> and <b>202</b>). The User component receives and saves the shared secret to assist in encryption key generation. Alternatively, the service provider <b>102</b> can receive a secret password (shared secret) from the user <b>100</b>. In either case, the service provider saves the shared secret in relation to this user.

0040After exchange of the shared secret, the service provider <b>102</b> can initiate (in this example) steps of generating key pairs (step <b>204</b>) and transferring key information to the user <b>100</b> (step <b>206</b>). In particular, the service provider <b>102</b> generates a short-term authentication key pair and a long-term encryption key pair (step <b>204</b>). This corresponds to step <b>108</b> in <figref idref="DRAWINGS">FIG. 4</figref>.

0041Once the service provider's short-term and long-term key pairs are generated, the public keys thereof are transmitted to the user (step <b>206</b>) to further generate the final master key (also referred to as a master secret). This transfer can take place over an insecure link as only the owner of the shared secret would be able to understand and use the short-term authentication key to generate the master key. The service provider's public keys are received by the user, and it checks memory to verify the service creation is expected and that it has a shared secret saved in memory (step <b>208</b>). The user recalls the shared secret for that service provider <b>102</b> and generates a short-term authentication key pair using the shared secret (step <b>210</b>). The user also generates a long-term encryption key pair (step <b>210</b>). Using the public keys generated and sent by the service provider <b>102</b> and using the shared secret, the user <b>100</b> generates a master encryption key (or master secret) as shown at step <b>212</b>. After generating the master key the user <b>100</b> also generates a key confirmation value by combining a known string(i.e., known to itself and the service offering) with the master key (step <b>212</b>). The user's short-term public authentication key the long-term public encryption key, and the key confirmation value are sent to the service provider (step <b>212</b>).

0042The service provider receives the user's public keys and key confirmation value and verifies the sender of the information (step <b>214</b>), and also recalls the shared secret for this user. With the received public key values of the user, the service provider recalls its own saved private key values for this user (step <b>214</b>). Using the received public keys of the user and the service provider's saved private keys, the service provider can now generate a master key (step <b>216</b>). After generating the master key, the service provider <b>102</b> verifies the key confirmation value by calculating its own key confirmation value, using the known string and the newly created master key, and comparing it against the received key confirmation value (step <b>216</b>). If the key confirmation value does not verify, the created master key is not trusted, and it is assumed that someone is trying to compromise the connection. If the key confirmation value does verify, the master encryption key is considered valid and the service provider <b>102</b> sends a final key confirmation value back to the user (step <b>218</b>). The user receives the message (step <b>220</b>), verifies the key confirmation value, and marks the service provider as ready to go (step <b>220</b>). This allows full data exchange to take place from the user's point of view (step <b>222</b>). On the service offering side, once the verification message is sent there would be a pause in transmission, but then full data exchange can begin (step <b>224</b>). In most cases it will be the user that initiates the first data exchange; so having the confirmation sent to the user does have some advantages.

0043Transmissions may comprise e-mail messages, HTTP (hypertext transfer protocol)-based traffic, such as XML (extensible markup language), WML (wireless markup language), etc., or other types of traffic.

0044<figref idref="DRAWINGS">FIG. 6</figref> is a data flow diagram of exemplary steps carried out by the user (e.g., within the user software) for carrying out the exemplary approach shown in <figref idref="DRAWINGS">FIG. 4</figref>, when the user is the initiator of the key exchange. The first step occurs when the user discovers a new service and wants to access it (step <b>300</b>). This might occur via a UDDI-like service, through a corporate Intranet service, through browsing the world-wide web, through conversation with a friend or through a phone call. Once the service and user have connected, they exchange a shared secret ‘s’ that only the two of them know (step <b>302</b>). Exemplary methods for this exchange have been described in detail already. This shared secret ‘s’ will be used later like a PIN (Personal Identification Number) to authenticate the user and the service to each other. When the user is ready to access the service, the user (e.g., in software) generates a long-term key pair for the requested service (step <b>304</b>). This long-term key pair is one of the key values used during all for future re-keying operations. For all of the mathematical calculations in the remainder of this application, we assume that all parties involved in the transactions have agreed beforehand on a group G, of size order(G), and an element g of G such that q=order(g) is a large prime number. G and g may be publicly known, i.e., they do not need to be kept secret. Exemplary mathematical calculations to create key values are as follows (using a SPEKE method), and while the exemplary calculations shown below utilize a multiplicative group, it will be apparent that suitable calculations could be carried out using an additive group:

0045Pick A Long-Term Key Pair (e.g., by User) <ul id="ul0003" list-style="none"><li id="ul0003-0001" num="0000"><ul id="ul0004" list-style="none"><li id="ul0004-0001" num="0046"> Pick Random a, 1<a<q−1; </li><li id="ul0004-0002" num="0047"> Calculate A=g^a; </li><li id="ul0004-0003" num="0048"> If A=1, keep choosing different a's until A< >1. </li></ul></li></ul>

0049The value ‘A’ is the user's long-term public key (or, more generally, first public key), and the value ‘a’ is the user's long-term private key (or, more generally, first private key).

0050In this example, the selected number ‘a’ is greater than 1 and less than the prime number q−1. Once the private key is selected (i.e. ‘a’) and the public key is generated (i.e. ‘A’), the private key ‘a’ is stored securely, and the public key ‘A’ is eventually transmitted to the service provider.

0051A short-term authentication key pair is also generated by the user based on the shared secret ‘s’ (step <b>306</b>). Using a similar calculation following a SPEKE key generation method, exemplary mathematical calculations for this step are (using, e.g., the same assumptions for q and for ‘a’ (as now applied to x) as before):

0052Pick A Short-Term Authentication Key Pair (e.g., by User) <ul id="ul0005" list-style="none"><li id="ul0005-0001" num="0000"><ul id="ul0006" list-style="none"><li id="ul0006-0001" num="0053"> Pick Random x, 1<x<q−1; </li><li id="ul0006-0002" num="0054"> Calculate X=s^x; </li><li id="ul0006-0003" num="0055"> If X=1 keep choosing new x's until X< >1. </li></ul></li></ul>

0056The value ‘X’ is the user's short-term public key (or, more generally, second public key), and the value ‘x’ is the user's short-term private key (or, more generally, second private key). The value ‘s’ is the shared secret.

