Secure session set up based on the Wireless Application Protocol.
Technical Field of the InventionThe Wireless Application Protocol (WAP) defines an industry-wide specification for developing applications that operate over wireless communication networks. The wireless market is growing very quickly, and reaching new customers and services. To enable operators and manufacturers to meet the challenges in advanced services, differentiation and fast/flexible service creation a set of protocols has been designed in transport, security, transaction, session and application layers.
Background of the InventionWAP security functionality includes the Wireless Transport Layer Security (WAPWTLS) and application level security, accessible using Wireless MarkupLanguage Script (WMLScript). For optimum security, some parts of the security functionality need to be performed by a tamper-resistant device, so that an attacker cannot retrieve sensitive data. Such data is especially the permanent private keys used in WTLS handshake with client authentication, and for making application level electronic signatures (such as confirming an application level transaction). In WTLS, also master keys (master secrets) are relatively long living-which could be several days-this is in order to avoid frequent full handshakes which are quite heavy both computationally and due to relatively large data transfer. Master secrets are used as a source of entropy, to calculate MAC keys and message encryption keys which are used to secure a limited number of messages, depending on usage of WTLS.
US 5.307.411 describe the set up of a secure communication session between two communication units, such as phones or facsimile machines.
The secure session is controlled by separate smart cards based verification units associated with a respective one of the communication units. These two verification units exchanges random number, encrypts these numbers by using private keys, returns the encrypted random numbers to their origin.
Then the encrypted random number is decrypted based on public keys. If the received numbers corresponds to the transmitted numbers, the parties verifies each other and the secure session may take place.
Summary of the InventionThe main object of the present invention is to establish a secure connection between a wireless communication apparatus and a data communication apparatus based on a wireless application protocol.
Another object is that establishing a secure connection is a heavy procedure both computationally and due to intensive data transfer. That is why, there is a need to use the mutually agreed master secret for a relatively long time.
The problem is to store the master key in a secure way. Partly due to that problem, it is common practice to restrict the lifecycle of the master secret and the associated secure session to e. g., 24 hours, after which it is required to perform the heavy key establishment procedure anew.
The main object is achieved in accordance with the present invention by connecting a wireless communication apparatus, e. g. a cellular phone, to a separate unit, e. g. a smart card, a SIM (Subscriber identity Module) card, etc., which may store sensitive data of a secure connection. This means that the wireless communication apparatus having some kind of contact means, for example wireless (e. g. infra-red, radio frequency, etc.) or physical (i. e. an electrical contact), for receiving information from the separate unit, i. e. the unit is provided with memory means. The memory means comprises information to control an access of the wireless communication apparatus through a wireless communication network, e. g. a cellular phone network, connected to a data communication apparatus, e. g. a server, which supports a WirelessApplication Protocol (WAP).
One advantage of using a separate unit, when establishing a secure connection, is that it will be much easier to re-establish a connection to the data communication apparatus. Thus, it is possible to save information, e. g. signatures, secret keys, etc., in the memory means, and may be re-used in another secure connection. In order to avoid fraud, the re-use of a secure connection can be restricted for limited period of time. By saving this information in the memory means the second object will be achieved.
Another advantage is that the user pays less when re-establishing a secure session, in case of the necessary information to re-establishing is saved.
To establish a connection, the wireless communication apparatus connects to the separate unit, accessing the wireless communication network connected to said data communication apparatus. Then the wireless communication apparatus transmits a request to the data communication apparatus. This request comprises information of which pre-defined algorithm (s) the wireless communication apparatus supports. When the data communication apparatus receives this request, it chooses at least one algorithm, associated with a public key and a private key, and transmits a message back to the wireless communication apparatus. This message comprises the public key and information about which algorithm the data communication apparatus has chosen. When the wireless communication apparatus receives the message, comprising the public key, it will generate a master secret code, and calculates a signature based on the chosen algorithm, the public key and the master secret code. Thereafter, the wireless communication apparatus will transmit a respond to the data communication apparatus. This respond comprises the calculated signature. When the data communication apparatus receives the respond, comprising the signature, it will calculate the master secret code based on the chosen algorithm, the signature received, and the private key. Finally, the data communication apparatus will be able to establish a secure connection to the wireless communication apparatus.
