EP2979390A1

EP2979390A1 - Method and device for establishing session keys

Info

Publication number: EP2979390A1
Application number: EP14709651.5A
Authority: EP
Inventors: Aymen BOUDGUIGA; Nouha OUALHA; Alexis Olivereau; Christophe Janneteau
Original assignee: Commissariat a lEnergie Atomique et aux Energies Alternatives CEA
Current assignee: Commissariat a lEnergie Atomique et aux Energies Alternatives CEA
Priority date: 2013-03-28
Filing date: 2014-03-12
Publication date: 2016-02-03
Also published as: US20160044007A1; WO2014154482A1; FR3004041B1; CN105075175A; JP2016514913A; FR3004041A1; US9787651B2

Abstract

The present invention relates to a method and a device for establishing a session key between a source entity and a target entity in a communication network comprising a plurality of communicating entities. The method which is based on the use of symmetric cryptographic primitives offers each entity of the session a protection against attacks by denial of service by the establishment of a session in four or five exchanges of messages.

Description

METHOD AND DEVICE FOR ESTABLISHING KEYS

SESSION

Field of the invention

The invention relates to the field of network communications and in particular that of encrypted communications between the entities of a network. State of the art

The establishment of a session key between two entities of a communication network is a fundamental prerequisite for the implementation of the vast majority of cryptographic services intended to secure the exchanges between these entities. Thus, the protection of data exchanged against attacks intended to modify them or simply to read them is generally based on symmetric cryptographic primitives, in which the entities communicating with each other use the same key when sending (encryption, integrity protection) and receiving (decryption, integrity checking) of a message.

The establishment of a session key is a procedure intended to intervene many times in the lifetime of a communicating entity. A new symmetric key must be used for any secure exchange with any new correspondent. These correspondents are all the more numerous as the entity participates in a scenario favoring interactions between members, for example of the type of sensor network (Wireless Sensor Network, WSN), Machine to Machine (M2M) or Internet of Things (loT ). In addition, a session key is characterized by a limited life, and should be regularly refreshed. Thus, the procedure implemented to establish a session key is particularly important, and requires a neat design.

Several parameters are to be considered to judge the quality of a session key establishment protocol. First, the security of the protocol must be guaranteed. Thus, the confidentiality of the established key and the mutual authentication of the two correspondents must be ensured. Added to this are other security settings, such as Denial of Service (DDoS) protection that protects entities involved in protocol execution from attacks aimed at depleting their energy resources and / or system resources. Second, the session key setup protocol must be efficient in terms of bandwidth requirements and power consumption, and in particular from the point of view of the cryptographic computations implemented. This second criterion is particularly important when the session key establishment protocol must be implemented by entities having only low energy resources, such as a battery, or a low computing capacity and / or memory . Finally, the protocol can offer additional functionalities, such as the interoperability of the authentication mechanisms between two nodes implementing it or the possibility of a centralized control over the exchange of keys and / or the possibility of a definition. centralized security policies accompanying established keys. Three main approaches have been developed to establish session keys between two nodes.

The first family is the key transport, known as Anglicism

"key transport" which consists in securely transporting by encryption one or more secret values from one of the two participants from one session to the other. These secret value transports can be made in one direction only and this mode is known by the Anglicism "one-pass key transport protocol" is to be made in both directions, and this mode is known then under the anglicism "two-pass key transport protocol". The session key is then derived from these secret values.

The second family of session keying solutions is the key agreement, known as the "key agreement". This approach consists of an exchange of public values between the two nodes from which a common session key is recovered by the two entities participating in the exchange without the public values exchanged having to be decrypted. The main known protocol of "key agreement" is the Diffie-Hellman protocol. In terms of resources consumed, mainly at the level of energy consumed in cryptographic operations, the "key agreement" is a costly approach.

A third family of session keying solutions is key distribution, known as key distribution. In this approach, a third entity often called "trusted third party" intervenes to provide the other two participants a secret value that allows them to calculate the session key or the session key itself. However, key distribution also involves direct exchanges between the two participants. Indeed, they must prove their involvement in the protocol, establish the freshness of the messages they send and prove their knowledge of the established secret. Key distribution, although a simple solution to implement, light in terms of cryptographic operations required and energy consumed, has disadvantages not yet solved by existing solutions.