0057In this example, the selection of ‘x’ is between <b>1</b> and the prime number q−1. The user software then sends the public key values ‘A’ and ‘X’ to the service offering (service provider) as shown at step <b>308</b>. This step proceeds to (A) where the service offering receives the values and performs additional calculations, shown in <figref idref="DRAWINGS">FIG. 7</figref>. Once the service offering has completed those calculations, it returns a similar pair of its own public key values ‘B’ and ‘Y’ with a key confirmation value to the user for verification (step <b>312</b>) as discussed further below in connection with <figref idref="DRAWINGS">FIG. 7</figref>. This is shown as input (B) in <figref idref="DRAWINGS">FIG. 6</figref> coming from <figref idref="DRAWINGS">FIG. 7</figref>. At this point the user is able to use ‘B’ and ‘Y’ to create a master key using, e.g., advanced SPEKE calculations. By using both ‘B’ and ‘Y’ together to generate the master key, the encryption method allows for the implementation of perfect forward secrecy. This is seen more clearly in the re-key sequence shown later. An exemplary master key calculation is as follows:

0058Calculate Master Key (e.g., by User) <ul id="ul0007" list-style="none"><li id="ul0007-0001" num="0000"><ul id="ul0008" list-style="none"><li id="ul0008-0001" num="0059"> k1=Y^x; </li><li id="ul0008-0002" num="0060"> k2=B^a; </li><li id="ul0008-0003" num="0061"> check whether k1, k2!=0, 1, or order(G)−1; </li><li id="ul0008-0004" num="0062"> k=hash(k1 ∥ k2) where ∥ is a concatentation function. Here, ‘x’ is the user's short-term private authentication key (or, more generally, second private key), and ‘Y’ is the received short-term public authentication key of the service offering (or more generally, fourth public key). Also, ‘a’ is the user's long-term private encryption key (or, more generally, first private key), and ‘B’ is the received long-term public encryption key of the service offering (or, more generally, third public key). </li></ul></li></ul>

0063The value ‘k’ represents the master key that can be used for encrypting data between the user and the service. The value ‘k’ is a combination of the intermediate keys ‘k1’ (based on the short-term authentication keys) and ‘k2’ (based on the long-term encryption keys). A check can be made on the intermediate key values of k1 and k2 at step <b>314</b> to verify that these two values are not 0, 1 or order(G)−1; otherwise it could mean there is a security attack being attempted This attack would result if the key were being forced into a small subset of total possible keys. If the attacker sends an X=0 or Y=0, the communicating parties could get a resulting key value of 0. This quick check will ensure that an attack is not being staged. If however the value of k1 or k2 does fall into one of these small subset groups, the negotiation for a key can be aborted (step <b>316</b>).

0064If a subset attack is not detected, the master key ‘k’ can be used by the user to test the key confirmation value sent by the service offering (step <b>318</b>). One method for generating a key confirmation value is to hash the key with a known string such as the bytes in the public key “A”. An exemplary calculation to test key confirmation value would be:

0065Test Key Confirmation Value <ul id="ul0009" list-style="none"><li id="ul0009-0001" num="0000"><ul id="ul0010" list-style="none"><li id="ul0010-0001" num="0066"> “Received h_A”=h_A=hash (k ∥ bytes of public key“A”) where “received h_A” came from the service offering, and ‘k’ is the local master key. </li></ul></li></ul>

0067If the software's generated key confirmation value for ‘A’ does not match (step <b>320</b>) the received key confirmation value, then it is incorrect (step <b>322</b>). An incorrect key confirmation value could mean that a man-in-the-middle attack, or some other attack is being attempted. The operation will be aborted in this case (step <b>322</b>). If the two confirmation values match, then it is assumed that a fully secure link has been established (step <b>324</b>). The link is marked as valid and after a short delay will be used for communications (step <b>324</b>). Using the newly generated verification key, the user sends this value back to the service (step <b>326</b>). This follows back to <figref idref="DRAWINGS">FIG. 6</figref> following label (C). After a few moments pause, i.e., to ensure the confirmation is received by the service offering, the user can begin to exchange data (step <b>328</b>).

0068Any suitable encryption and decryption methods can be used to encrypt and decrypt messages using the master key, such as symmetric-key encryption/decryption methods like the Advanced Encryption Standard (AES) (Federal Information Processing Standards Publication 197, Nov. 26, 2001, National Institute of Standards and Technology).

0069<figref idref="DRAWINGS">FIG. 7</figref> is a data flow diagram of exemplary steps carried out by the service offering (e.g., within the service provider software) for carrying out the exemplary approach shown in <figref idref="DRAWINGS">FIG. 4</figref> when the user is the initiator of the key exchange as shown in <figref idref="DRAWINGS">FIG. 4</figref>. The process starts when a user contacts a service provider ‘out-of-band’ to exchange a shared secret (step <b>398</b>). This corresponds with step <b>302</b> in <figref idref="DRAWINGS">FIG. 6</figref> on the user's device. This out-of-band exchange has been discussed several times and also provides a level of authentication that the user and service are who they say they are. Once this exchange is complete, the user is free at any point in time to contact the service to begin the process. Once the user does contact the host service, shown with message (A) arriving from the user's flow chart in <figref idref="DRAWINGS">FIG. 6</figref>, the new user is verified (step <b>400</b>). Since a service provider might have tens or hundreds of users wanting to start using their service at any time, the service provider is passive until the user decides he wants to start the service. Even though a shared secret has been exchanged this may mean very little, and stale shared secrets might even be purged after some number of days if the user fails to connect over that time period. The arrival of the message allows the service provider to find the new user and verify that a shared secret exists (step <b>400</b>). In the message is the user's public short-term authentication key, which is based on the shared secret (step <b>400</b>). The message also contains the user's public long-term encryption key (step <b>400</b>), which can be used in the implementation to create perfect forward secrecy when re-key operations take place, <figref idref="DRAWINGS">FIGS. 7 and 8</figref>.

0070The service offering generates a long-term encryption key pair for this user, in a manner similar to the long-term encryption key-pair created by the user (step <b>402</b>). Exemplary mathematical calculations to create the service offering's long-term encryption key pair are as follows (e.g., using a SPEKE method):

0071Pick A Long-Term Key Pair (e.g., by Service Provider) <ul id="ul0011" list-style="none"><li id="ul0011-0001" num="0000"><ul id="ul0012" list-style="none"><li id="ul0012-0001" num="0072"> Pick Random b, 1<b<q−1; </li><li id="ul0012-0002" num="0073"> Calculate B=g^b</li><li id="ul0012-0003" num="0074"> If B=1, keep choosing different b's until B< >1. </li></ul></li></ul>

0075The value ‘B’ is the service offering's (service provider's) long-term public key (or more generally, third public key), and the value ‘b’ is the service offering's long-term private key (or, more generally, third private key).

0076In this example, the selected number ‘b’ is greater than 1 and less than the prime number q−1. Once the private key ‘b’ is selected and the public key ‘B’ is generated, the private key ‘b’ is stored securely, and the public key ‘B’ is eventually transmitted back to the user so he can use it in his calculations.

0077The service offering also generates a short-term authentication key pair based on the shared secret (step <b>404</b>). Using a similar calculation following a SPEKE key generation method, exemplary mathematics for this step are (using, e.g., the same assumptions for q and for x (as now applied to y) as before):

0078Pick A Short-Term Authentication Key Pair (e.g., by Service Provider) <ul id="ul0013" list-style="none"><li id="ul0013-0001" num="0000"><ul id="ul0014" list-style="none"><li id="ul0014-0001" num="0079"> Pick Random y, 1<y<q−1; </li><li id="ul0014-0002" num="0080"> Calculate Y=s^y; </li><li id="ul0014-0003" num="0081"> If Y=1, keep choosing y's until Y< >1. </li></ul></li></ul>

0082The value ‘Y’ is the service offering's (service provider's) public short-term authentication key (or, more generally, fourth public key), and the value ‘y’ is the service offering's private short-term authentication key (or, more generally, fourth private key).