Further advantages of the vane arrangement according to the present invention will be apparent from the dependent claims.
Brief Description of the DrawingFig. 1 schematically illustrates a preferred embodiment of a hand portable phone according to the invention.
Fig. 2 schematically shows the essential parts of a telephone for communication with a cellular or cordless network.
Fig. 3 schematically shows how the secure session is set up between a client /phone and a server according to the invention.
Fig. 4 illustrates the message structure for setting up a secure connection according to the invention.
Detailed Description of EmbodimentsFig. 1 shows a preferred embodiment of a phone according to the invention, and it will be seen that the phone, which is generally designated by 1, comprises a user interface having a keypad 2, a display 3, an on/off button 4, a speaker 5, and a microphone 6. The phone 1 according to the preferred embodiment is adapted for communication via a cellular network, but could have been designed for a cordless network as well. The keypad 2 has a first group 7 of keys as alphanumeric keys, by means of which the user can enter a telephone number, write a text message (SMS), write a name (associated with the phone number), etc. Each of the twelve alphanumeric keys 7 is provided with a figure"0-9"or a sign"#"or"*", respectively. In alpha mode each key is associated with a number of letters and special signs used in text editing.
The keypad 2 additionally comprises two soft keys 8, two call handling keys 9, and a navigation key 10.
The two soft keys 8 have a functionality corresponding to what is known from the phones Nokia 21 1 OTM, Nokia 8110w and Nokia 3810w. The functionality of the soft key depends on the state of the phone and the navigation in the menu by using a navigation key. The present functionality of the soft keys 8 is shown in separate fields in the display 3 just above the keys 8.
The two call handling keys 9 according to the preferred embodiment are used for establishing a call or a conference call, terminating a call or rejecting an incoming call.
The navigation key 10 is an up/down key and is placed centrally on the front surface of the phone between the display 3 and the group of alphanumeric keys 7. Hereby the user will be able to control this key with his thumb. This is the best site to place an input key requiring precise motor movements. Many experienced phone users are used to one-hand handling. They place the phone in the hand between the finger tips and the palm of the hand. Hereby the thumb is free for inputting information.
Fig. 2 schematically shows the most important parts of a preferred embodiment of the phone, said parts being essential to the understanding of the invention. The preferred embodiment of the phone of the invention is adapted for use in connection with the GSM network, but, of course, the invention may also be applied in connection with other phone networks, such as cellular networks and various forms of cordless phone systems or in dual band phones accessing sets of these systems/networks. The microphone 6 records the user's speech, and the analog signals formed thereby are A/D converted in an A/D converter (not shown) before the speech is encoded in an audio part 14. The encoded speech signal is transferred to the controller 18, which supports the GSM terminal software. The controller 18 also forms the interface to the peripheral units of the apparatus, including a RAM memory 17a and a Flash ROM memory 17b, a SIM card 16, the display 3 and the keypad 2 (as well as data, power supply, etc.). The controller 18 communicates with the transmitter/receiver circuit 19. The audio part 14 speech-decodes the signal, which is transferred from the controller 18 to the earpiece 5 via an D/A converter (not shown).
The controller 18 is connected to the user interface. Thus, it is the controller 18 which monitors the activity in the phone and controls the display 3 in response thereto.
Therefore, it is the controller 18 which detects the occurrence of a state change event and changes the state of the phone and thus the display text. A state change event may be caused by the user when he activates the keypad including the navigation key 10, and this type of events is called entry events or user events. However, the network communicating with the phone may also cause a state change event. This type of event and other events beyond the user's control are called non user events. Non user events comprise status change during call set-up, change in battery voltage, change in antenna conditions, message on reception of SMS, etc.
An example of a tamper-resistant device is a smart card (SC). In the phone, it can be the Subscriber Identity Module (SIM) or an extemal smart card.
The way which a phone and a smart card interact is specified as a commandresponse protocol. The goal of this protocol is to provide means for a WAP handset to utilize smart cards in performing WTLS and application level security functions. The functionality presented here is based on the requirement that sensitive data, especially keys, can be stored in the card, and all operations where these key are involved can be performed in the card.
Different classes of the cards are introduced in order to define how widely the functionality is implemented.
This specification is based on IS07816 series of standards on smart cards. In particular, it uses the IS07816-8 standard (draft) [IS07816-8]. When this functionality is applied to GSM SIM there may be a need to extend also the related GSM specifications [GSM11.11], where applicable.