Well known 'key distribution' solutions are those of Needham and Schroeder, or the Kerberos protocol or the MIKEY-Ticket approach. The Needham and Schroeder key transport protocol, presented in the document "Using encryption for authentication in large networks of computers," Communication of the ACM, Volume 21, Number 12, 1978, contains five message exchanges between two entities, one initiator (I) and an answering machine (R) that each share an encryption key with a trusted third party (TC). The exchange of the five messages is shown in FIG. One of the major issues in the Needham and Schroeder protocol is an impersonation attack by an attacker posing as the initiator and replaying the third message (Message3) between the initiator and the responder. The attacker, who can know the key generated by the trusted third party, can then decipher the fourth message and impersonate the session by sending the last message.

The standard Kerberos protocol described in the Kohi and Neuman document The Kerberos network authentication service, September 1993, is illustrated in Figure 2 and is based on the exchange of four messages. The first two messages are exchanged between the initiator and the trusted third party. The last two messages are exchanged between the initiator and the answering machine. A disadvantage of the Kerberos protocol is that it is not symmetrical with respect to the initiator and the receiver. The trusted third party has no assurance that the receiver has been contacted or has agreed to participate in the secure transaction. In fact, the Kerberos protocol is limited in its applications and rather intended to allow access from a client initiator to a resource supposed not to be subject to malicious behavior, such as a printer or a file server.

A known improvement of the Needham and Schroeder protocol is the Otway and Rees protocol illustrated in Figure 3 and described in the authors document "Efficient and timely mutual authentication", 1987. However, this protocol still has the disadvantage that the third party trust has direct interaction only with one of the two participants, here the answering machine.

WO 2009/070075 A1 to Blom et al. Entitled "Key management for secure communication" presents a method for establishing session keys for secure communications between two entities. The method relies on the MIKEY-Ticket key distribution protocol specified in RFC 6043 of Mattsson and Tian, "MIKET-Ticket: Ticket-based modes of key distribution in Multimedia Internet KEYing (MIKEY)". The key management is based on a centralized key management trust service where the initiator and the responder each share a common key with a trusted third party.

But by design, this protocol can be the target of denial of service attacks. The skilled person knows the different forms of denial of service attacks. An attacker can create a denial of service by exploiting implementation errors in communication protocols, for example, if the protocol used is implemented to block nodes when they receive unknown data. An attacker can then make a denial of service attack by sending messages containing erroneous fields to the responder, his target. Another way to perform a denial of service is to initiate a communication with the target and then stop sending messages to block the target in an acknowledgment state and saturate its receiving stack. Finally, denial of service can be distributed by using several attackers at the same time to saturate the target as quickly as possible and to make it more difficult for the attacker to trace.

Thus, the known approaches although offering alternative solutions for establishing session keys, by their disadvantages do not meet all the security needs expected for a session key establishment protocol.

The proposed invention makes it possible to meet these needs. Summary of the invention

An object of the present invention is to provide a method for setting session keys that is secure and protected against denial of service attacks.

Advantageously, the invention offers both an initiating node and an answering node protection against denial of service attacks.

Another object of the present invention is to provide an efficient session key setting method in terms of consumed resources. Thus the method of the invention allows the establishment of a session in four or five message exchanges and relies on the use of symmetric cryptographic primitives.

Advantageously, the present invention will be implemented in the fields of machine-to-machine (M2M) communications security, or in the context of networks of nodes constrained in resources such as sensors and / or actuators, which are among the nodes most resource-intensive and may need to dynamically set session keys.

To obtain the desired results, a method, a device and a computer program product are provided. In particular, in a communication network comprising a plurality of communicating entities, a method for establishing a session key between a source entity (I) and a target entity (R), the source entity sharing with a third-party trusted entity (TC) a first source key K | _e for data encryption and a second source key K | _a for calculating a message authentication code of the source entity, the target entity sharing with said third-party trusted entity a first target key K _Re for data encryption and a second target key K _Ra for calculating a message authentication code of the target entity, the method comprises the steps of:

receive at the third trusted entity, a message from the source entity for restoration of a session with the target entity, the message containing at least identifiers of the source entity, of the target entity, of said entity third-party trust and an authentication code based on the second source key K | _a ;