0083In this example, the selection of ‘y’ is between <b>1</b> and the prime number q−1. The public key values ‘B’ and ‘Y’ will eventually be sent to the user to generate the user's own master key.

0084The service offering then uses the public keys ‘A’ and ‘X’ received from the user, and the private keys just calculated to generate a master key (step <b>406</b>). By using both ‘A’ and ‘X’ together to generate the master key the encryption method provides perfect forward secrecy. To provide perfect forward secrecy, if desired, the implementation can also use the private keys in the re-generation of subsequent keys during any re-key sequence. An exemplary master key calculation is as follows:

0085Calculate Master Key (e.g., by Service Provider) <ul id="ul0015" list-style="none"><li id="ul0015-0001" num="0000"><ul id="ul0016" list-style="none"><li id="ul0016-0001" num="0086"> k1=X^y; </li><li id="ul0016-0002" num="0087"> k2=A^b; </li><li id="ul0016-0003" num="0088"> check that k1, k2!=0, 1, or order(G)−1; </li><li id="ul0016-0004" num="0089"> k=hash(k1 ∥ k2). </li></ul></li></ul>

0090Here, ‘y’ is the service offering's short-term private encryption key (or, more generally, fourth private key), and ‘X’ is the received short-term public encryption key of the user (or, more generally, second public key). Also, ‘b’ is the service offering's long-term private key (or, more generally, third private key), and ‘A’ is the received long-term public encryption key of the user (or, more generally, first public key).

0091The value ‘k’ represents the master key generated by the service offering, and it is the same as the master key generated by the user. This master key can be used for encrypting data between the service and the user. The value ‘k’ is a combination of the intermediate keys ‘k1’ (based on the short-term authentication keys) and ‘k2’ (based on the long-term encryption keys). A check can be made on the intermediate key values of k1 and k2 at step <b>408</b> to verify that these two values are not 0, 1 or order(G)−1; otherwise it could mean there is a security attack being attempted. This attack would result if the key were being forced into a small subset of total possible keys. If the attacker sends an X=0 or Y=0 the communicating parties could get a resulting key value of 0. This quick check will ensure that an attack is not being staged. If however the value of k1 or k2 does fall into one of these small subset groups the negotiation for a key can be aborted (step <b>410</b>).

0092If a subset attack is not detected, the master key ‘k’ can be used by the service offering to test the key confirmation value sent by the user (step <b>416</b>). One method for generating a key confirmation value is to hash the key with a known string such as the bytes in the public key “B”. An exemplary calculation to test the string (key confirmation value) would be:

0093Test Key Confirmation Value <ul id="ul0017" list-style="none"><li id="ul0017-0001" num="0000"><ul id="ul0018" list-style="none"><li id="ul0018-0001" num="0094"> h_B=hash (k ∥ bytes of public key “B”). </li></ul></li></ul>

0095The service offering would then transmit the test string to the user so that it can verify that the master key generated by the user matches the master key created by the service offering. The service offering then sends the long-term public encryption key ‘B’, the short-term public authentication key ‘Y’ (or, fourth public key) and the verification string h_Bto the user (step <b>414</b>).

0096Once the user has generated its own master key ‘k’ it sends back a final key confirmation value to ensure the service offering knows that everything has worked correctly (C). This final step (C) is shown in <figref idref="DRAWINGS">FIG. 7</figref> as an input to the service offering at step <b>416</b>. If the key confirmation value was calculated based upon ‘A’ and sent to the service offering (step <b>416</b>), then this is what the test looks for (step <b>418</b>). If the key confirmation value does not match the expected value the operation is aborted (step <b>420</b>). If the key confirmation value is matched then it is assumed that a full two-way encrypted and secure data communication path exists (step <b>422</b>).

0000The Re-Key Data Flow Sequence

0097<figref idref="DRAWINGS">FIG. 8</figref> is a data flow diagram showing exemplary steps within the user (e.g., within software) for a re-key sequence when regenerating another key in the environment illustrated in <figref idref="DRAWINGS">FIGS. 1, 2</figref> and <b>3</b>. This procedure illustrates the utility of using the long-term encryption key to enable the implementation of perfect forward secrecy. The process starts when either the user or the service offering decide a new key is desired or required. For this example we will assume the host (service provider) is running an encryption key expiry timer. However, there are many other ways the encryption key might be re-generated. The user might decide that it is time for a new key, the user or service might have fears that someone has been trying to attack and determine the current key value. Whatever the case, a new key is desired or required, and a unique method, not based on the original shared secret, can be used to generate the new key.

0098As shown in the example of <figref idref="DRAWINGS">FIG. 8 a</figref> re-key request is received by the user, or the user decides to cut a new key (step <b>430</b>). Of course, step <b>430</b> could be executed by the service provider instead of the user. The user software generates a new short-term encryption key (step <b>432</b>). An exemplary mathematical calculation is based on SPEKE and uses the same sequence as shown before:

0099Pick A New Short-Term Encryption Key Pair (e.g., by User) <ul id="ul0019" list-style="none"><li id="ul0019-0001" num="0000"><ul id="ul0020" list-style="none"><li id="ul0020-0001" num="0100"> Pick Random x, 1<x<q−1; </li><li id="ul0020-0002" num="0101"> Calculate X=g^x; </li><li id="ul0020-0003" num="0102"> If X=1, keep choosing x's until X< >1. </li></ul></li></ul>

0103Here, ‘x’ is a “new” value generated for the user's short-term private encryption key. The value ‘x’ can be referred to either as an “encryption” key or as an “authentication” key (as was done previously) because the value ‘x’ contributes to both aspects. In this example, the selection of ‘x’ is between <b>1</b> and the prime number q−1. The user software then sends the newly generated public key value ‘X’ to the service provider <b>434</b>. This step proceeds to (D) where the service provider receives the value and performs additional calculations. Step (D) is taken into <figref idref="DRAWINGS">FIG. 9</figref> as input on the service provider side of the connection.

0104Once the service provider has completed those calculations with (D) shown in <figref idref="DRAWINGS">FIG. 9</figref>, it returns a similar new public encryption key ‘Y’ (discussed further below) with a key confirmation value for verification by the user (E). This is shown as input (E) in <figref idref="DRAWINGS">FIG. 8</figref>. At this point the user is able to use the service provider's new ‘Y’ key with the service provider's older long-term public ‘B’ key to create a master key following advanced SPEKE calculations, for example. By using both the existing ‘B’ and the new ‘Y’ together to generate the key, the encryption method can provide implementation of perfect forward secrecy. Perfect forward secrecy can be achieved because neither the existing ‘B’ nor the new ‘Y’ are based on the original shared secret, and the existing ‘B’ is combined with the new ‘Y’ to create a new key not directly based on the previous key. Additionally, the existing ‘B’ key carries some of the authentication generated with the original shared secret. Only an authenticated service user, that originally possessed the shared secret, would have been able to have the private key ‘b’ saved to disk. This is seen more clearly in the exemplary re-key mathematical calculation to create a new master key ‘k’:

0105Calculate Master Key (e.g., by User) <ul id="ul0021" list-style="none"><li id="ul0021-0001" num="0000"><ul id="ul0022" list-style="none"><li id="ul0022-0001" num="0106"> k1=Y^x; </li><li id="ul0022-0002" num="0107"> k2=B^a; </li><li id="ul0022-0003" num="0108"> check that k1, k2!=0, 1, or order(G)−1; </li><li id="ul0022-0004" num="0109"> k=hash(k1 ∥ k2). </li></ul></li></ul>

0110Here, ‘x’ is the user's new short-term private encryption key, and ‘Y’ is the new received short-term public encryption key generated by the service provider. The value ‘a’ is the user's existing long-term private encryption key, and ‘B’ is the service provider's existing long-term public encryption key.