According to the invention the smart card 16 is used to enhance security of the implementation of the Security Layer and certain functions of theApplication Layer. The smart card 16 can be used for several purposes forWTLS. The major purpose of the smart card 16 is to perform cryptographic operations during the handshake, especially when the handshake is used for client authentication. Furthermore the memory of the smart card 16 is used for securing a master secret, a public key and other type of confidential material during long-living WTLS sessions. Finally the memory of the smart card 16 is used for recording the level security of the sessions. According to the invention the WTLS support in a smart card 16 can be described with reference to the following three embodiments.
First embodiment.
According to this embodiment, the smart card 16 is used for storage of permanent, typically certified, private keys and for performing operations using these keys. The operations includes signing operation (e. g., ECDSA orRSA) for client authentication when needed for the selected handshake scheme; key exchange operation using a fixed client key (e. g., ECDH key, in ECDSA-ECDSA handshake).
The smart card 16 is not required to perform the calculation of the master secret or operations using the master key. These calculations may advantageously be performed by the controller 18 of the phone. However, the smart card 16 may act as a persistent storage for WTLS secure session (and connection) data, including master secrets. In this case, master secrets would be calculated and used for key derivation in the volatile phone memory (theRAM 17a) but erased from there when not needed at that moment, e. g., when the user exits from secure WAP applications. Not storing session data persistently in phone 1 may improve security, e. g., in the case of a stolen phone 1. It also brings better usability in the case of changing the smart card 16 from one phone 1 to another.
Additionally, for portability, the smart card 16 may store needed certificates.
Storage of trusted root certificates (or public keys) has significance also from security point of view: they must not be altered-but they can be exposed without danger.
Note that when public key encryption based key exchange (e. g., RSA) is used according to the first embodiment of the invention, there is no advantage in doing public key encryption on the smart card 16 when the pre-master secret would anyway be retumed to the phone1, for master secret calculation in the controller 18.
When client authentication is not supported in WTLS, at the minimum, the smart card 16 only acts as a storage for session data. If client authentication is supported, the card would be able to perform a signing operation based on a private key (e. g., ECDSA or RSA) stored in the card, or key agreement calculation (e. g., ECDH) based on a fixed key stored in the card.
Second embodiment.
According to the second embodiment, the smart card 16 is used as a tamper resistant device for all crypto-critical functionality : storage of all persistent keys and operations using these keys. Besides the operations performed according the first embodiment, the smart card 16 now also supports the calculation (ECDH key exchange) or generation (RSA key exchange) of the pre-master secret; calculation and storage of the master secret for each secure session; and derivation and output of key material (for MAC, encryption keys, IV, finished check), based on the master secretThe phone 1 stores MAC and message encryption keys as long as they are currently needed. These keys have a limited lifetime which may be negotiated during the WTLS handshake-in the extreme case they are used for a single message only. The phone 1 has to delete the from its RAM memory 17a when the user exits from the secure WAP applications. These keys can always be derived anew from the master secret if needed.
An attacker who obtains a message encryption key can read as many messages as is agreed in the key refresh configuration (in the extreme case, a single message). An attacker who obtains a MAC key can impersonate the compromised party during as many messages as is agreed in the configuration (in the extreme case, a single message).
Third embodiment.
Certain specialized smart cards 16 may act as full-blown security engines forWTLS. This requires that the smart card 16 is equipped with its own processing unit and only uses the phone 1 as an interface to the cellular network during the secure session set up or the handshake procedure.
Besides the operations according to the second embodiment, the smart card 16 may store MAC and encryption keys for each secure connection ; and perform MAC calculation/verification and encryption/decryption of messages.
Furthermore the smart card 16 may be responsible for the verification of certificates and the verification of digital signatures.
Note that having message encryption in the smart card 16 does not necessarily bring any additional security because in any case the data is as plain text in the phone 1. The same is true for MAC calculation : the phone 1 must be trusted to input and output data in a correct way. The only advantage here would be not having to take encryption keys out of the card 16. However, the keys have a limited lifetime which is negotiated during the WTLS handshake-in the extreme case they are used for single message only.
According to the third embodiment, the smart card 16 will contain all algorithms so that they could be controlled by smart card issuers.