generate a pair K | _R of keys (K | _Ra , K | _Re ) for the source entity and for the target entity;

encrypting the key pair K _{! R} (K | _Ra , K | _Re ) by the first source key Κ | _θ and by the first target key K _Re ;

send the target entity a message containing at least the key K | _R encrypted and an authentication code based on the second target key K _Ra ;

receiving from the target entity, a message containing at least the identifiers of the source entity, the target entity, said third-party trusted entity and an authentication code based on the second target key K _Ra ; and sending to the source entity a message containing at least the key K | _R encrypted and an authentication code based on the second source key

Kia-

Advantageously, the source key K _{! E} for data encryption and the source key K | _a for the authentication code are derived from an initial key obtained after a step of authenticating the target entity with the third-party trusted entity.

In a variant, the message received from the source entity further comprises a source nonce (Ni) to prove the freshness of the message of the source entity. In another variant, the step of generating the key K | _Ra for the source entity includes the steps of concatenating the key K | _Ra to the identifiers of the target entity, the third-party trusted entity, and the NI nonce, and to encrypt the concatenated set with the first target key K _Re .

Advantageously, the message sent to the target entity further comprises the nonce (Ni) and the authentication code based on the second target key K _Ra is calculated on the nonce. Advantageously, the message received from the target entity further comprises the source nonce (Ni) and a target nonce (N _R ) to prove the freshness of the target message.

In an implementation variant, the step of calculating an authentication code using the second target key K _Ra , the identifiers, the source and target nonces (Ni, N _R ) and the received authentication code based on the second target key K _Ra and the step of checking whether the calculated authentication code is equal to the authentication code received. Advantageously, the message sent to the source entity further comprises the source and target nonces (Ni, N _R ).

Advantageously, the identifiers of the source and target entities are either IPv6 addresses, MAC addresses or URLs.

According to the variants, the nuncio is either a timestamp information, a random number or a counter.

In another variant, the message sent by the third-party trusted entity to the target entity also contains a key encrypted by the keys K | _e and K | _a and where the message received from the target entity and the message sent to the source entity are a single message sent from the target entity to the source entity.

In another variant, the method further comprises a step of sending a message from the source entity to the target entity that contains the identifiers of the source entity and the target entity, the source and target nonces (Ni , N _R ) and an authentication code calculated with the key K | _Ra shared between the source entity and the target entity. The invention further relates to a system for establishing a session key between a source entity and a target entity, the source entity sharing with a third-party trusted entity a first source key K | _e for data encryption and a second source key K | _a for calculating a message authentication code of the source entity, the target entity sharing with said third-party trusted entity a first target key K _Re for data encryption and a second target key K _Ra for computing a code of message authentication of the target entity, the system comprising means for implementing all the steps of the claimed method.

The invention may operate in the form of a computer program product that includes code instructions for performing the claimed process steps when the program is run on a computer.

Description of figures

Various aspects and advantages of the invention will appear in support of the description of a preferred embodiment of the invention, but not limiting, with reference to the figures below: Figure 1 illustrates the message exchanges according to the method of Needham and Schroeder; Figure 2 illustrates the message exchanges according to the method of

Kerberos

Figure 3 illustrates the message exchanges according to the method of Otway and Rees;

Figure 4 is a topological representation of a network infrastructure in which to advantageously implement the invention;

FIG. 5 shows the procedures executed between the initiator, responder and trusted third party entities of the network of FIG. 4 according to the MIKEY-Ticket method;

Figure 6 shows the procedures performed between initiator, responder and trusted third party entities of the network of Figure 4 in an advantageous implementation of the invention;

FIG. 7 shows the procedures executed between the initiator, responder and trusted third party entities of the network of FIG. 1 in an implementation variant of the invention.

Detailed Description of the Invention Figure 4 illustrates an example of an infrastructure of communication 100 in which advantageously implement the invention. For the sake of simplicity of description and not limitation of the invention, the example of FIG. 4 only shows a finite number of entities (or nodes) and connections, but the person skilled in the art will extend the principles described to a plurality and a variety of entities and types of connections (wireless, mobile, very high speed).

