USE OF QUANTUM-SAFE KEYS IN A NETWORK
Patent Information
- Application Number
- DE502022001985
- Authority / Receiving Office
- DE · DE
- Patent Type
- Patents
- Current Assignee / Owner
- Filing Date
- 2022-01-12
- Publication Date
- 2024-11-14
- Estimated Expiration
- 2042-01-12
AI Technical Summary
Current quantum key distribution (QKD) technologies are limited by high costs and restricted range, making them unsuitable for widespread use in network elements without quantum mechanical means, particularly in access networks, and there is a need for secure key distribution in mobile and IoT applications.
A method and system for generating and distributing quantum-safe keys using high-entropy random numbers, where these keys are generated locally in network elements without quantum mechanical means, using a hierarchical structure with a top-level entropy source and stream multipliers to distribute random numbers and identifiers, allowing for secure communication via symmetrical encryption.
Enables secure communication across a large number of network elements, maintaining security for extended periods with minimal maintenance, as the pre-shared secret can last for over 1000 years with 256-bit AES keys, and the system is scalable and cost-effective for widespread implementation.
Description
[0001] The invention relates to a technical solution for using quantum-safe keys in a network, which in particular makes it possible to make such quantum-safe keys widely available for extensive use in securing communication within a network. The solution presented below relates to a special method of generating quantum-safe keys, which enables their use even by a large number of network elements not themselves equipped with quantum mechanical means, as well as to a special method of distributing the high-entropy random numbers processed for this purpose. The invention includes a corresponding method and a system designed as part of the network for carrying out this method.
[0002] Securing the infrastructure of public networks and the data exchanged over them against potential attacks and manipulation is of ever-increasing importance. A significant part of this security is achieved through the encryption of control data—considered part of the network infrastructure—that manages communication and data traffic, as well as the user data transmitted over the networks by communication participants. Numerous encryption mechanisms have become known for encrypting such data, which can essentially be divided into the categories of asymmetric and symmetric encryption.
[0003] In asymmetric encryption, a public key known to all authorized communication participants is used to encrypt the data, whereas a private key of a participant or institution receiving the data encrypted with their public key is used for later decryption of the data encrypted in this way; this private key was previously generated there as part of a key pair comprising this public and the private key.
[0004] Symmetric encryption methods are significantly more secure than asymmetric encryption, as they use the same secret key for both encrypting and decrypting data. However, a challenge with symmetric encryption is distributing this secret key to authorized users without compromising its security by unauthorized or potential attackers.
[0005] Quantum cryptography offers new possibilities for distributing secret keys usable for symmetric encryption. It makes it possible to generate keys with very high entropy, meaning truly random keys, and to distribute them securely to multiple participants and / or institutions. Quantum Key Distribution (QKD) methods allow such high-entropy keys to be made available simultaneously at at least two geographically distant locations, in such a way that any attempt to attack these keys is immediately detected and cannot remain concealed. However, this requires a quantum channel—that is, a channel for transmitting quantum mechanical states—between the participating parties and their respective institutions using the keys.
[0006] However, the requirements for applying the QKD principle result in a number of disadvantages for this technology. Systems and facilities for using QKD technology are very expensive. Furthermore, the range of the necessary quantum channels is very limited, at least insofar as they are physically implemented using optical fibers that typically transmit quantum states in the networks. The possible range is typically around 100 km, at most a few hundred km, but not, according to current technology, 1000 km or more.
[0007] This results from the attenuation to which the quantum mechanical states transmitted via the quantum channels (optical fibers) are subject. On the other hand, the use of amplifiers within the quantum channels is fundamentally impossible, since amplifying the corresponding quantum mechanical states of signals transporting these states would require reading them out, which would immediately lead to the destruction of these quantum mechanical states. Furthermore, the QKD principle is not applicable to key generation and distribution in mobile networks.
[0008] For the aforementioned reasons, and especially due to the high costs, it can be assumed that the use of QKD methods and the necessary technology will remain essentially limited, at least for the time being, to the infrastructure of core networks or to point-to-point connections with special security requirements (e.g., bank server connections, medical service connections, police database connections, and the like). Accordingly, in the foreseeable future, network elements in access networks will likely only benefit from QKD technology in exceptional cases, and other network elements, such as network access devices like Integrated Access Devices (IADs), as well as end devices or mobile terminal equipment, especially mobile devices like smartphones or tablet PCs, will probably not benefit directly from QKD technology.
[0009] Nevertheless, even in connection with the use of these network elements, there is an urgent need for secure, or rather quantum-safe, data transmission. This is especially true since industrial goods or high-value consumer goods in which mobile communication devices are used, for example in car-to-car or car-to-X communication, Industry 4.0 or the IoT (Internet of Things), are often used for long periods, so that these should preferably be made capable now of withstanding an attack with quantum computers on data they send or receive, i.e., user or control data.
[0010] One possibility for this is to generate the keys using a device that uses quantum mechanical means, such as a Quantum Random Number Generator (QRNG), and to transmit these keys to the network elements that require them for later use using encryption methods that are also considered to be very secure, such as Post Quantum Cryptography, for example based on the principle of key encapsulation.
[0011] However, there is still a risk that a corresponding quantum-safe key could be intercepted during its transmission across a network to a network element intended for its receipt. The consequence would be that any subsequent encrypted communication using this key, the encrypted exchange of control data concerning the network infrastructure, or any other securing of connections using such keys would no longer be secure, and this might even go unnoticed by authorized participants or the operators of the relevant infrastructure for an extended period.
[0012] Furthermore, encrypting communication using PQC instead of quantum-secure keys could potentially represent an alternative. PQC is a group of cryptography methods that can be considered classical in that it does not rely on quantum mechanical effects. Rather, it is a group of methods based on complex mathematical problems that are assumed to be able to withstand the computational capabilities of quantum computers. While absolute security against attempted attacks, at least in theory, has been proven for QKD methods, this does not apply to PQC methods. They are based solely on the assumption that they cannot be circumvented even using quantum computers, although no evidence to the contrary has actually been provided to date.For particularly sensitive areas, the use of quantum-secure keys will therefore continue in the future, despite the associated costs.
[0013] The object of the invention is to make quantum-secure keys widely available in a network and to enable their use even by network elements that do not themselves possess quantum-mechanical means. For this purpose, a suitable method and a system designed to implement this method must be provided.
[0014] The problem is solved by a method having the features of patent claim 1. A system that solves the problem is characterized by the independent claim. Advantageous embodiments and further developments of the invention are provided by the respective subclaims.