0111The value ‘k’ represents the new master key that can be used for encrypting data between the user and the service provider. The value ‘k’ is a combination of the intermediate keys ‘k1’ (based on the short-term encryption key) and ‘k2’ (based on the long-term encryption keys). A check can be made on the intermediate key values of k1 and k2 (step <b>442</b>) to verify that these two values are not 0, 1 or order(G)−1; otherwise or it could mean there is a security attack being attempted (step <b>442</b>). If however the value of k1 or k2 does fall into one of these small subset groups the negotiation for a key can be aborted (step <b>444</b>).

0112If a subset attack is not detected, the new master key ‘k’ can be used to test the key confirmation value sent by the service offering (service provider) as shown at step <b>446</b>. One method for generating a key confirmation value is to hash the key with a known string like the bytes of the public key of “A”. The approach for calculating a key confirmation value can be the same as previously described. If the calculated key confirmation value does not match what was received (step <b>448</b>), the key is assumed to be in error (step <b>450</b>). An incorrect key confirmation value would mean that a man-in-the-middle attack, or some other attack is being attempted. Otherwise the user generates a final key confirmation value using the master key ‘k’ (step <b>452</b>). The key confirmation value is sent to the service provider (step <b>454</b>) as a final confirmation; as shown at point (F) in <figref idref="DRAWINGS">FIG. 8</figref>. Then after a short pause the new encryption key is used within the user software (step <b>456</b>). During a short period of time there is also a window where messages that were previously transmitted could arrive in. During this period of several minutes the old key is kept and tried if decryption errors occur (step <b>456</b>).

0113Turning now to <figref idref="DRAWINGS">FIG. 9</figref> this represents a data flow diagram of exemplary steps within the service provider for a re-key sequence when regenerating another key in the environment illustrated in <figref idref="DRAWINGS">FIGS. 1, 2</figref> and <b>3</b>. This procedure shows the utility of using the long-term encryption key in conjunction with the short-term encryption key for implementing perfect forward secrecy. In this embodiment we have assumed the user has started the process and has already created a new short-term encryption (or authentication) key pair as shown in <figref idref="DRAWINGS">FIG. 8</figref>. The arrival of the short-term public encryption key ‘X’ is shown as input (D). The public key is received and the user's configuration information is recalled and checked (step <b>460</b>). The service offering then generates a new short-term encryption key pair for use over the next segment of time (step <b>462</b>). Exemplary mathematics to create a new short-term encryption key is similar to what has been shown before, except the shared secret ‘s’ is not used.

0114Pick A New Short-Term Encryption Key Pair (e.g., by Service Provider) <ul id="ul0023" list-style="none"><li id="ul0023-0001" num="0000"><ul id="ul0024" list-style="none"><li id="ul0024-0001" num="0115"> Pick Random y, 1<y<q−1; </li><li id="ul0024-0002" num="0116"> Calculate Y=g^y; </li><li id="ul0024-0003" num="0117"> If Y=1, keep choosing random y's until Y< >1. </li></ul></li></ul>

0118In this example, the selection of ‘y’ is between 1 and the prime number q−1. The value ‘Y’ will eventually be sent to the user to generate a master key (step <b>472</b>).

0119After picking a new short-term encryption key pair, a master key is generated by the service provider using the value ‘X’ that was just received from the user and the newly generated value ‘y’. By using both ‘A’ and ‘X’ together to generate the key, the encryption method provides for perfect forward secrecy. An exemplary master key calculation is as follows:

0120Calculate Master Key (e.g., by Service Provider) <ul id="ul0025" list-style="none"><li id="ul0025-0001" num="0000"><ul id="ul0026" list-style="none"><li id="ul0026-0001" num="0121"> k1=X^y; </li><li id="ul0026-0002" num="0122"> k2=A^b; </li><li id="ul0026-0003" num="0123"> check that k1, k2!=0, 1, or order(G)−1; </li><li id="ul0026-0004" num="0124"> k=hash (k1 ∥ k2). </li></ul></li></ul>

0125Here, ‘y’ is the service provider's new short-term private encryption key, and ‘X’ is the new received short-term public encryption key generated by the user. The value ‘b’ is the service provider's existing long-term private encryption key, and ‘A’ is the user's existing long-term public encryption key.

0126The value ‘k’ represents the master key for the service offering (step <b>464</b>). This will be used for encrypting all data between the service offering and the user. The value ‘k’ is a combination of the intermediate keys ‘k1’ (based on the new short-term encryption keys) and ‘k2’ (based on the long-term encryption keys). The calculation of ‘k’ is not directly dependent on the original shared secret ‘s’, but the values ‘A’ and ‘b’ carry some of the authentication originally provided by ‘s’. A check can be made on the intermediate key values of k1 and k2 (step <b>466</b>) to verify that these two values are not 0, 1 or order(G)−1; otherwise it could mean there is a security attack being attempted. If k1 or k2 do fall into one of these small subset groups the negotiation for a key can be aborted (step <b>468</b>).

0127If a subset attack is not detected, the master key ‘k’ can be used to test the key confirmation value sent by the service offering (step <b>470</b>). One method for generating a key confirmation value is to hash the key with a known string like the bytes in the public key “B” (step <b>470</b>). This calculation can be similar to those already described. The service offering would then transmit its new short-term public encryption key ‘Y’ and the key confirmation value h_Bto the user (step <b>472</b>). This transfer of the key values and the key confirmation value is shown at transfer box (E) in <figref idref="DRAWINGS">FIG. 9</figref>.

0128Once the user has generated its own master key ‘k’, it sends back a final key confirmation value to ensure the service offering knows that everything has worked correctly (step <b>454</b> of <figref idref="DRAWINGS">FIG. 8</figref>) as shown at (F). This final step at (F) is shown in <figref idref="DRAWINGS">FIG. 9</figref> as an input to the service offering (step <b>474</b>). If the key confirmation value was calculated for ‘A’ and sent to the service offering (step <b>474</b>), then this is what the test looks for (step <b>476</b>). If the key confirmation value does not match the expected value the operation is aborted (step <b>478</b>). If the key confirmation value verifies, then it is assumed that a full two-way encrypted and secure data communication path exists (step <b>480</b>). The server keeps the previous key for several minutes just in case packets were on route during this new key generation stage (step <b>480</b>).

0129The above-described approach can be modified to utilize other exemplary calculations for generation (e.g., over-the-air) of secure keys. For example, a user (e.g., a mobile station such as mobile stations <b>30</b> or <b>32</b>) and a service provider (e.g., a host system such as host systems <b>20</b> or <b>22</b>) can first share a small secret key (e.g., cryptographically weak secret key), obtained through out-of-band methods such as those previously described, and then bootstrap to a larger cryptographically strong key. The bootstrapping is done in such a way that online and offline dictionary attacks are prevented.