Smartcard.
The term"smartcard"covers a card-like unit having some memory means in which some secret information identifying the card holder is stored. The memory means may be a magnet strip that may be read by a magnet reader, or it may be provided as discrete memory components as a ROM, EEPROM etc. When the user inserts the smart card in a more or less public apparatus he may become authorized to perform some operations such as banking operations. Presently the user of a GSM phone is identified by a so-calledSubscriber Identity Module or a SIM card 16, and the structure of this type of smart card is defined in the GSM specification"Specification of the SubscriberIdentity Module-Mobile Equipment (SIM-ME) interface", GSM 11.11 version 5.5.0, published by European Telecommunications Standards Institute ; ETSI.
The present type of smartcards will be able to support the first embodiment explained above.
Gemplus has recently launched a smartcard, GemXpresso RAD, based on a 32-bit chip from Texas Instruments using ARM7 RISC core technology. This 32 bit RISC processor has a 32 kbyte of non volatile flash memory and 8 kbyte of ROM. When the mechanical interface of the Gemplus card is adapted to fulfill the GSM specification this type of smartcard will be able to support the second and the third embodiment.
Network.
Fig. 3 schematically shows how the secure session, i. e. a secure connection, between a data communication apparatus and a wireless communication apparatus, e. g. a cellular phone 1. Basically the WAP content and applications are specified in a set of well-known content formats based on the familiar WWW content formats. Content is transported using a set of standard communication protocols based on the WWW communication protocols. A browser in the phone 1 co-ordinates the user interface and is analogous to a standard web browser.
The wireless communication apparatus 1 is a client 1 who wants to establish a secure connection to a server 20,30,40, which is the data communication apparatus 20,20,30. The client is provided in an environment, which make it possible to reach a wide variety of different wireless platforms, e. g. world wide web (WWW). The environment provided may be referred to as WirelessApplication Environment (WAE). This means that the client 1 may be supported by some kind of browser, e. g. a micro-browser, to access the different services connected to the server. In order to access these services the browser may comprise following functionalities :Wireless Markup Language (WML)-a lightweight markup language, similar to HTML, but optimised for use in hand-held mobile terminals ; WMLScript-a lightweight scripting language, similar to JavaScript ; -Wireless Telephony Application (WTA, WTAI)-telephony services and programming interfaces; andContent Formats-a set of well-defined data formats, including images, phone book records and calendar information.
The server 20 is using a wireless application protocol, and may comprise a gateway 30 and an origin server 40. The gateway 30 is also a server, which may identify and encrypt/decrypt information between the client 1 and the origin server 40. This means that the gateway is provided with encoders and decoders (not shown). Also, the server 20 comprises different algorithms to make the encryption/decryption. The encryption/decryption itself may be performed by well-known methods, e. g. RSA, Diffie-Hellman, etc. The origin server 40 comprises different scripts to support WAP and data to be accessed by the client. This data may be all kind of information, e. g. weather reports, news, information from stock markets, etc.
In order to access the server 20, from the client 1, the server has to be connected to a wireless communication network 50, e. g. a cellular phone network. Therefore, in accordance with the present invention, the client is provided with contact means (not shown) for receiving information from a separate unit (not shown) provided with memory means. This separate unit may be a smart card, subscriber identity module (SIM), or the like. The memory means may be a random access memory (RAM), read only memory (ROM), or the like. Further, the memory means comprises information to control the access of the server 20 through the wireless communication network 50.
To establish a secure connection, the client 1 connects to the separate unit, accessing the wireless communication network 50 connected to the server 20. Then the client 1 transmits an encrypted request 60 through the gateway 30. This encrypted request 60 comprises information of which pre-defined algorithm (s) the client 1 supports. When the gateway 30 receives this encrypted request 60, it sends 70 the encrypted request to the origin server 40. The origin server 40 chooses at least one algorithm, associated with a public key and a private key, and transmits a message 80 back to the gateway 30. The gateway encrypts the message and send it 90 to the client 1.