[0015] According to the proposed method for using quantum-safe keys in a network to solve the problem, each quantum-safe key used by at least two network elements, which themselves do not have quantum mechanical means, to secure their communication via the network, is generated locally by the aforementioned at least two network elements, i.e., locally in each of these network elements.This is achieved by these network elements generating a quantum-secure key by processing an identical random number obtained via the network with a high entropy relative to its randomness due to the use of quantum mechanical means. This is done using a pre-shared secret previously stored as a shared secret by at least two network elements, for example, through exchange or implementation by the manufacturer or a provider of the network elements, according to an identical processing rule applied by each element. Further details regarding the processing rules suitable for this purpose (hereinafter also referred to as KDF = Key Derivation Function) will follow later.
[0016] With regard to the previously mentioned use of locally generated quantum-safe keys to secure communication between the network elements generating these keys, the use of quantum-safe keys refers in particular to securing a transmission path used by network elements of the network for data exchange and / or to the symmetric encryption of data exchanged between the network elements.
[0017] High-entropy random numbers, which can be used to locally generate quantum-secure keys, can originally be generated and made available by one or more entropy sources located centrally within the network. Such an entropy source could, in principle, also transmit the random numbers it generates using quantum mechanical means directly to the network elements that use them to derive quantum-secure keys at the local level. However, if the aim is to enable broad availability of random numbers for their extensive use by a large number of network elements in a large network that generate quantum-secure keys from them, a hierarchical approach to distributing random numbers, as described below, appears to be significantly more efficient, particularly with regard to good scalability.
[0018] According to the proposed method, a stream consisting of generated random numbers and identifiers representing each of these random numbers is sent to at least one stream multiplier that multiplies this entropy stream, namely forwards it to several network elements of a subordinate structure level, by means of at least one entropy source of high entropy generating random numbers using quantum mechanical means at the top level of a hierarchical structure formed in the network for the execution of the method.Accordingly, each of the network elements generating at least two keys from random numbers at the local level, which simultaneously form the lowest structural level of the hierarchical structure, is supplied with at least one stream of random numbers and identifiers (hereinafter also referred to synonymously as entropy stream) via one or more structural levels of the hierarchical structure superior to the respective network elements below the top structural level, each comprising at least one stream multiplier.
[0019] The method is preferably designed such that at least all units communicating with each other above the lowest structural level during the transmission of streams of random numbers and associated identifiers authenticate each other. In particular, at least one, or each of, if applicable, preferably several entropy sources, authenticates itself to the stream multiplier(s) immediately below the highest structural level when sending an entropy stream to the entropy source(s). Depending on the conditions / criteria defined for the reception of a specific entropy stream by the network elements of the lowest structural level that generate locally quantum-safe keys from the contained random numbers, it may also be provided that all facilities and network elements of different structural levels involved in its generation and transmission authenticate each other.Possible procedures for this will be discussed later.
[0020] In connection with a possible process design and the concept of conceivable distribution programs for the distribution of entropy streams with random numbers used for local key generation within the network, it will become clear later that the stream of random numbers and identifiers received by a respective network element generating keys at the local level does not have to be identical to a stream of random numbers and identifiers generated by (at least one or a single) entropy source. With regard to the aforementioned distribution programs, we will refer to them synonymously below as streaming programs or entropy programs (entropy is distributed, in a sense, in the form of entropy streams with random numbers of high entropy, leading to quantum security with regard to the generated keys).
[0021] With regard to the distribution programs mentioned above, according to which different entropy streams, namely different entropy streams in particular with regard to the type of their accessibility, can be distributed to network elements of the lowest structural level for the purpose of local key generation there, the following should first be pointed out at this point.From random numbers of one of possibly several received entropy streams, a first network element together with at least one second network element generates quantum-secure keys to secure the communication between them via the network, depends on free or authentication-bound access to at least one stream multiplier of the hierarchical structure for both the first network element and the respective second network element and on at least one common pre-shared secret stored at the first network element and at least one of the second network elements.
[0022] It is therefore immediately obvious that a given network element can only generate quantum-safe keys from random numbers derived from entropy streams, provided that these keys are also received by the network element via a stream multiplier accessible to it. On the other hand, it is useless for a first network element to receive a stream and convert a random number from that stream into a key using a pre-shared secret stored within it, if there is not at least one second network element that can both receive the same entropy stream and possess the same pre-shared secret as the aforementioned first network element. In the latter case, the network elements could not generate shared quantum-safe keys, meaning that the possibility of generating shared keys also depends on the presence of the same pre-shared secret in both network elements.
[0023] Regarding the type of entropy stream from whose random numbers a first network element, together with another network element, can generate quantum-safe keys, and its distribution within the network, it can be, for example, an entropy stream of at least one of the following categories: a pure entropy stream, a mixed entropy stream, a closed entropy stream, an open entropy stream, a globally available entropy stream, a nationally available entropy stream, or a regionally available entropy stream. With regard to the distribution of the aforementioned different entropy streams, we will also refer to different distribution programs, or—synonymously—different programs, namely streaming programs or entropy programs, and discuss their details later.A single entropy stream can belong to several of the aforementioned categories simultaneously.
[0024] The processing rule (KDF) already mentioned several times, according to which the at least two network elements calculate a respective quantum-safe key from a respective random number using the pre-shared secret exchanged between them, can, for example, be a bitwise XOR operation between the random number and the pre-shared secret.
[0025] In view of the problems inherent in a bitwise XOR operation (for example, when combining two identical numbers or two bit sequences in which all individual bits of one of the bit sequences are inverse to the individual bits of the other bit sequence) and the relative simplicity of this operation, in order to further increase security, a Key Derivation Function (KDF) standardized by NIST (National Institute of Standards and Technology) should preferably be used as the processing rule for calculating the high entropy random number with the pre-shared secret.
[0026] It is possible, for example, to combine the respective random number with the pre-shared secret using an encryption algorithm such as the Advanced Encryption Standard (AES). In this process, the respective random number is applied as a key to the pre-shared secret, and the resulting ciphertext, generated using the block cipher, then forms the shared quantum-safe key, which can later be used by the at least two network elements to secure their communication.