0130For example, the exemplary protocols described below can allow the user (e.g., mobile station) and the service provider (e.g., host system) to negotiate cryptographic keys over the air, which is assumed to be an insecure channel. In this example, the first protocol is for the initial key exchange, and the second is for key regeneration (re-keying).

0131According to one example, the first protocol runs the SPEKE and Diffie-Hellman (DH) key exchange protocols in parallel. The resulting master key is a combination of the keys generated by the SPEKE and DH protocols. SPEKE combines a small shared secret with the DH key agreement protocol to negotiate a shared key, in such a way that precludes an offline dictionary attack. Specifically, SPEKE replaces the base point of the DH key agreement with the small shared secret.

0132The second protocol is the MQV technique (Menezes-Qu-Vanstone key agreement scheme) known in the art. An exemplary version of this technique is described in <i>Guide to Elliptic Curve Cryptography </i>by D. Hankerson, A. Menezes, and S. Vanstone (New York: Springer-Verlag, 2004); the MQV discussion of this reference is hereby incorporated herein by reference for the purpose of demonstrating the generally known state of the art of this technique.

0133In exemplary implementations, the user <b>100</b> (e.g., mobile station) and service provider (or host system) can share the following cryptosystem parameters: <ul id="ul0027" list-style="none"><li id="ul0027-0001" num="0000"><ul id="ul0028" list-style="none"><li id="ul0028-0001" num="0134"> F_q—a finite field of prime order q; </li><li id="ul0028-0002" num="0135"> E(F_q)—an elliptic curve over F_q, such as the NIST (National Institute of Standards and Technology) approved 512 bit random elliptic curve over F_q, which curve has a cofactor of one (see Federal Information Processing Standards Publication, “Digital Signature Standard (DSS)”, Jan. 27, 2000, Appendix 6, located at http://csrc.nist.gov/publications/fips/fips 186-2/fips186-2-change1.pdf); </li><li id="ul0028-0003" num="0136"> P—any point of E that generates a subgroup of E(F_q) of prime order r; </li><li id="ul0028-0004" num="0137"> xR—represents elliptic curve scalar multiplication, where x is the scalar and R is a point on E(F_q); and <maths id="MATH-US-00001" num="1"><math overflow="scroll"><mrow><mover><mi>R</mi><mi>_</mi></mover><mo>=</mo><mrow><mrow><mo>(</mo><mrow><mover><mi>x</mi><mi>_</mi></mover><mo>⁢</mo><mstyle><mtext> </mtext></mstyle><mo>⁢</mo><mi>mod</mi><mo>⁢</mo><mstyle><mtext> </mtext></mstyle><mo>⁢</mo><msup><mn>2</mn><mrow><mstyle><mtext> </mtext></mstyle><mo>⁢</mo><mrow><mo>⌈</mo><mfrac><mi>f</mi><mn>2</mn></mfrac><mo>⌉</mo></mrow></mrow></msup></mrow><mo>)</mo></mrow><mo>+</mo><msup><mn>2</mn><mrow><mstyle><mtext> </mtext></mstyle><mo>⁢</mo><mrow><mo>⌈</mo><mfrac><mi>f</mi><mn>2</mn></mfrac><mo>⌉</mo></mrow></mrow></msup></mrow></mrow></math></maths><br /> where x is the integer representation of the x-coordinate of the elliptic curve point R, and f is the bit length of r (i.e., f=└log₂r┘+1). </li></ul></li></ul>

0138The user <b>100</b> and host system <b>102</b> agree to use the above-noted parameters, or other suitable parameters, which can be either publicly known or private parameters. All mathematical operations can be done in the group E(F_q), except that some MQV operations can be done in Z_r(where Z_rrefers to integers modulo r).

0000Initial Key Exchange

0139As noted above, the first exemplary protocol utilizes SPEKE and DH together for initial key exchange. In describing the example below, reference will be made to the previously described flow chart of <figref idref="DRAWINGS">FIG. 4</figref>, which illustrates exemplary process steps where a user <b>100</b> (e.g., mobile station) is the initiator of data exchange. Of course, <figref idref="DRAWINGS">FIG. 5</figref> is likewise applicable, which illustrates exemplary process steps where a service provider (e.g., host system) is the initiator of data exchange. Also, the exemplary steps need not occur in the order described.

0140To generate and exchange a shared secret (step <b>104</b>), the user <b>100</b> and the host system <b>102</b> can use any suitable approach such as previously described above. Either party can generate the secret and share it with the other via an out-of-band communication.

0141The user <b>100</b> (e.g., mobile station <b>30</b> or <b>32</b>) can generate a long term key pair (step <b>108</b>) by picking a random number a where 1&lt;a&lt;r−1, and by calculating A=aP (elliptic curve scalar multiplication), where a is the user's long term private key (or first private key), A is the user's long term public key (or first public key), and P and r are as described above. Thus, in this example, the user's long term public key is generated using an elliptic curve calculation. The user <b>100</b> (e.g., mobile station) can then securely store a and A.

0142To generate its short term key pair (step <b>108</b>), the user <b>100</b> can first convert the shared secret s to a point on E(F_q), and this point can be denoted as S. The user can then choose a random number x, where 1&lt;x&lt;r−1, and can calculate X=xS (elliptic curve scalar multiplication). The quantity x is the user's short term private key (or second private key) of the short term key pair (also called its short term authentication key pair), and X is the user's short term public key (or second public key).

0143The user <b>100</b> can then send (step <b>110</b>) the second public key X and the first public key A to the host system <b>102</b> along with first auxiliary data (‘Aux Data_D’). As discussed further below, the first auxiliary data will be used in generating a key confirmation value. The first auxiliary data can be essentially any data string such as a random data string, or it can be other useful data, such as device configuration information that can be used by the host system in configuring one or more devices at the host system's end. Its size is agreed upon in advance by the communicating parties.

0144The host system <b>102</b> receives the long term and short term public keys of the user <b>100</b> and recalls the shared secret s (step <b>112</b>). The host system can generate its own long term key pair (step <b>114</b>) by picking a random number b, where 1&lt;b&lt;r−1 and by calculating B=bP (elliptic curve scalar multiplication), where b is the host system's long term private key (or third private key), B is the host system's long term public key (or third public key), and P and r are as described above. Thus, in this example, the host system's long term public key is also generated using an elliptic curve calculation. The host system <b>102</b> can then securely store b and B. Of course, various steps can occur in any suitable order, e.g., the host system can calculate its long term public and private keys prior to receiving any public key information from the user <b>100</b>.

0145To generate its short term key pair (step <b>114</b>), the host system can first convert the shared secret s to a point on E(F_q), and this point can be denoted as S. Since the user <b>100</b> and the host system <b>102</b> are operating in framework of shared parameters, the user <b>100</b> and host system <b>102</b> both generate the same value for S. The host system can then choose a random number y, where 1&lt;y&lt;r−1, and can calculate Y=yS (elliptic curve scalar multiplication). The quantity y is the host system's short term private key (or fourth private key) of the short term key pair (also called its short term authentication key pair), and Y is the host system's short term public key (or fourth public key).