The communication network (1 00) comprises fixed or mobile entities that can form an object network (102). Entities can be heavily resource constrained (102-1, 102-n) or resource constrained (1 12-1, 112-m).

The entities with strong resource constraints may be wireless sensors or actuators, having limited computing and / or storage capacities. They can also be active tags. However, an entity that is not intrinsically resource-limited may be temporarily so long as it uses a large portion of its CPU resources for another task, or its battery level reaches a critical threshold value. And this entity can be brought to implement less energy-efficient protocols such as that of the invention.

Entities with lesser resource constraints may be mobile phones equipped with an internet connection and a camera. It can also be interconnection gateways between a network of constrained entities and the Internet. These entities offer more computing power and storage capacity, can have a higher energy reserve (battery, mains power supply) and can communicate over a network, either directly to an internet network (104) such as illustrated or through gateways and intermediate servers (not shown).

The node network (102) may be based on level 2 (e.g., 802.1 5.4 or 802.11) and / or level 3 (e.g., IP) communications between the entities of which it is composed. Following the protocols on which it relies, multicast or broadcast communication schemes can be used.

The present invention may be advantageously applied in the environment of FIG. 4 between two nodes of the network, a source entity that is called an 'initiator' and a target entity that is called a 'responder'. Both entities need to establish a security association with each other. A central key distribution server (106) that is known as a trusted third party is responsible for authenticating the nodes. It may be remote and accessible via a third party communications network (104) which may be a cellular network or the Internet. The trusted third party stores cryptographic data necessary for the authentication of each of the nodes.

In operation, each initiator and responder node authenticates to the central server using its own credentials, identity templates, and independent authentication methods most in accordance with each their own specificities and constraints. Thus, two nodes can for example establish a session key and associate with their respective identities while the latter are respectively validated by means of a smart card for one and a biometric authentication for the other. The trusted server distributes the same two-node session key that wants to establish a security association between them, allowing decorrelated authentications for each node, as well as centralized control over the key establishment and / or the policies that accompany it. these last.

FIG. 5 shows the procedures performed according to the known Mikey-Ticket protocol between an initiator node (I), a responder node (R) and a trusted third party (TC) of the network of FIG. 4.

To facilitate the understanding of the invention, identical notations are used to describe the following figures 5 to 7. Thus, subsequently: Datai _, Data ₂ , ... etc respectively designate the concatenation of the data in the messages 1, 2, ... etc. Those skilled in the art will appreciate that the present invention is not limited to the order given by way of example in the description of the messages for the concatenated data. Thus, for the same message, transmissions of the same data in the form (Datai, Data ₂ ) and (Data ₂ , Datai) are equivalent.

E {K, (Datai, Data ₂ , ...)} designates the encryption of the concatenated data (Datai, Data ₂ , ...) with an encryption algorithm using a key K.

MAC {K, (Datai, Data ₂ , ...)} designates the Message Authentication Code (MAC) on the concatenated data (Datai, Data ₂ , ...) using a K key As explained above, the MIKEY-Ticket protocol has been defined to extend the MIKEY protocol by the use of a trusted third party and is based on the exchange of six messages between the three initiator entities (I), responder (R) and trusted third party (TC).

An initiating node sends the trusted third party a first message in the form of an initialization request "Request_init". The request contains a MAC authentication code which is calculated with its key "Kia". On receipt of this first message, TC checks the validity of the MAC and the authenticity of the data (Datai) sent by the initiator. This data mainly contains an identifier of the answering node with which the initiator node wants to establish a session, a nonce and information on the MAC encryption and computation algorithms supported by the initiator.

The trusted third party generates a "KIR" key and transmits it to the initiator in a "Request_resp" response message. The key K | _R is encrypted by the encryption key "Ki _e " of the initiator. The second message also contains a ticket "Ticketi" for the initiator and a MAC calculated with the key K _{! a} .

On receipt of the message, the initiator verifies the validity of the MAC and retrieves his key K _{! R} which will allow him to derive with the answering machine two keys "KiRa" and "K | _Re ". The key K | _Ra will be used to calculate MACs while the KiRe key will encrypt data.

The initiator then sends the responder a message "Transfert_init" which contains its ticket "Ticketi". This third message contains a MAC calculated with the key K | _Ra .