[0027] With an AES key, 1 Tbit of random numbers can be securely converted into application keys. When these quantum-safe keys are used to encrypt user data, or to encrypt communication between two network elements generating a common key, 1 Tbit of network traffic can again be securely encrypted using AES. For core networks with a bandwidth of 1 Tbit, the AES key must therefore be changed once per second. This means that, using the method according to the invention, one key per second must be generated to secure the communication between at least two network elements, which simultaneously corresponds to the required speed of the entropy stream provided to these at least two network elements for generating the quantum-safe keys.At this speed, it is possible to encrypt 10^12 random number bits, that is, 4 * 10^10 (256 bits) of random number blocks, using a 256-bit pre-shared secret. A year has approximately 3 * 10^7 seconds, so a pre-shared secret can securely calculate 256-bit AES application keys (shared keys of at least two network elements for encrypting their communication) from 256 random numbers for about 1000 years. This, in turn, means that the proposed method can be considered maintenance-free – a single distribution of a 256-bit pre-shared secret is sufficient for at least 1000 years.
[0028] Regardless of the KDF used for local key generation from the random numbers, the network elements store each key generated in the manner described above in association with the identifier obtained with the stream using the random number used for this purpose for later use in a hardware security module (HSM = Hardware Security Module) or encrypted in a main memory of the respective network element (device).
[0029] For the aforementioned authentication of at least the units above the lowest structural level, a PQC (Post Quantum Cryptography) signature method can be used, for example. Furthermore, the use of PQC certificates for authentication is also possible. In the context of forwarding an entropy stream to network elements using random numbers from this stream to generate (shared) quantum-secure keys, authentication between an entropy source at the highest structural level and a stream replicator immediately below the highest structural level can be carried out, for example, as follows. Using an appropriate protocol, the stream replicator requests the entropy source connected to it to send it the SHA256 sum of the PQC key with ID X. The entropy source transmits the SHA256 sum of the corresponding PQC key to the requesting stream replicator.The stream multiplier checks the sum and thus authenticates the entropy source.
[0030] According to one possible embodiment of the method, two or more streams of random numbers and associated identifiers can be fed to a stream multiplier for distribution from the devices of the next higher structural level (superordinate stream multipliers or entropy sources). To increase the security of the transmission, it is proposed in this case that such a stream multiplier concatenate the at least two incoming streams before they are concatenated and forwarded to the structural level below it. This is also achieved, for example, by a bitwise XOR operation of the respective streams or by applying another key derivation function (KDF). In this way, two or more streams consisting of random numbers and identifiers are, in a sense, intermixed.As mentioned earlier, this implies that a stream of random numbers and identifiers received by a network element that generates locally quantum-safe keys from the random numbers does not necessarily have to be an entropy stream as originally emitted in this form by the at least one entropy source or by one of the entropy sources of the top-level structure.
[0031] This means that an entropy stream arriving at a network element at the lowest structural level may only acquire its final form with regard to the specific random numbers it contains in one of the stream multipliers, namely precisely when the respective stream multiplier not only multiplies the stream but also merges at least two entropy streams. When mixing streams through bitwise XORing, appropriate measures can be taken to ensure that the entropy of the random numbers does not deteriorate. For example, ensuring the same length of random numbers in the streams and avoiding XORing identical random numbers, which would result in a null string.However, as already explained, the stream linking can also be done using a different KDF, which has nothing to do with the KDF used for the local generation of quantum-safe keys by network elements of the lowest structure level.
[0032] According to the previously explained process design for stream mixing, the entropy transported by the streams in the form of random numbers is mixed, as shown. If multiple streams are fed and mixed only at some stream multipliers of a structural level, the streams arriving at the elements below this structural level, i.e., at downstream stream multipliers or at network elements that locally generate quantum-secure keys from them, are no longer identical at all transfer points. Therefore, with a corresponding process design, the elements of the lowest structural level, i.e., the network elements that locally generate quantum-secure keys, must communicate not only via the pre-shared secret but also via a stream they share (the program).
[0033] This approach thus offers a concept according to which network elements can choose between different entropy streams, or, as previously mentioned, between different (streaming) programs they use. The individual programs can be distinguished, for example, by the nature of their distributed entropy streams, such as the data rate, their territorial or network-related coverage and availability, and the intended use of the keys generated locally based on the distributed stream(s). In this respect, the following streaming programs, or entropy programs, in particular, can be offered for free or restricted use by network elements for the purpose of key generation: Pure programs: These programs involve forwarding and multiplying a pure entropy source, i.e., the streams emitted by an entropy source with random numbers. The endpoints, especially the users, or their network elements at the lowest structural level, can be spatially distributed worldwide. Every user of a pure program receives the same random numbers at all endpoints of the pure stream. An example of a pure program would be a proprietary entropy stream from a mobile phone provider. Mixed programs: These are programs in which the entropy streams of different entropy sources are multiplied by at least one stream multiplier at at least one structural level, for example, using bitwise XOR or linked using another KDF.Random numbers distributed across mixed programs are only available at a few endpoints (network elements of the lowest structural level), but can also be distributed geographically worldwide. Closed programs: These are programs that can only be used by (PQ)-authenticated user groups, meaning those authenticated and (PQ)-authorized using principles of Post Quantum Cryptography. Closed programs serve specific user groups, such as the police, fire department, companies with special security requirements, intelligence agencies, or the military. If a closed program has its own entropy source, it can also be called an independent closed program.If a closed program does not have its own entropy source, it can be considered a dependent closed program that only runs in a part of the hierarchical structure, namely below the top level of the entropy sources, and is characterized as closed by the requirement for authentication. A closed program can be operated in either a "pure" or a "mixed" mode, using at least one stream multiplier. The performance of closed programs is easily controlled because the number of authorized users is known. Open programs: These are programs that accept any user. No special user authorization is necessary. Requests from users, or rather from the network facilities they operate that are configured to generate keys locally, must be managed via load management.Global programs: These are programs that can be used by globally distributed endpoints (i.e., network elements at the lowest structural level). In global programs, network elements, such as telecommunications terminals, can supply themselves with the same entropy, i.e., the same random numbers, at geographically distributed locations worldwide. Regional programs: This category refers to programs that use regionally distributed endpoints (endpoints as defined in the explanations for global programs). Regional programs are only available within a specific region. Example: European Union, Eifel region. National programs: These are programs that are executed via nationally distributed endpoints (according to the understanding previously discussed in the sections on global and regional programs). National programs are only available within a single country or state. Example: Germany.
[0034] It is immediately apparent from the preceding explanations that an entropy stream can be assigned to several of the aforementioned types of streams, or distribution programs. For example, it is readily understandable that a global or national program can simultaneously be an open or a closed program. A significant advantage of the proposed solution is that each organization, each state, each region can operate its own entropy service, with each service operating independently of the others. Network elements, such as telecommunications terminal equipment, can select one of the entropy channels (entropy streams) for service and, if necessary, change (switch) it. The quantum-secure keys used by the network elements (devices) that generate the shared entropy stream simply need to coordinate accordingly.