0146The host system <b>102</b> can then calculate (step <b>116</b>) a master key (or master secret) using the user's long term and short term public keys (first and second public keys) using the following exemplary calculation:

0147Calculate master key (e.g., by host system): <ul id="ul0029" list-style="none"><li id="ul0029-0001" num="0000"><ul id="ul0030" list-style="none"><li id="ul0030-0001" num="0148"> k1=yX; </li><li id="ul0030-0002" num="0149"> k₂=bA; </li><li id="ul0030-0003" num="0150"> if k₁=0, 1, −1, then set k₁=random number; </li><li id="ul0030-0004" num="0151"> if k₂=0, 1, −1, then set k₂=random number; and </li><li id="ul0030-0005" num="0152"> (k, k_conf)=SHA-512(k₁∥ k₂). <br /> In the last of the above-noted steps, SHA-512 refers to the 512 bit “secure hash algorithm” that is known to those of ordinary skill in the art, and ‘∥’ is an agreed upon concatentation function (see Federal Information Processing Standards Publication 180-2, “Secure Hash Standard”, Aug. 1, 2002, located at http://csrc.nist.gov/publications/fips/fips180-2/fips180-2withchangenotice.pdf). Of course, other secure hash algorithms known in the art, such as those that utilize a different number of bits could also be used. The reference to (k, k_conf) means the first half of the result (first 256 bits) from the SHA-512 calculation is the calculated master secret key k, and the second half of the result (second 256 bits) is a quantity k_confthat will be used in calculating the host system's key confirmation value h_B(or test string) as discussed below. The steps at which k1 and k2 are tested to check whether either of them is equal to 0, 1 or −1 is similar to the check made at step <b>314</b> of <figref idref="DRAWINGS">FIG. 6</figref> discussed previously to test whether there is a possible attack being staged. If the condition is met, such that either or both of k1 and k2 are set to random numbers, the calculated master key k will not match the user's master key, and two-way secure communication based upon a shared master (secret) key will not occur. Of course, the process could simply be aborted at this stage if the condition is met. </li></ul></li></ul>

0153The host system <b>102</b> can calculate its key confirmation value h_B(step <b>116</b>) using the following calculation: <br /><i>h</i>_B<i>=HMAC</i>-256_k_{<sub2>conf</sub2>}(Aux Data_D∥ Aux Data_B<i>∥ A ∥ B ∥ X ∥ Y </i>∥ ASCII₁). <br /> In this calculation, HMAC-256 refers to the 256 bit “hashed message authentication code” algorithm known to those of ordinary skill in the art (see “HMAC: Keyed-Hashing for Message Authentication”, H. Krawczyk et al., Network Working Group Request for Comments: 2104, February 1997, 11 pages, accessible from the Internet at http://www.faqs.org/rfcs/rfc2104.html). Of course, other versions of a hashed message authentication code, such as those that use a different number of bits, could also be used. The quantity k_conf(referred to previously) is a bit string provided as input to the HMAC-256 algorithm. Aux Data_Dis the first auxiliary data provided by the user <b>100</b> as described above. Aux Data_Brefers to second auxiliary data generated by the host system. The second auxiliary data can be essentially any data string such as a random data string, or it can be other useful data, such as device configuration information that can be used by the user in configuring one or more devices at the user's end. Its size is agreed upon in advance by the communicating parties and is the same size as the first auxiliary data. The symbol ‘∥’ is an agreed upon concatentation function. A, X, B, and Y are the first through fourth public keys, and ASCII₁is any predetermined ASCII code (e.g., for the letter “B”) programmed into the argument of the HMAC algorithm.

0154The host system <b>102</b> can then send its short term public key (fourth public key) Y, long term public key (third public key) B, and key confirmation value (test string) h_Bto the user <b>10</b> (step <b>116</b>). The host system <b>102</b> can also send its second auxiliary data (Aux Data_B) to the user <b>102</b> at this stage. After receiving this information (step <b>118</b>), the user can perform its calculation of the master key.

0155The user <b>100</b> can perform its calculation (step <b>120</b>) of the master key (or master secret) based upon the host system's long term and short term public keys (third and fourth public keys) using the following exemplary calculation:

0156Calculate master key (e.g., by user): <ul id="ul0031" list-style="none"><li id="ul0031-0001" num="0000"><ul id="ul0032" list-style="none"><li id="ul0032-0001" num="0157"> k₁=xY; </li><li id="ul0032-0002" num="0158"> k_{2 =aB;}</li><li id="ul0032-0003" num="0159"> if k₁=0, 1, −1, then set k1=random number; </li><li id="ul0032-0004" num="0160"> if k₂=0, 1, −1, then set k₂=random number; and </li><li id="ul0032-0005" num="0161"> (k, k_conf)=SHA-512(k₁∥ k₂). <br /> In the last of the above-noted steps, SHA-512 refers to the 512 bit “secure hash algorithm” that is known to those of ordinary skill in the art, and ‘∥’ is an agreed upon concatentation function. The reference to (k, k_conf) means the first half of the result (first 256 bits) from the SHA-512 calculation is the calculated master secret key k, and the second half of the result (second 256 bits) is the quantity k_confthat will be used in calculating the users key confirmation value h_A(or test string) as discussed below. Based on the above, it is apparent that the user's calculations are intended to produce the same values for k and k_confas produced by the host system's calculations. </li></ul></li></ul>

0162The user <b>100</b> can then check (step <b>120</b>) the host system's key confirmation value (test string) according to the following test: <br />if <i>h</i>_B<i>!=HMAC</i>-256_k_{<sub2>conf</sub2>}(Aux Data_D∥ Aux Data_B<i>∥ A ∥ B ∥ X ∥ Y </i>∥ ASCII₁), <br /> then abort. <br /> As discussed above, the quantity ASCII₁is a predetermined ASCII code. It is not only programmed into the argument of the HMAC algorithm at the host system's end for generating h_B, but also is programmed into argument of the HMAC algorithm at the user's end for checking h_B. The symbol “!=” means “not equal to” as known to those of ordinary skill in the art.

0163The user <b>100</b> can then generate its own key confirmation value (test string) h_Athat is different from the host system's key confirmation value h_B(step <b>122</b>) according to the following exemplary calculation: <br /><i>h</i>_A<i>=HMAC</i>-256_k_{<sub2>conf</sub2>}(Aux Data_D∥ Aux Data_B<i>∥ A ∥ B ∥ X ∥ Y </i>∥ ASCII₂) <br /> In this calculation, the various quantities are as described above, and ASCII₂is a another predetermined ASCII code (e.g., for the letter “A”) programmed into the argument of the HMAC algorithm for generating h_A. ASCII₂can be set to a different value than ASCII₁to ensure that h_Ais not the same as h_B. The user <b>100</b> can then send h_Ato the host system <b>102</b> (step <b>122</b>).