It should be noted that at this level of the execution of the protocol, the answering node (R) has not yet received the key K | _R and therefore not received the keys K | _Ra and K | _Re . Therefore, (R) can not verify the validity of the MAC that it has just received from the initiating node, and must send to the trusted third party (TC) a message "Resolve_init" to ask it to communicate to it upon receipt of the fourth message, the trusted third party verifies the validity of the answering machine's MAC using the key K _Ra . If the MAC is valid, it sends the answering machine the key K _{! R} encrypted by the key K _Re , in a fifth message "Resolve_resp".

On receipt of the fifth message, the answering machine checks the MAC calculated by the trusted third party and retrieves the key K | _R. Once the key K | _R received, the answering machine calculates the key K | _Ra and checks the value of the MAC calculated by the initiator in the third message "Transfer_init". If the value of the MAC is valid, the responder sends a sixth message "Transfert_resp" to the initiator to acknowledge receipt of the KIR key.

Thus, as already discussed above, the answering node only checks the MAC sent by the initiator in the third message, after having exchanged the fourth and fifth messages with the trusted third party, it can then be the target of denial of service attack. An attacking node may bombard the answering node with an unlimited number of "Transfe nit" messages (third message) to force him to compute a large number of messages "Resolve_init" (fourth message) and thus exhaust all energy and computing resources of the answering node. Figure 6 shows the procedures performed between entities

Initiator (I), Responder (R) and Trusted Third Party (TC) network of Figure 4 in an advantageous implementation of the invention.

An important advantage of the invention lies in the energy efficiency of the method implemented which is obtained by means of a reduced number of exchanged messages. The proposed method makes it possible to establish secure communication between an initiating node and an answering node in five message exchanges in the implementation of FIG. 6, or in four message exchanges according to the implementation of FIG. note that the last message described for both examples is exchanged between the initiator and the responder and corresponds to a confirmation of key by the initiator which can be implicit. The secure exchange of data between the initiator and the responder can begin respectively after the fourth and the third message of Figures 6 and 7.

Those skilled in the art will appreciate that the message exchanges give roles equivalent to the initiating node (I) and the receiving node (R) with respect to the trusted third party (TC). Indeed, the interactions are of type (l) e (TC), (TC) el (l), (R) tt (TC) and (TC) al (R) and thus offer the trusted third party a better control of session key establishment.

The implementation illustrated in FIG. 6 can advantageously be applied when establishing a secure communication between two nodes that do not share any secret directly, but which each share one or more secrets with a trusted third party (TC) . The trusted third party and the initiator initially share two keys (K | _e , K | _a ) where K | _e is a key used for data encryption, and where K | _a is a key used to calculate a code message authentication (MAC). K _{! E} and K _{! A} can be derived from a master key following authentication of the initiator with the trusted third party. For its part, the answering machine (R) also shares a key pair (K _Re , K _Ra ) with the trusted third party.

Those skilled in the art will understand that in order to simplify the description of the variant of FIG. 6, the initiator and the responder each designate a simple entity, but that in a more general case, it may be a group of initiating nodes and / or a group of answering nodes using individual keys and / or group keys.

In order to establish a secure communication channel with the responder (R), the initiator (I) contacts the trusted third party (TC) to create one or more keys for them. To simplify the description of the protocol, in the variant of Figure 6, the trusted third generates a single key KIR for both nodes (I) and (R). Then, the trusted third party sends the encrypted key KIR with the key K | _E to the initiator and send the encrypted key KIR with the key K _RE to the answering machine. The integrity of the messages exchanged is ensured by the addition of MACs, respectively calculated using the keys Kia and K _RA .

In more detail with reference to FIG. 6, the initiator starts the process by sending a first "Messagel" message to the trusted third party. This message contains the identifiers of the initiator (IDi), the trusted third party (ID _T c) and the responder (ID _R ). It also contains a nuncio (Ni) that serves to prove the freshness of the session and to avoid replay attacks. In addition, the initiator adds to the messagel a MAC (MACn) calculated on (ID ,, ID _T c, ID _R and N,) with the key K _! _A. The MAC makes it possible to prove to the trusted third party the authenticity of the data received in the first message.