[0035] A streaming or entropy program is a dedicated stream of independent random numbers and associated identifiers. An entropy program can be received by multiple endpoints, i.e., by multiple network elements generating quantum-secure keys at the local level, depending on the configuration of the stream multipliers. An entropy program is, in a sense, analogous to a television program or a video stream. Different programs therefore have different content.
[0036] In the case of an implementation of the method in a hierarchical structure with only one entropy source and thus using only one pure (streaming) program, a protocol for the use by network elements (e.g. terminal devices) of the lowest structural level of quantum-secure keys already generated from the corresponding stream by two of these network elements to secure the data traffic between them is designed, for example, as follows: (1) A network element randomly selects a private application key, or rather, one of its locally generated and stored quantum-safe keys, from its available supply of such keys. (2) The network element sends the key's identifier (ID) to the other network element. (3) Based on the ID, the other network element selects the matching key from its key storage. (4) Both network elements use the identical quantum-safe keys (private application keys) to encrypt traffic or applications.
[0037] This means that the two network elements generate common keys from random numbers exclusively from the same entropy stream received from only one entropy source and therefore do not need to first communicate about an entropy program to which these keys are derived when using them.
[0038] If the hierarchical structure, as preferred for the practical implementation of the invention, comprises multiple entropy sources and within this structure, multiple, i.e., different entropy programs are used to distribute the streams of random numbers, both network elements that locally generate quantum-secure keys and then jointly use them must agree on an entropy program and thus on the stream used to generate the keys they are to use. For this purpose, the protocol is extended by several steps at the beginning of the protocol. a. One network element sends another network element a list of available entropy programs, specifically those already being processed. b. The other network element selects one or more entropy programs that both network elements can receive. It then transmits the selected program(s) back.
[0039] The protocol is then executed further according to the previously specified steps (1) to (4).
[0040] According to a particularly advantageous development of the method, the quality of the random numbers distributed across the described hierarchical structure, i.e., their entropy, is monitored and repeatedly validated for this purpose. Such validation and monitoring of the entropy can be performed, for example, by the appropriately designed stream multipliers located below the topmost structural level of the at least one entropy source. However, it can also be performed by a respective entropy source itself and the units directly assigned to it. The validation of the random numbers, or rather their entropy, is performed, for example, using a conventional computer according to statistical methods using special statistical tools and algorithms available for this purpose.
[0041] Possibilities for measuring entropy quality are addressed, for example, in Killmann et al., "A proposal for Functionality classes for random number generators," 2011. According to Killmann et al., different classes are defined with regard to entropy quality, whereby the random numbers in the context of the technical solution described here should preferably be assigned to class PTG.3. If this is the case, a corresponding stream whose random numbers meet the requirements of this class is provided with a confirming signature by the validating element, for example, the classical computer included in a stream multiplier. Otherwise, a corresponding signal is preferably transmitted to the entropy source that originally generated the stream, which evaluates the entropy source in question and thereby prompts it to take appropriate control measures to increase entropy.
[0042] With regard to the recently presented further development of quality assurance of the entropy of the random numbers of the stream used for the local generation of the quantum-safe keys, one can additionally distinguish between managed entropy programs and programs that are basically available to everyone in the programs already presented for their generation and distribution, whereby the latter underlying streams, which are typically not checked for quality in this way, should also be referred to here as "community streams" for the sake of linguistic simplification.
[0043] Commercial, managed entropy streams could differ from "community streams" in that a commercial entropy stream is characterized by: a measured and validated quality of the entropy of the generated random numbers, a certification of the entire process chain, increased reliability.
[0044] Therefore, managed entropy streams can achieve a higher level of confidentiality for security-critical applications. If an entropy stream were corrupted by an attacker, the keys generated by the two participants, or network elements, using their pre-shared secret would no longer be identical, and a secure connection using the keys generated after applying the KDF would no longer be possible. Such an attack would result in a connection disruption, which would be immediately noticed and reported to higher-level management, which can and must take immediate action to counteract the corruption.
[0045] A system capable of solving the task and implementing the previously described procedure for using quantum-safe keys in a network initially comprises at least one entropy source. This entropy source generates random numbers with a high degree of entropy relative to their randomness due to the use of quantum mechanical means. It generates a stream of such random numbers and corresponding identifiers, each representing one of these random numbers.
[0046] The system further comprises at least two network elements, not themselves possessing quantum mechanical means, for example, end devices or internet access points, which nevertheless use quantum-safe keys to secure their communication over the network. These two network elements are configured to process random numbers, which arrive via the network as a stream of random numbers and identifiers representing each of these random numbers, to locally generate quantum-safe keys using a pre-shared secret previously exchanged between them, according to a processing rule applied by the two network elements.
[0047] The possibilities available for such a processing rule (KDF) have already been explained in connection with the explanation of the procedure.
[0048] The at least one entropy source and the at least two network elements of the system proposed for solving the problem are components of a hierarchical structure designed for generating and distributing random numbers within the network. Within this hierarchical structure, the at least one entropy source is part of the top level of the structure or constitutes it (if only one entropy source exists in the network). The at least two network elements, which generate locally quantum-safe keys from incoming random numbers, form the lowest level of the hierarchical structure.
[0049] Between the top and bottom levels of the hierarchical structure, one or more further levels, each with at least one stream multiplier, are formed. Streams of high-entropy random numbers and identifiers assigned to these random numbers are fed to the at least two network elements of the bottom level (which generate quantum-safe keys at the local level) for key generation via this level or these levels, i.e., via at least one stream multiplier. All elements or units located above the bottom level of the aforementioned hierarchical structure are preferably configured to authenticate each other.
[0050] The at least one entropy source is, for example, a quantum random number generator (QRNG) or a quantum computer. Preferably, the QRNG or quantum computer in question is part of a server service that can be addressed by the units in the underlying structural level. This server service can be embedded in a cloud infrastructure.
[0051] Typically, the system is designed so that the network elements that use the random numbers to generate keys and obtain them via the at least one stream multiplier have essentially free access to the stream multipliers located directly above their structural level. However, for special process configurations (i.e., in the context of implementing various streaming programs), it may be necessary for network elements at the lowest structural level to first authenticate themselves to the stream multipliers communicating directly with them in order to obtain streams containing random numbers from these streams. Accordingly, these network elements can also be configured accordingly.