0164The host system <b>102</b> can then check (step <b>124</b>) the user's key confirmation value (test string) according to the following test: <br />if <i>h</i>_A<i>!=HMAC</i>-256_k_{<sub2>conf</sub2>}(Aux Data_D∥ Aux Data_B<i>∥ A ∥ B ∥ X ∥ Y </i>∥ ASCII₂), <br /> then abort. <br /> To carry out this calculation, it is evident that the predetermined ASCII code ASCII₂is programmed into the argument of the HMAC algorithm at the host system's end for checking h_A. If the user <b>100</b> (e.g., mobile station) makes more than a predetermined number (e.g., 10) of failed attempts, then the host system <b>102</b> can deny service to that user.

0165Finally, the user <b>100</b> can zero out the quantities s, x, k₁, k₂, and k_conf, and can securely store B (host system's long-term public key) and k (shared master key). Likewise, the host system <b>102</b> can zero out the quantities s, y, k₁, k₂, and k_conf, and can securely store A (user's long-term public key) and k (shared master key).

0166A discussion of several security considerations with respect to the above approach is now provided.

0000Eavesdropping Attack

0167The shared secret is a component of k₁=yX=xY=xyS. The attacker is faced with calculating xyS knowing only X and Y. This is a Diffie-Hellman problem and is assumed to be hard.

0000Man-in-the-Middle Attack

0168Diffie-Hellman key agreements are traditionally prone to man-in-the-middle attacks. However, the base points used in the protocol are the shared secret key. To mount a man-in-the middle attack, all parties must be using the same base point. Since the attacker does not know the shared secret key he will not be able to successfully calculate k₁, and hence cannot calculate k.

0000Impersonating a Mobile Station

0169Since an attacker does not know the secret s, he or she must send X=xP, instead of xS, to the host system. The host system will calculate k₁=yX=yxP. To calculate the same key, the attacker needs to determine y from Y. This is the discrete logarithm problem and is assumed to be hard.

0000Impersonating a Host System

0170Since an attacker does not know the secret s, he or she must send Y=yP, instead of yS, to the user (e.g., mobile station). The user's device will calculate k₁=xY=xyP. To calculate the same key the attacker needs to determine x from X. This is the discrete logarithm problem and is assumed to be hard.

0000Preventing an Offline Dictionary Attack

0171The best position to attack the password is to impersonate the mobile station, as the host system sends the first key confirmation value. This allows the attacker to try different passwords, and check if he or she has made the correct choice.

0172Assume the attacker sends X=xP, instead of xS, to the host system. The host system will reply with Y=xS and calculate k_{1 HOST}=yX=yxP. Meanwhile, the attacker will calculate k_{1 attacker}=xY=yxS=yxzP, for some z such that S=zP. Using the key confirmation value h_B, the attacker needs to find a value w such that wk_{1 attacker}=wxY=wyxzP=xyP=k_{1 HOST}. Notice however, that w=z⁻¹mod r. Therefore finding w amounts to finding z, which corresponds to solving the discrete logarithm problem for S. This is assumed to be hard.

0000Preventing an Online Dictionary Attack

0173The best position to attack the password is to impersonate the mobile station, as the Host System sends the first key confirmation value. During each run the mobile station tries a different password, and checks if he or she has made the correct choice against h_B. This is averted by limiting the mobile station to a predetermined number (e.g., 10) of attempts to determine the correct secret s.

0000Small Subgroup Attack

0174A small subgroup attack is one in which an attacker tries to force the key agreement result to come from a small set of values. For example if the attacker chooses X to be the point at infinity then k₁would be the point at infinity no matter what the host system chose for Y. Therefore by checking that X is not the point at infinity, one or minus one this threat is averted.

0000Regeneration of New Master Key (Re-Keying using MQV)

0175As noted above, regeneration of a new master key (re-keying) in the present exemplary framework can be done using the MQV algorithm known to those of ordinary skill in the art. In the example below, reference will be made to certain general process steps of previously described <figref idref="DRAWINGS">FIGS. 8 and 9</figref>, a modifications thereto.

0176As shown in <figref idref="DRAWINGS">FIG. 8</figref>, a re-key request is received by the user <b>100</b>, or the user <b>100</b> decides to cut a new key (step <b>430</b>). Of course, step <b>430</b> could be executed by the service provider instead of the user.

0177The user generates a new short-term encryption key pair (step <b>432</b>). According to an exemplary mathematical calculation, the user <b>100</b> chooses a new random number x to be its new private short term authentication key (new second private key), where 1&lt;x&lt;r-1, and calculates X=xP, where P is as previously described, and where X is the user's new public short term authentication key, which can also be referred to as a new public short term encryption key (new second public key). The user <b>100</b> sends X to the host system <b>102</b> (step <b>434</b>) and also sends first auxiliary data (Aux Data_D) along with X, where the first auxiliary data is as previously described.

0178The host system <b>102</b> receives the user's new public short term authentication key (step <b>460</b>) and generates its own new short term authentication (or encryption) key pair (step <b>462</b>). According to an exemplary calculation, the host system chooses a new random number y to be its new private short term authentication key (new fourth private key), where 1&lt;y&lt;r−1, and calculates Y=yP, where Y is the host system's new public short term authentication key (new fourth public key).

0179The host system <b>102</b> can then calculate (step <b>464</b>) a new master key using the following exemplary calculation:

0180Calculate new master key (e.g., by host system): <ul id="ul0033" list-style="none"><li id="ul0033-0001" num="0000"><ul id="ul0034" list-style="none"><li id="ul0034-0001" num="0181"> S_B=(y+{overscore (Y)}b) mod r; </li><li id="ul0034-0002" num="0182"> Z=S_B(X+{overscore (X)}A); </li><li id="ul0034-0003" num="0183"> if Z is the point of infinity, then choose a new y and re-calculate s_Band Z; </li><li id="ul0034-0004" num="0184"> (k, k_conf)=SHA-512(x_Z), where x_Zis the x-coordinate of Z. <br /> In the above, {overscore (X)} and {overscore (Y)} are calculated from X and Y, respectively, using a known approach in the MQV protocol. Namely, to calculate {overscore (X)}, the abscissa (x-coordinate) value of X is obtained (noting that X is an elliptic curve point having an x-coordinate value and a y-coordinate value), the leading half (left half) of the abscissa's bit values are dropped, and a leading “1” is inserted in place of the dropped bits. The calculation of {overscore (Y)} from Y is done in the same manner. The first half of the bits of the result of the SHA calculation is the master key k, and the second half of the bits is the value of k_conf, such as described previously. In this example, separate k1 and k2 values are not generated during re-keying, and a step analogous to step <b>466</b> is not carried out. </li></ul></li></ul>

0185The host system <b>102</b> can calculate (step <b>470</b>) its key confirmation value h_B(test string) using the following calculation in the manner discussed previously: <br /><i>h</i>_B<i>=HMAC</i>-256_k_{<sub2>conf</sub2>}(Aux Data_D∥ Aux Data_B<i>∥ A ∥ B ∥ X ∥ Y </i>∥ ASCII₁).

0186The host system <b>102</b> then sends Y and h_B(step <b>472</b>) to the user <b>100</b> along with the second auxiliary data (Aux Data_B). After receiving this information, the user can perform its calculation of the new master key.