The identifiers (ID |, ID _T c, ID _R ) may depend on the technology used and the type of network deployed. These identifiers may be, for example, IPv6 addresses, MAC addresses or URLs. In the particular case of IP networks, the identifiers ID | and ID _T c can correspond respectively to the source address and the destination address of the packet I P.

The nonce can be a timestamp information, a random number, or a counter (sequence number). It must be a variable and unique information in time that distinguishes the different executions of the known protocol of Menezes, Van Oorschot and Vanstone, described in the document "Handbook of Applied Cryptography", Chapter 10. The Nuncio can also be formed by the combination of the techniques mentioned above. For example, a nonce can be formed by timestamp information and a random number.

On receipt of the first message, the trusted third party checks the freshness of the Nuncio Ni and uses the key K _{! A} to calculate the MAC (MAC _T ci) on (ID |, I DJC, I DR and Ni). Then, it checks the equality between the MACn received from the initiator and its own MAC (MAC _T ci) in order to verify the integrity of the received message. If the two values are equal, the trusted third party will generate a second message "Message2" to send it to the receiver. The trusted third party starts by generating a key K | _R for initiator and responder. Then he sends the answering machine this key K | _R encrypted by encryption key K _Re . The message sent to the answering machine also contains a MAC (MACTC2) which allows the answering machine to check the integrity of the received data.

In a preferred implementation, the trusted third concatenates the key K | _R to identifiers ID | and ID _T c, and Nuncio N | before encrypting the whole with the key K _Re to obtain an encryption E {K _Re , (K | _R , ID _T c, ID |, N |)}. This second message also contains the identifiers (IDi, I DTC, I DR), the nonce N, and the MAC (MACTC ₂ ). The MAC (MACTC ₂ ) is calculated with the key K _Ra on the following fields: ID _T c, I DR, I DI, NI, and E {K _Re , (K | _R , _TC ID, ID |, N |) }.

On receipt of the second message, the answering machine checks the equality between the received MAC _T c2 and its own MAC value (MAC _R2 ) which it calculates using the key K _RA . If the value of MAC _T C2 is good, the answering machine decrypts the content of E {K _Re , (K | _R , ID _T c, ID |, N |)} to recover the key K | _R.

The answering machine generates a nonce (N _R ) and calculates a MAC (MAC _R |) with the key KIR on the identifier data (ID _R , I DTC, I DI) and nonce (Ni, NR). The MAC (MACRI) will be transmitted by the trusted third party to the initiator to enable him to verify that the answering machine has received the

In a first variant, the responder uses the nonces (Ni, N _R ) and the key K _{! R} as inputs to a pseudo-random function to generate two keys "K | _Ra "and" K | _Re ". The key K | _Ra is used for calculating the MACs, the KiR _e key is used to encrypt data between the initiator and the responder.

In a second variant, the trusted third party generates two keys, "K | _Ra "and" K | _Re "for the initiator and the answering machine. These keys will be sent by the trusted third party to the initiator and the answering machine and will be used respectively to calculate MACs and to encrypt the data.

To simplify the notations, the notation K | _R is subsequently used to designate both the key used to calculate the MACs and for the encryption, independently of the methods which made it possible to generate the keys K | _Ra and K | _Re .

The responder generates a third message "Message3" for the trusted third party. The message contains the identifiers (ID _R ) and (ID _T c), the nuncios (N _R ) and (N,), the MAC _R | and a new MAC (MAC _R3 ), calculated with the key K _Ra on all previous data.

On receipt of the third message, the trusted third party checks the freshness of the nonce N _R and calculates the MAC _T c3 using the key K _Ra and the identifier data (ID _R , ID _T c), of nonces (N _R , Ni) and MAC (MACRI).

The trusted third party then verifies that the computed MAC _T c3 is equal to the MAC _R3 received from the receiver. If the value of the MAC _T c3 received is valid, the trusted third party generates a message "Message4" to the initiator. The fourth message contains the nonces N _R and N |, the MACIR and the key K _{! R} encrypted by the key K _! E.

In a preferred implementation, the key KIR is concatenated with the identifiers of the responder ID _R and the trusted third party ID _T c, and the nonces Ni and N _R before being encrypted with the key K | _e in encryption

E {K | e, (K | _R , _R ID, I DTC, N |, N _R )}).