This message 90 comprises the public key and information about which algorithm the server 20 has chosen. When the client 1 receives the encrypted message 90, comprising the public key, it will generate a master secret code, and calculates a signature based on the chosen algorithm, the public key and the master secret code. Thereafter, the client 1 will transmit an encrypted respond 65 to the gateway 30. This encrypted respond 65 comprises the calculated signature. When the gateway 30 receives the encrypted respond 80, comprising the signature, it will decrypt the respond 75 and send it to the origin server 40. The origin server will calculate the master secret code based on the chosen algorithm, the signature received, and the private key. Finally, the origin server 40 sends a final message 85 to the client through the gateway 30. If the origin server 40 has accepted the clients 1 request 60, the server will be able to establish a secure connection between the origin server 40 and the client 1, else the connection will be terminated.
Setting up a secure connection.
Fig. 4 illustrates the message structure for setting up a secure connection according to the invention.
The cryptographic parameters of the secure session are produced by theWTLS Handshake Protocol, which operates on top of the WTLS RecordLayer. When a WTLS client and server first start communicating, they agree on a protocol version, select cryptographic algorithms, optionally authenticate each other, and use public-key encryption techniques to generate a shared secret.
The WTLS Handshake Protocol is described Wireless Transport LayerSecurity Specification dated 30. April 1998 and is a part of the WirelessApplication Protocol.
The WTLS Handshake Protocol involves the following sequence of steps.
When the a WAP session has been set between the phone 1 (the client) and the server 20 (e. g. a bank), and the client (phone 1) wants to establish a secure connection he sends a client hello message 100 as his first message.
This message includes a key exchange list that contains the cryptographic key exchange algorithms supported by the client in decreasing order of preference. In addition, each entry defines the certificate or public key the client wishes to use. The server will select one or, if no acceptable choices are presented, retum a handshakefai)ureaiert and close the secure connection.
In response to the client hello message 100 the server 20 will send a server hello message 101 when it was able to find an acceptable set of algorithms. If it cannot find such a match, it must respond with a handshake-failure alert.
The server hello message 101 will identify the session and set up the parameters need for the session.
The server 20 will furthermore transmit a server certificate message 102. The server certificate message 102 will always immediately follow the server hello message 101, and the purpose of this server certificate message 102 identify the cryptation algorithm selected by the server from the key exchange list included in the client hello message 100. The server certificate message 102 will include a so-called certificate carrying a public key for the selected encryption algorithm. The server certificate message 102 includes information about issuer of the certificate, the beginning and the end of the validity period, and parameters relevant or the public key. The server controls the validity period and when the granted validity period is expired the client has to renew the secure connection. The length of the validity period will typically be in the level of a week or more. The maximum number of session will also have to be defined.
A Server Key Exchange Message 103 will be send as a third message immediately after the server certificate message 102. The server key exchange message 103 is optionally and will be sent by the server 20 only when the server certificate message102 does not contain enough data to allow the client 1 to exchange a pre-master secret. This message 103 conveys cryptographic information to allow the client to communicate the premaster secret: either an RSA public key to encrypt a secret with, or EllipticCurve Diffie-Hellman parameters with which the client can complete a key exchange (with the result being the pre-master secret). As additional KeyExchange Suites are defined for WTLS which include new key exchange algorithms, the server key exchange message will be sent if and only if the certificate type associated with the key exchange algorithm does not provide enough information for the client to exchange a pre-master secret.
Also a forth message-a Server Certificate message 104-is optionally. This message 104 requests a certificate from the client, if appropriate for the selected cipher suite. This message will immediately follow the ServerCertificate message 102 and Server Key Exchange message 103.
In order to inform the client that the server has ended of the Server Hello session, it transmits a Server Hello Done message 105. After sending this message 105 the server 20 will wait for a client response. This message indicates that the server 20 has send messages to support the key exchange, and that the client 20 can proceed with its phase of the key exchange.
Upon receipt of the server hello done message the client should verify that the server provided a valid certificate if required and check that the server hello parameters are acceptable.
If the server 20 asks for an Client Certificate message 107, the client 1 has to transmit such a after receiving a Server Hello Done message 105. This message is only sent if the server 20 requests a certificate. If no suitable certificate is available, the client must send a certificate message containing no certificates. If client authentication is required by the server for the handshake to continue, it may respond with a fatal handshake-failure alert.
Client certificates are sent using the Certificate structure defined previously for server certificates.