0187The user <b>100</b> can perform its calculation (step <b>440</b>) of the new master key (or master secret) using the following exemplary calculation:

0188Calculate new master key (e.g., by user): <ul id="ul0035" list-style="none"><li id="ul0035-0001" num="0000"><ul id="ul0036" list-style="none"><li id="ul0036-0001" num="0189"> s_A=(x+{overscore (X)}a)mod r; </li><li id="ul0036-0002" num="0190"> Z=s_A(Y+{overscore (Y)}B); </li><li id="ul0036-0003" num="0191"> if Z is the point of infinity, then set Z to a random point of E(F_q); </li><li id="ul0036-0004" num="0192"> (k, k_conf)=SHA-512(x_Z), where x_Zis the x-coordinate of Z. <br /> In the above, {overscore (X)} and {overscore (Y)} are calculated from X and Y as described previously. Other aspects of the calculation are like those previously described. Also, in this example, separate k1 and k2 values are not generated during re-keying, and a step analogous to step <b>442</b> is not carried out. During this time, if a received message cannot be decrypted with the current master key or with the previous master key (if still available), then pending master key (i.e., the newly calculated quantity k intended to become the new master key) can be tried. </li></ul></li></ul>

0193The user <b>100</b> can then check (step <b>466</b>) the host system's key confirmation value (test string) according to the following test, such as described previously: <br />if <i>h</i>_B<i>!=HMAC</i>-256_k_{<sub2>conf</sub2>}(Aux Data_D∥ Aux Data_B<i>∥ A ∥ B ∥ X ∥ Y </i>∥ ASCII₁). <br /> then abort. <br /> If the quantity h_Bmatches the expected value, the user <b>100</b> sets the newly calculated value of k to be the new master key.

0194The user I <b>00</b> can then generate its own key confirmation value (test string) h_Athat is different from the host system's key confirmation value h_B(step <b>452</b>) according to the following exemplary calculation, as previously described: <br /><i>h</i>_A<i>=HMAC</i>-256_k_{<sub2>conf</sub2>}(Aux Data_D∥ Aux Data_B<i>∥ A ∥ B ∥ X ∥ Y </i>∥ ASCII₂). <br /> The user <b>100</b> can then send h_Ato the host system <b>102</b> (step <b>454</b>).

0195The host system <b>102</b> can then check (step <b>474</b>) the user's key confirmation value (test string) according to the following test: <br />if <i>h</i>_A<i>!=HMAC</i>-256_k_{<sub2>conf</sub2>}(Aux Data_D∥ Aux Data_B<i>∥ A ∥ B ∥ X ∥ Y </i>∥ ASCII₂), <br /> then abort. <br /> If the quantity h_Amatches the expected value, the host system <b>102</b> sets the newly calculated value of k to be the new master key.

0196Finally, the user <b>100</b> can zero out the quantities s_A, x, and k_conf, and can securely store k (shared master key). Likewise, the host system <b>102</b> can zero out the quantities s_B, y, and k_conf, and can securely store k (shared master key).

0197A discussion of several security considerations with respect to the above approach is now provided.

0000Man-in-the-Middle Attack.

0198According to the National Institute of Standards and Technology (NIST), this attack will not be successful against an MQV key generation scheme. As long as the long-term private keys a and b remain secret there can be no man-in-the-middle attack. (See, <i>NIST Special Publication </i>800-56<i>: Recommendation on Key Establishment schemes, Draft </i>2.0, January 2003, National Institute of Standards and Technology, available at http://csrc.nist.gov/CryptoToolkit/tkkevmgmt.html).

0000Perfect Forward Secrecy

0199Notice that the shared secret varies from one key establishment transaction to the next. If both the static (long term) and ephemeral (short term) private keys from a particular transaction are lost, the shared secrets from other legitimate transactions are not compromised because different ephemeral private keys are used in each transaction.

0000Masquerade Attack

0200NIST indicates that if an attacker compromises a victim's long-term private key, in MQV, the attacker cannot masquerade as a third party to the victim. (See, <i>NIST Special Publication </i>800-56<i>: Recommendation on Key Establishment schemes, Draft </i>2.0, January 2003, National Institute of Standards and Technology, available at http://csrc.nist.gov/CryptoToolkit/tkkeymgmt.html). The ability to prevent the masquerade attack is an advantage of the above-described approach utilizing MQV in re-keying or over-the-air key generation, since if this feature is missing, and if the host system's long-term private key is compromised, the attacker holding the key can potentially masquerade as any mobile station to the host system.

0201Having described the above-noted additional exemplary first and second protocols for key generation and re-keying, respectively, further exemplary aspects will now be described. For example, a method for secure bidirectional communication between two systems can be used. A first key pair and a second key pair are generated, the latter including a second public key that is generated based upon a shared secret. First and second public keys are sent to a second system, and third and fourth public keys are received from the second system. The fourth public key is generated based upon the shared secret. A master key for encrypting messages is calculated based upon a first private key, a second private key, the third public key and the fourth public key. For re-keying, a new second key pair having a new second public key and a new second private key is generated, and a new fourth public key is received. A new master key is calculated using elliptic curve calculations using the new second private key and the new fourth public key.

0202As another example, a first system is used for establishing a secure bidirectional communication path to a second system for exchanging one or more messages. The first system includes a memory and a processing unit coupled to the memory. The processing unit is configured to execute steps of: generating a first key pair having a first public key and a first private key; generating a second key pair having a second public key and a second private key, the second public key being generated based upon a shared secret known to the first system and the second system; sending the second public key and the first public key to the second system; receiving a third public key and a fourth public key generated by the second system, the fourth public key being generated based upon the shared secret; calculating a first master key based upon the first private key, the second private key, the third public key and the fourth public key, wherein the first master key is configured to be used in encryption of one or more messages; generating a new second key pair having a new second public key and a new second private key; receiving a new fourth public key from the second system, and calculating a new master key based upon the first private key, the new second private key, the third public key, and the new fourth public key using an elliptic curve calculation.

0203According to another aspect, any form of computer readable carrier can contain processing instructions adapted to a cause a processing unit to execute the methods described herein. The computer readable carrier can be any suitable type of carrier, such as solid-state memory (e.g., read only memory (ROM), random access memory (RAM), etc.), magnetic memory, optical memory, other type of memory, or modulated waves/signals (such as radio frequency, audio frequency, or optical frequency modulated waves/signals) containing an appropriate set of computer instructions that would cause a processing unit to carry out the techniques described herein.

0204According to another aspect, a computer readable carrier contains processing instructions adapted to cause a processing unit to execute the above-noted steps.

0205An exemplary password-based encryption communication system, in wireless or wired environments, having, if desired, perfect forward secrecy has been described. It includes using a long-term key-pair in combination with a short-term authentication key-pair, generated using a shared secret, to allow for the implementation of perfect forward secrecy. The long-term public key is piggy-backed with the authentication public key to enable an authentic exchange of long-term keys. This enables the corresponding party that is in possession of the shared secret, to receive and be able to use the long-term public key.

0206Having described in detail the exemplary embodiments of the present invention, including exemplary methods of operation, it is to be understood that the operations described herein could be carried out with different elements and steps. The exemplary embodiments are presented only by way of example and are not meant to limit the scope of the present invention, which is defined by the following claims.