The message sent by the trusted third party to the initiator also contains a MAC (MAC _T c ₄ ) calculated with the key K _{! A} which allows the initiator to check the integrity of the data received.

Upon receipt of the fourth message, the initiator calculates its MAC (MAC | ₄ ) with its key K | _was then compares the MAC _(MAC-T c ₄₎ sent by trusted third party. If the two MACs are equal, the initiator decrypts the content of E {K | _e, (K | _R, _R ID, ID _c, N |, N _R) with its key K} | _e to recover the KIR key.

Then, the initiator uses the key K | _R to check the value of the MAC (MACRI) that was calculated and sent by the receiver in the third message. For this purpose, the initiator locally calculates a MAC (MACIR) on the same fields as those used by the receiver when calculating the (MACRI), namely the identifiers (ID _R , ID _T c, ID |) and the nuncios (Ni, N _R ). Then, the initiator compares if there is a tie between his MACIR and the MAC ™. If both MACs are equal, it means that the answering machine has received the key K | _R from the trusted third party. To acknowledge the good reception of the fourth message, the initiator generates a message "Message5" to the answering machine.

The initiator sends directly to the responder the fifth message 5 which contains the respective identifiers of the initiator and the responder (I Di and ID _R ), their nuncios (N |, N _R ) and a MAC (MAC | ₅ ) calculated with the key K | _R.

On receipt of the fifth message, the answering machine checks the value MAC received (MAC | ₅ ) by calculating its own MAC (MAC _R5 ). If the two MACs are equal, the responder R concludes that the initiator has indeed received the KIR key from the trusted third party. FIG. 7 shows the procedures executed between the initiator, responder and trusted third party entities of the network of FIG. 4 in an implementation variant of the invention consisting of four exchanged messages.

The initiator starts the process by sending a first message "Messagel" to the trusted third party. The content of the first message is identical to that described with reference to FIG.

After receiving the Messagel, the trusted third party TC generates a nonce (N _T c), as well as the key K | _R for initiator and responder. The nonce N _T c is an optional parameter that allows to add freshness to the second message in the case where the responder combines the freshness of the message to the receipt of a new nuncio of the initiator but also the trusted third party.

The addition of the Nunc (N _T c) can be decided during the establishment of the security policy by the network administrator.

The trusted third digit, with the key K _RE , the nuncio, the key KIR, and this same key K | _R encrypted and protected in integrity by K keys | _e and K _! shared with the initiator. The cipher is sent in a second message "Message2bis" to the answering machine. The message further contains a MAC calculated with the key K _Ra , to allow the responder to verify the integrity of the data received.

After receiving the second message, the responder verifies the integrity and authenticity of the received message using its key K _Ra . If the result of the check is positive, the answering machine decrypts the message to retrieve the shared key K | _R.

Then, the answering machine generates a nunc (N _R ) then sends in a third message "Message3bis", the data received from the third party. trust in the second message TC and containing the key K | _R encrypted and authenticated with K keys | _e and K | _is shared between the TTP and the initiator. To allow the initiator to verify that the answering machine has received the shared key KIR, the answering machine also sends a calculated MAC with this key KIR.

Upon receipt of the third message, the initiator verifies the integrity of the first part of the message generated by the trusted third party by using the key K | _A. If the verification is successful, the initiator retrieves the shared key K _{! R} using its key K _{! E.} Then, it checks the MAC sent by the answering machine using the key K | _R.

The initiator then sends directly to the responder a fourth message which contains the identifiers of the initiator and the responder (ID |, ID _R ), their nonces (N |, N _R ) and a MAC (MAC | ₅ ) calculated with the K key | _R.

When the fourth message is received, the answering machine checks the value of the MAC | ₅ received by calculating his own MAC _R5 . If the two MACs are equal, the responder concludes that the initiator has received the key K | _R from the trusted third party.

Those skilled in the art will appreciate that variations can be made on the method as preferentially described while maintaining the principles of the invention. Thus, the examples described are based on a preferred protocol but it is possible to use other authentication protocols.