[0052] The network elements that generate quantum-safe keys locally by processing streams of incoming random numbers can be, for example, telecommunications terminal equipment, such as two computers located in different local networks that exchange symmetrically encrypted data over the network (wide area network, internet). However, such a terminal, designed to carry out the process of generating quantum-safe keys by processing incoming random numbers, can also be a mobile device equipped with a Subscriber Identity Module (SIM), such as an IoT device.The telecommunications terminal device in question and its SIM are designed to store at least one pre-shared secret used for the local generation of quantum-safe keys in the SIM (or as part of an eSIM = embedded SIM).
[0053] Furthermore, the network elements configured to implement the process, i.e., to generate quantum-secure keys at the local level, can also be, for example, an Integrated Access Device (IAD) or a network node configured as a peering point, via which the aforementioned IAD and the end devices operated on it are connected to the network, forming a VPN tunnel protected by the generated quantum-secure keys. A corresponding network element can also be a base station of a mobile network.
[0054] The proposed solution presented earlier offers several advantages. As a method for global entropy distribution to generate shared randomness for secure communication, it can be implemented using quantum random number generators and incorporating PQC encryption and authentication. The system's architecture is scalable, specifically globally scalable, and can be implemented modularly.
[0055] It can be used as a community platform or as a commercial product with a corresponding quality of service. Integration into existing network architectures, cloud architectures, and edge cloud architectures is transparent, cost-effective, and minimally invasive. Furthermore, the solution is network-agnostic, meaning it works in wired fiber optic or copper networks as well as wireless networks, such as mobile networks. It is also cost-effective and suitable for the mass market.
[0056] The system is expandable to secure all telematics solutions in traffic, offices, transport, factories, logistics and to secure the Internet of Things (IoT) as well as communication within the framework of Industry 4.0 and other areas.
[0057] The presented method distributes quantum-secure private application keys across public network environments without exchanging secret keys. The key that protects network traffic is generated, processed, and stored locally on the end device and at the remote endpoint in the network—servers, routers, or other mobile devices. It is no longer transferred across unsecured networks. Therefore, technologies such as quantum key exchange are no longer necessary. The quality of the random numbers generated centrally by QRNGs / quantum computers can be verified before transmission and improved if necessary.
[0058] The technical solution presented protects network traffic, provided that it is ensured that The entropy stream originates from an authenticated source. The entropy stream is immutable. The pre-shared secrets and the locally computed private application keys are securely stored and protected from third-party access. The pre-shared secret is distributed in a tamper-proof manner (e.g., using dongles and multiple factors).
[0059] Since these requirements are already feasible today, this patent idea describes a disruptive but easily implementable method for quantum-safe key distribution in digital networks - where the availability of different entropy programs and especially their combination significantly increases the security of the entire architecture.
[0060] Among other things, the following safety considerations are of particular importance in this context: a.) Immutability of the entropy stream: The entropy stream must be immutable so that the same private application keys, or quantum-safe keys, can be calculated by different network elements at different locations. b.) Quality of the pre-shared secret: The pre-shared secret must have high entropy and remain secret, as it is used to translate the random numbers of the entropy stream into secret keys. Each network element that locally calculates keys should have a unique, secret, pre-shared key to ensure the security and uniqueness of the endpoint. c.) Storage of private application keys: The locally generated quantum-safe keys must be kept in secure environments to protect them from unauthorized access.For example, in the case of mobile network deployments, data should be stored on the SIM card of a mobile device, in HLRs / Visitor Location Registers (HLRs / VLRs). The proposed solution also meets the requirements for a legal intercept. If, for instance, only data transfer within the mobile network (from the device to the peering point on the internet) is to be secured, the operator knows the appropriate key and can provide it to the legal intercept stations (legal intercept means that the police can intercept voice and data communications, provided a corresponding court order exists).
[0061] The reliability of entropy distribution can be increased by using disjoint network paths between stream multipliers. This allows part of the entropy to be transmitted via one network path, and another part via a different network path. These network paths can be different fiber optic networks, fiber optic and mobile networks, or fiber optic and satellite networks. This means the method also works globally. It is the responsibility of the receiving stream multiplier to correctly reassemble the different parts of the entropy stream before further processing.
[0062] Temporal multiplexing also increases security, especially in the case of (PQ)-encrypted entropy distributions. In this case, the entropy components are transmitted over disjoint networks at different times and correctly reassembled by the receiving stream multiplier. The possibilities of spatial multiplexing are limited only by the availability of disjoint networks, while temporal multiplexing has no such limitations.
[0063] The security of the process can be further enhanced by encoding control bits, that is, bits that have meaning for the streaming programs and protocols used. For example, specific bits of the entropy stream could determine the subsequent steps. These control bits define the key function (KDF) to be used for the locally generated keys shared by network elements at the lowest structural level. If the entropy value begins with "00", this could mean "permutation". If the entropy value begins with "01", AES encryption could be applied to calculate the application key from the entropy value.
[0064] The data rate of the entropy sources (entropy rate) and the throughput rate of the stream multipliers are generally not uniform and depend on the program. Depending on the required key rate, there can be programs with high and programs with low data rates in relation to the entropy stream. Possible data rates range from once a day for mobile applications to several Mbit / second for specialized closed-source applications. The applications register for an entropy stream that provides a sufficient data rate to secure the application. Encryption of up to 64 GB of data with a key using the AES-256 algorithm is currently considered secure.
[0065] When dealing with mixed entropy streams, a decision must be made as to whether the data rate of the faster or the slower entropy stream takes precedence. Either random numbers from the faster entropy stream are deleted or not forwarded, or a combination of original random numbers from the faster stream and random numbers from the two streams (multiplied by XOR or combined using a different KDF) are passed to the subsequent levels of the stream multipliers. The first method should be chosen for higher random number quality. Controlling the clock rate based on unambiguous time signals is essential for mixing streams with different entropy rates.
[0066] As already explained, a wide variety of applications are possible for the quantum-safe keys generated according to the proposed solution. Some of these are described again below.
[0067] To secure a network, the distributed entropy, provided it is sent encrypted and authenticated using PQC algorithms, is written directly to a secure key store, a Hardware Security Module, for further use. Alternatively, the random numbers can be further modified after receipt in a network element. This requires an additional, pre-distributed, secret but identical key (not to be confused with the pre-shared secret) to be present in two network elements. The random numbers of the entropy stream are secretly modified using a mathematical operation, such as permutation, or any conventional encryption method, such as AES-256, and only then are they written to a secure key store (HSM). The advantage of this method is a significant increase in security, since in addition to the PQC encryption, a secret modification of the distributed entropy also occurs.This allows distributed private application keys to be generated to encrypt any traffic without having previously exchanged the actual application keys over the internet. This is a significant difference from established quantum optical key exchange methods, which potentially expose the application key to public networks.