Now the phone 1 or the client starts to calculate a 20 byte random number to be used as a Master Secret 106 for the secure sessions. The master secret 106 is used to derive key material needed for Message Authentication Code (MAC) keys and data encryption keys. MAC and data encryption provide data integrity and privacy between co data transfer. That is why, there is a need to use the mutually agreed master secret 106 for a relatively long time.
The processor or the controller 18 of the phone 1 calculates the master secret. A smart card, e. g. the SIM card16, which can be regarded as a tamper resistant device, is used for storage of the sensitive data of the secure session, and performing operations using that sensitive data, so that this data never leaves the card. In practice the secure information will be transferred from the SIM card 16 to the working RAM 17a of the processor 18 but these information will be overwritten when no session is ongoing or when the phone 1 is switched off.
According to the first embodiment of the invention the controller 18 performs the operations needed for the key establishment, e. g., Diffie-Hellman calculation or RSA encryption and complementary calculations. Then the controller 18 persistently stores the resulting secret key (master secret 106) in the SIM card 16. Then the controller 18 performs the key derivation based on the master secret 106 and additional data (e. g., seed), producing key material for MAC calculation and encryption. The key derivation function is security protocol specific. It is typically based on some secure hash function, e. g.,SHA-1.
Preferably the SIM card 16 is provided as a smart card having its own processor, whereby both the operations needed for performing the key establishment and the key derivation based on the master secret may be performed inside the smart card. Then the master secret, and data used to calculate it, would never have to leave smart card. So, the secure session associated with the master secret can be used during a long periodA Client Key Exchange Message 108 will immediately follow the client certificate message 107, if it is sent. Otherwise it will be the first message sent by the client 1 after it receives the Server Hello Done message 105. With this message 108, a pre-master secret is set, either through direct transmission of the RSA-encrypted secret, or by the transmission of EC Diffie-Hellman public key which will allow each side to agree upon the same pre-master secret.
Then the Master Secret 106 is encrypted by using the public key from the server's certificate and the agreed RSA algorithm. The result is send to the server 20 in an encrypted master secret message 109.
A Certificate Verify message 110 is used to provide explicit verification of a client certificate. This message is only sent by the client following a client certificate Message 107 that has signing capability (i. e., RSA certificates).
Both ends has to send finished messages 111 and 112 at the end of the handshake to verify that the key exchange and authentication processes were successful.
The finished messages 111 and 112 is the first messages protected with the just-negotiated algorithms, keys, and secrets. Recipients of finished messages must verify that the contents are correct. Once a side has sent itsFinished message and received and validated the Finished message from its peer, it may begin to send and receive application data 113 over the secure connection. It is a critical or fatal error if a finished message is not preceded by a change cipher spec message at the appropriate point in the handshake.
The value handshake-messages includes all handshake messages starting at client hello up to, but not including, this finished message. The handshake-messages for the finished message sent by the client will be different from that for the finished message sent by the server, because the one which is sent second will include the prior one.
As long as a secure connection is valid application data session 113 may be initiated just by using Client Hello messages 100 and Server Hello messages 101.
Acronyms.
APDU Application Protocol Data UnitAPI Application Programming InterfaceCA Certification Authority CBC Cipher Block ChainingDF Dedicated FileDH Diffie-HellmanEC Elliptic CurveECC Elliptic Curve Cryptography ECDH Elliptic Curve Diffie-HellmanECDSA Elliptic Curve Digital Signature AlgorithmEF Elementary FileGSM Global System for Mobile CommunicationIV Initialization VectorMAC Message Authentication CodeME Management EntityOSI Open System Interconnection PDU Protocol Data UnitPRF Pseudo-Random FunctionSAP Service Access PointSDU Service Data UnitSHA-1 Secure Hash Algorithm SIM Subscriber Identity Module SMS Short Message ServiceSSL Secure Sockets Layer TLS Transport Layer SecurityWAP Wireless Application ProtocolWML Wireless Markup LanguageWMLScript Wireless Markup LanguageScriptWDP Wireless Datagram ProtocolWSP Wireless Session ProtocolWTLS Wireless Transport Layer Security VVTP Wireless Transaction ProtocolThe list above includes the acronyms used in the present text. Detailed discussion and explanation of the acronyms may be found in the technical specifications defining the Wireless Application Protocol on the Internet homepage for WAPFORUM, http ://www. wapforum. org/.