The present invention can be implemented from hardware and / or software elements. It may be available as a computer program product on a computer readable medium. The support can be electronic, magnetic, optical, electromagnetic or be an infrared type of diffusion medium. Such media are, for example, Random Access Memory RAMs (ROMs), magnetic or optical tapes, floppies or disks (Compact Disk - Read Only Memory (CD-ROM)). ROM), Compact Disk - Read / Write (CD-R / W) and DVD).

Claims

In a communications network comprising a plurality of communicating entities, a method for establishing a session key between a source entity (I) and a target entity (R), the source entity sharing with a third-party trusted entity (TC) a first source key K| _e for data encryption and a second source key K _!a for calculating a message authentication code of the source entity, the target entity sharing with said third-party trusted entity a first target key K _Re for encryption of data and a second target key K _Ra for calculating a message authentication code of the target entity, the method comprising the steps of:

- receive (602) to the third-party trusted entity, a message from the source entity for the establishment of a session with the target entity, the message containing at least identifiers of the source entity, of the target entity, said third-party trusted entity and an authentication code based on the second source key K _!a ;

- generate (604) a KIR pair of keys (K _Ra , K| _RE ) for the source entity and for the target entity;

- encrypt (604) the pair K| _R of keys (K| _RA , K| _RE ) by the first source key K _e and by the first target key K _RE ;

- send (606) to the target entity a message containing at least the encrypted KIR key and an authentication code based on the second target key K _RA ;

- receive (608) from the target entity, a message containing at least the identifiers of the source entity, the target entity, said third-party trusted entity and an authentication code based on the second target key K _RA ; And send (61 0) to the source entity, a message containing at least the encrypted KIR key and an authentication code based on the second source key K _!A .

2. The method according to claim 1 in which the source key K| _e for data encryption and the source key K _!a for the authentication code are derived from an initial key obtained after a step of authentication of the target entity with the third-party trusted entity.

3. The method according to claim 1 or 2 wherein the message received from the source entity further comprises a source nonce (N|) to prove the freshness of the message from the source entity.

4. The method according to claim 3 wherein the step of generating the key KiRa for the source entity comprises the steps of concatenating the key K| _Ra to the identifiers of the target entity, the third-party trusted entity, and the nonce Ni, and to encrypt the concatenated set with the first target key K _Re .

The method according to claim 4 wherein the message sent to the target entity further comprises the nonce (Ni) and the authentication code based on the second target key KR ₃ is calculated on the nonce.

The method of claim 5 wherein the message received from the target entity further comprises the source nonce (Ni) and a target nonce (N _R ) to prove the freshness of the target message.

The method according to claim 6 comprising a step of calculating an authentication code using the second target key K _ra , the identifiers, the source and target nonces (Ni, N _R ) and the received authentication code based on the second target key K _Ra and a step of checking whether the calculated authentication code is equal to the received authentication code.

8. The method according to claim 7 wherein the message sent to the source entity further comprises the source and target nonces (Ni, N _R ).

9. The method according to any one of claims 1 to 8 in which the identifiers of the source and target entities are either IPv6 addresses, MAC addresses or URLs.

10. The method according to any one of claims 3 to 9 in which the nonce is either timestamp information, a random number or a counter.

1 1. The method according to any one of claims 1 to 10 in which the message sent (706) to the target entity additionally contains a key encrypted by the keys K _!e and K _!a and where the message received from the target entity and the message sent to the source entity are a single message (708) sent from the target entity to the source entity.

12. The method according to any one of claims 6 to 11 further comprising the step of sending (612) a message from the source entity to the target entity which contains the identifiers of the source entity and the the target entity, the source and target nonces (Ni, N _R ) and an authentication code calculated with the key pair for the source entity and the target entity.

13. A system for establishing a session key between a source entity and a target entity, the source entity sharing with a third-party trusted entity a first source key Κ| _θ for data encryption and a second source key K _!a for calculating a message authentication code of the source entity, the target entity sharing with said third-party trusted entity a first target key KR _e for encryption of data and a second target key K _Ra for calculating a message authentication code of the target entity, the system comprising means for implementing the steps of the method according to any one of claims 1 to 12.

14. The system according to claim 13 where the source entity is a group of source nodes and/or the target entity is a group of target nodes using individual keys and/or group keys.

15. A computer program product, said computer program comprising code instructions for carrying out the steps of the method according to any one of claims 1 to 12, when said program is executed on a computer.