[0068] Furthermore, communication between gateway routers in home or office networks can be secured. Two routers, or gateway routers, and a network-side counterpart share a common randomness via distributed entropy. This allows the derived traffic to be protected with quantum-safe encryption using the same private application key.
[0069] Mobile devices can also be supplied with entropy via a closed program within the mobile network. This utilizes the OTA / FOTA channel, which has its counterpart at an application on the SIM card. The mobile devices can then secure their communication with each other, with a network counterpart, or with a private server service, provided that pre-distributed, secret keys have been exchanged beforehand. Furthermore, the described function can be integrated into any app on mobile devices to perform application-level encryption. The apps use the entropy services to generate private application keys from the distributed entropy for secure client-server communication with their application server.Communication apps (voice, video, chat) use distributed entropy to implement quantum-safe end-to-end encryption - and provide an interface to negotiate the necessary "secret", for example locally & peer-to-peer via QR codes, NFC communication or Bluetooth technology.
[0070] This can also include securing telematics modules and thus communication with intelligent vehicles, or communication between vehicles themselves. Communication between robots and the edge cloud in a campus network can also be secured using entropy programs and pre-distributed, secret keys. The application of the proposed solution can be extended to securing communication between logistics chains, ships, aircraft, trucks, and trains.
[0071] Furthermore, applications can be found beyond those explicitly aimed at security. Although the primary focus of entropy distribution is on security applications, there are other fields of application in the field of communications. In the field of network coding, for example, to determine the identification of a message, for example, "Was an alarm sent with the message? (yes or no)," the data transmission required for identification can be significantly reduced by using identical random numbers previously distributed at two or more locations. H. Boche and C. Deppe, "Secure identification for wiretap channels; robustness, super-additivity and continuity," IEEE Trans. Inf. Forensics Security, vol. 13, no. 7, pp. 1641-1655, July 2018, discuss such network coding possibilities.
[0072] Data can also be transmitted very securely in mobile networks without dedicated encryption using shared random coding, as discussed in M. Wiese and H. Boche, "Semantic Security via Seeded Modular Coding Scheme and Ramanujan Graphs," IEEE TRANSACTIONS ON INFORMATION THEORY, VOL. 67, NO. 1, JANUARY 2021. Attacks on mobile networks can be fended off with appropriate random coding of messages; see H. Boche and C. Deppe, "Secure identification under passive eavesdroppers and active jamming attacks," IEEE Trans. Inf. Forensics Security, vol. 14, no. 2, pp. 472-485, Feb. 2019.
[0073] All these extended functionalities require a previously securely exchanged set of random numbers, which the solution described here can provide in various sizes, data rates, and access options (including global access).
[0074] The following drawings illustrate exemplary implementations of the technical solution previously described in terms of its basic characteristics, and some aspects of this solution will be discussed in more detail. The drawings show: Fig. 1: a form of the hierarchical structure formed according to the proposed solution for random number distribution in a network with one entropy source generating the random numbers, Fig. 2: a form of the hierarchical structure for the worldwide distribution of random numbers generated by several entropy sources, Fig. 3: the formation of closed groups within the hierarchical structure for random number distribution using closed streaming programs.
[0075] The Fig. 1This shows a form of the hierarchical structure formed according to the proposed solution for distributing random numbers in a network. Each entropy source 1 1 , 1 2 , ... 1 n, arranged in the top level L 0 of this structure, generating a stream of random numbers and associated identifiers - in the example of the Fig. 1It is only one – provides authorized access for stream multipliers 2 11 , 2 12 , ... 2 1n , 2 21 , 2 22 , ... 2 mn at the next lower structure level (the structure comprises structure levels L 0 , L 1 , ... L x with a top structure level L 0 and a bottom structure level L x ). Through (PQ)-authorized access (access requiring authentication using a defined Post Quantum Cryptography procedure), it is prevented that the stream multipliers 2 11 , 2 12 , ... 2 1n , 2 21 , 2 22 , ... 2 mn are hindered in their main task, namely the multiplication and transmission of streams of random numbers. End devices, i.e., network elements 31, 32, ... 3n of the lowest structural level Lx, which generate quantum-safe keys locally using these random numbers, can also access the stream multipliers 211, 212, ... 21n, 221, 222, ... without authentication.2 mn of the structural level immediately superior to the lowest level L x may be permitted. However, it may, on the other hand, particularly in connection with closed streaming programs, be subject to the requirement of authentication.
[0076] The installation of an entropy source takes place in cloud infrastructures across different availability zones (i.e., availability regions) to ensure high availability of the server-side application. This also includes operational monitoring. The number of entropy sources that can be deployed globally when implementing the proposed solution is unlimited.
[0077] Each entropy source 1 1 , 1 2 , ... 1 n must authenticate itself via a DNS (Domain Name Service) name and an associated PQ certificate (Post Quantum Cryptography (hybrid) certificate).
[0078] The layered architecture allows the distribution of random numbers to be scaled across a client-server architecture. The number of users served (network elements operating for the local generation of quantum-secure keys) is specified depending on the structure levels (L1, L2, etc. - the entropy source forms the topmost structure level L0) and the multiplication factor (V-factor) of a stream multiplier 211, 212, ... 21n, 221, 222, ... 2mn. Structural level V-factor: 100 V-factor: 1000 V-factor: 10,000 Level 1 10 2< 10 3< 10 4< Level 2 10 4< 10 6< 10 8< Level 3 10 6< 10 9< 10 12< Level 4 10 8< 10 12< 10 16<
[0079] Thus, a structure comprising 4 structural levels with stream multipliers 2 11 , 2 12 , ... 2 1n , 2 21 , 2 22 , ... 2 mn, each containing 100 structural elements of the next structural level (namely stream multipliers 2 11 , 2 12 , ... 2 1n , 2 21 , 2 22 , ... 2 mn or, with respect to the last level of multipliers 2 11 , 2 12 , ... 2 1n , 2 21 , 2 22 , ... 2 mn , network elements 3 1 , 3 2 , ... 3 n for local key generation), can already have 100 million network elements of the lowest structural level L x with entropy, that is, with high entropy random numbers suitable for generating quantum-safe keys. take care of.
[0080] If the multiplication factor increases to 1,000, 1 quintillion users are reached. An architecture with a multiplication factor of 10,000 achieves this value with just three levels of stream multipliers: 211, 212, ..., 21n, 221, 222, ..., 2mn.
[0081] A stream multiplier 2 11 , 2 12 , ... 2 1n , 2 21 , 2 22 , ... 2 mn with a high multiplication factor (V-factor) can, for example, be set up in a cloud operating center using stateless microservices behind multiple load balancers, i.e. a network device for load distribution. A stream multiplier 2 11 , 2 12 , ... 2 1n , 2 21 , 2 22 , ... 2 mn can also be distributed on the Internet as a community project. If the software required for its function is provided as a container, then in principle any Internet participant could operate a stream multiplier 2 11 , 2 12 , ... 2 1n , 2 21 , 2 22 , ... 2 mn. On the lower structural levels with stream multipliers 2 11 , 2 12 , ... 2 1n , 2 21 , 2 22 , ...2. Within the hierarchical structure formed in the network, P2P (peer-to-peer) architectures could also be implemented (analogous to streaming concepts for media content) or architectures within the framework of security projects (analogous to TOR or VPN).
[0082] The Fig. 2 This shows a form of hierarchical structure for the global, that is, cross-network distribution of random numbers – including across national networks. The corresponding random numbers, or rather the streams transporting them, are generated and transmitted by several entropy sources 11, 12, ... 1n at the top level of the hierarchical structure. In this example, these are entropy sources 11, 12, ... 1n of the national German and Chinese networks, as well as the US network. As can be seen from the Fig. 2As can be seen, below the highest structural level of entropy sources 1 1 , 1 2 , ... 1 n there are several structural levels (two in the example here) with stream multipliers 2 11 , 2 12 , ... 2 1n , 2 21 , 2 22 , ... 2 mn . Among them are some where multiple streams with random numbers arrive, either from different entropy sources 1 1 , 1 2 , ... 1 n or from stream multipliers 2 11 , 2 12 , ... 2 1n , 2 21 , 2 22 , ... 2 mn of the next higher structural level. The stream multipliers 2 11 , 2 12 , ... 2 1n , 2 21 , 2 22 , ... 2 mn , which receive multiple entropy streams, mix these together during their multiplication and transfer to the next lower structure level by means of a bitwise XOR operation or by applying another KDF.
[0083] The following possible sequence of events occurs, for example: Entropy source 11 (D0) generates a random number 0101 and an identifier 4711 that uniquely identifies this random number and transmits it in a stream. Entropy source 12 (CN0) generates a random number 1110 and an identifier 0001 that uniquely identifies this random number and transmits it in a stream. From the aforementioned random numbers and identifiers from both entropy sources 11 and 12, the stream multiplier D1B calculates, for example, the identifier (4711 XOR 0001) and the new random number (0101 XOR 1110) and passes this on to the stream multipliers of the underlying structure level. As already explained several times, the bitwise XOR operation is only one of several possible KDFs (Knowledge Functions) that can be used for this purpose. All the stream multipliers receiving the random numbers and identifiers linked together in the above manner from D1B 2 21 , 2 22 , ...The two sub-levels below now pass on the bitwise multiplied random number from the Chinese and German entropy sources 1 1 , 1 2.
[0084] By linking multiple entropy sources 1 1 , 1 2 , ... 1 n during the multiplication operations, random numbers can be generated that are independent of the entropy source 1 1 , 1 2 , ... 1 n and thus independent of the operator of the entropy source 1 1 , 1 2 , ... 1 n, which consequently increases the security of the overall system. The degree of use of different, independent entropy sources 1 1 , 1 2 , ... 1 n could therefore be a quality characteristic of an entropy multiplication chain.
[0085] The Fig. 3illustrates the formation of closed groups within the hierarchical structure using closed streaming programs for distributing streams with random numbers and identifiers. The part of the hierarchical structure circled on the right in the figure is a closed group that includes an entropy source 1 2 (F0) and distributes random numbers for key generation according to a closed program that requires authentication at all structural levels. Such a closed group is used, for example, to generate quantum-secure keys for the fire department and the police.
[0086] On the left side of the Fig. 3- also enclosed by a circle - this is, for example, the closed user group of a company, within which various locations of the company in question generate quantum-safe keys and use them to secure their communication. This group, however, does not have its own entropy source. This group, which can also be described as a dependent closed group, obtains the random numbers from the entropy source 1 n (D0) of an institutional provider it considers trustworthy, such as an internet service provider.
Claims
1. Method for using quantum-secured keys in a network, in which at least two network elements (31, 32, ... 3n) that do not have their own quantum-mechanical means allow communication that takes place between them over the network to be secured with the aid of common quantum-secured keys, wherein the at least two network elements (31, 32, ... 3n) each locally generate a respective common quantum-secured key from an identical high-entropy random number that they receive via the network by way of an entropy stream consisting of random numbers and identifiers designating in each case one of these random numbers, wherein these network elements (31, 32, ... 3n) a.) receive the entropy stream from a stream multiplier (211, 212, ... 21n, 221, 222, ... 2mn) of a hierarchical structure formed in the network that comprises, in addition to this stream multiplier (211, 212, ... 21n, 221, 222, ... 2mn) and the network elements (31, 32, ... 3n) forming a bottom structural level (Lx) therein, at least one entropy source ( 11, 12, ... 1n), which generates high-entropy random numbers with the aid of quantum-mechanical means, of a top structural level (L0), and b.) generate the quantum-secured key by processing the random number using a pre-shared secret stored beforehand at each of the network elements (31, 32, ... 3n) as a common secret in accordance with a processing rule that they apply, that is to say in accordance with said processing rule.
2. Method according to Claim 1, wherein at least all units communicating with one another above the bottom structural level (Lx) during the transmission of entropy streams consisting of random numbers and associated identifiers authenticate one another.
3. Method according to Claim 2, wherein units authenticating one another in order to transmit entropy streams authenticate themselves by way of a PQC signature method, that is to say post quantum cryptography signature method, or with the aid of PQC certificates.
4. Method according to one of Claims 1 to 3, wherein a stream multiplier (211, 212, ... 21n, 221, 222, ... 2mn) that receives at least two different entropy streams containing random numbers and identifiers from different units from the structural level superordinate thereto links these at least two entropy streams before they are multiplied and forwarded to the structural level subordinate thereto in accordance with a defined mathematical operation.
5. Method according to one of Claims 1 to 4, wherein a first network element (31, 32, ... 3n) and at least one second network element (31, 32, ... 3n) generate quantum-secured keys for securing communication that takes place between them over the network from a random number on the basis of existing free access or access linked to an authentication, for the first network element (31, 32, ... 3n) and for this at least one second network element (31, 32, ... 3n), to at least one stream multiplier (211, 212, ... 21n, 221, 222, ... 2mn) of the hierarchical structure and on the basis of the presence of at least one common pre-shared secret both at the first network element (31, 32, ... 3n) and at the at least one second network element (31, 32, ... 3n), the first network element (31, 32, ... 3n) and the at least one second network element (31, 32, ... 3n) receiving this random number by way of the same entropy stream comprising random numbers and associated identifiers, this entropy stream being an entropy stream from at least one of the following categories, namely: - a pure entropy stream, that is to say an entropy stream that is generated by an entropy source (11, 12, ... 1n) of the top structural level (L0) and is forwarded to the network elements (31, 32, ... 3n) in unchanged form via one or more stream multipliers (211, 212, ... 21n, 221, 222, ... 2mn), - a mixed entropy stream, that is to say an entropy stream that is forwarded to the network elements (31, 32, ... 3n) via one or more stream multipliers (211, 212, ... 21n, 221, 222, ... 2mn) and is formed by a link, carried out at at least one stream multiplier (211, 212, ... 21n, 221, 222, ... 2mn), between at least two different entropy streams provided by one or more entropy sources (11, 12, ... 1n) of the top structural level (L0), - a closed entropy stream, that is to say an entropy stream that is forwarded only to a closed group of network elements (31, 32, ... 3n) comprising the first network element (31, 32, ... 3n) and the at least one second network element (31, 32, ... 3n) following successful authentication with a stream multiplier (211, 212, ... 21n, 221, 222, ... 2mn) and that is generated by an entropy source (11, 12, ... 1n) that supplies entropy streams to only this closed group of network elements (31, 32, ... 3n) or else other network elements (31, 32, ... 3n) via at least one stream multiplier (211, 212, ... 21n, 221, 222, ... 2mn), - an open entropy stream, that is to say an entropy stream that is freely accessible to all network elements (31, 32, ... 3n) accessing the same stream multiplier (211, 212, ... 21n, 221, 222, ... 2mn) as the first network element (31, 32, ... 3n), - a globally available entropy stream, that is to say an entropy stream that is accessible to multiple network elements (31, 32, ... 3n) at locations distributed globally in space in each case via a stream multiplier (211, 212, ... 21n, 221, 222, ... 2mn), - a nationally available entropy stream, that is to say an entropy stream that is accessible to multiple network elements (31, 32, ... 3n) at locations distributed within a country in space in each case via a stream multiplier (211, 212, ... 21n, 221, 222, ... 2mn), - a regionally available entropy stream, that is to say an entropy stream that is accessible to multiple network elements (31, 32, ... 3n) in a geographical region in each case via a stream multiplier (211, 212, ... 21n, 221, 222, ... 2mn) .
6. Method according to one of Claims 1 to 5, wherein the entropy of the random numbers generated by the at least one entropy source (11, 12, ..., 1n) is validated and monitored.
7. Method according to Claim 6, wherein the entropy of the random numbers is validated and monitored by at least one stream multiplier (211, 212, ... 21n, 221, 222, ... 2mn) that receives the random numbers directly from the at least one entropy source (11, 12, ... 1n), that is to say by at least one stream multiplier (211, 212, ... 21n, 221, 222, ... 2mn) in the structural level directly below the top structural level (L0) of the at least one entropy source (11, 12, ... 1n).
8. System for using quantum-secured keys in a network, comprising at least one entropy source (11, 12, ... 1n) that generates random numbers, based on the use of quantum-mechanical means, that exhibit high entropy with regard to their randomness and comprising at least two network elements (31, 32, ... 3n) that do not have their own quantum-mechanical means but still use quantum-secured keys to secure their communication that takes place over the network, wherein a.) the at least two network elements (31, 32, ... 3n) are designed to process random numbers that reach them over the network as an entropy stream furthermore comprising identifiers in relation to the random numbers in order to locally generate quantum-secured keys using a pre-shared secret stored at the at least two network elements (31, 32, ... 3n) as a common secret in accordance with a processing rule applied by the at least two network elements (31, 32, ... 3n), that is to say in accordance with said processing rule, so as to generate quantum-secured keys, and to store each key obtained in the process in association with the identifier of the used random number in a security module; b.) the at least one entropy source (11, 12, ... 1n) and the at least two network elements (31, 32, ... 3n) are part of a hierarchical structure for generating and distributing entropy streams consisting of high-entropy random numbers and associated identifiers, wherein the at least one entropy source (11, 12, ... 1n) is part of a top structural level (L0) or forms same and the at least two network elements (31, 32, ... 3n) form a bottom structural level (Lx) of the hierarchical structure; c.) one or more further structural levels each containing at least one stream multiplier (211, 212, ... 21n, 221, 222, ... 2mn) are formed between the top structural level (L0) and the bottom structural level (Lx) of the hierarchical structure, using which entropy streams consisting of high-entropy random numbers and identifiers associated with these random numbers are forwarded to the at least two network elements (31, 32, ... 3n) of the bottom structural level (Lx) for key generation purposes.
9. System according to Claim 8, wherein the at least one entropy source (11, 12, ..., 1n) is a server service that comprises a quantum random number generator QRNG or a quantum computer.
10. System according to Claim 9, wherein the server service is embedded in a cloud infrastructure.
11. System according to Claim 8 or 9, wherein the at least two network elements (31, 32, ... 3n) locally generating quantum-secured keys are designed to authenticate themselves with stream multipliers (211, 212, ... 21n, 221, 222, ... 2mn) of the structural level superordinate to them in the hierarchical structure.
12. System according to Claim 8 or 11, wherein at least one of the at least two network elements (31, 32, ... 3n) locally generating quantum-secured keys is a telecommunications terminal.
13. System according to Claim 12, wherein the at least one telecommunications terminal is a device equipped with a subscriber identity module SIM for mobile use, wherein this telecommunications terminal and its SIM are designed to store at least one pre-shared secret used to locally generate quantum-secured random numbers in the SIM.
14. System according to Claim 8 or 12, wherein at least one of the at least two network elements (31, 32, ... 3n) locally generating quantum-secured keys is a base station of a mobile radio network.
15. System according to Claim 8 or 11, wherein the at least two network elements (32, 32, ... 3n) locally generating quantum-secured keys are an integrated access device IAD and a network node designed as a peering point, by way of which the IAD and terminals operated thereon are coupled to the network so as to form a VPN tunnel connection protected by way of the generated quantum-secured keys.
