NSF/ENG/ECCS-BSF: Collaborative Research: Random Channel Cryptography

使用多模光纤的物理层密钥分发在当今的互联世界中，信息安全至关重要。目前，信息是通过公开密钥加密来保护的，这是基于最先进的计算机无法以有效的方式解决某些数学问题，如质数分解。因此，这些加密方法在计算范式和计算能力的进步下并不安全。因此，量子密钥分发（QKD）近年来受到了学术界和商界的广泛关注。由于光的量子特性，包括不可克隆定理和不确定性原理，量子密钥分配从根本上是安全的。然而，QKD不能满足商业应用日益增长的容量（密钥速率和距离）需求。同时，尽管经典密钥分配（CKD）可以提供更高的容量，但目前提出的光学密钥分配方法都不能保证安全性。相反，现有的光学CKD方法只能通过施加黑客不对称来提供对黑客的威慑：使用于窃听的设备比用于合法用户之间密钥分发的设备复杂得多。鉴于其在当今信息化经济中的重要性，物理层安全密钥生成和分发代表了一个只能通过变革性研究来解决的技术差距。我们提出了一种使用多模光纤的物理层密钥分发方法，我们称之为随机信道加密（RCC），它提供了两个世界的最佳：CKD的容量和黑客不对称性，以及QKD的安全性。随机信道密码学是基于信息论安全的一个核心结果，即只要合法用户能够通过比黑客通道噪声更小的通道访问公共随机源，就可以保证物理层密钥分发的安全性。我们利用通信信道，如具有分布式模式耦合的多模光纤，其本质上是随机的，但由于互易性而确定性对称，用于同时生成和分发密钥。在RCC中，Alice和Bob通过空间中的任意自由度将连续波单模激光发送到随机的、空间的、频谱的和时间的多维通道中，例如多模光纤，并且在相同的空间自由度中都接收到时变的强度。根据测量到的强度，可以在Alice和Bob之间建立一个公共密钥，因为连续波光穿过互反路径，所以它们彼此相关。Alice和Bob都不需要生成密钥。相反，安全密钥是沿着多维通道以分布式方式生成的，并同时可供Alice和Bob使用。RCC的安全性由于黑客不对称而得到增强。在RCC中，Alice和Bob只需要进行少量的测量，而为了破解密钥，窃听者必须同时进行M次测量，其中M是光纤模式的数量，可以是几百个。因此，如果Alice和Bob进行的测量代表了最先进的技术，那么窃听者所需要的测量将比最先进的技术高出几个数量级。我们认为，使用现成的组件，RCC的密钥速率和距离可以分别达到10Gb/s和300公里。我们建议进行研究：确定RCC的性能极限，设计达到这些极限的方法，并进一步证明RCC对一般被动和主动攻击的安全性。总之，RCC有潜力成为商业应用程序的安全、高容量密钥分发方法。该奖项反映了美国国家科学基金会的法定使命，并通过使用基金会的知识价值和更广泛的影响审查标准进行评估，被认为值得支持。

Physical-Layer Key Distribution Using Multimode FibersInformation security is of paramount importance in today's connected world. Currently, information is secured through public-key cryptography, which is based on the inability of the state-of-the-art computers to solve certain mathematical problems such as prime number factorization in an efficient manner. Therefore, these cryptographic methods are not secure against advances in computing paradigms and computing power. As a result, quantum key distribution (QKD) has received significant academic and commercial attention in recent years. QKD is fundamentally secure by virtue of the quantum properties of light including the no-cloning theorem and the uncertainty principle. However, QKD cannot satisfy the increasing capacity (key rate and distance) demand of commercial applications. In the meantime, even though classical key distribution (CKD) can provide higher capacity, none of the optical CKD methods proposed so far can guarantee security. Instead, existing optical CKD methods can only provide deterrence to hacking by imposing hacking asymmetry: making equipment for eavesdropping prohibitively more complicated than that for key distribution between legitimate users. Given its importance in today's information-based economy, physical-layer secure key generation and distribution represents a technology gap that can only be addressed by transformative research. We propose a physical-layer key distribution method using multimode fibers, which we call Random Channel Cryptography (RCC), that offers the best of both worlds: capacity and hacking asymmetry of CKD, and security of QKD.Random Channel Cryptography is based on a central result in information-theoretic security that the security of key distribution at the physical-layer is guaranteed as long as the legitimate users have access to a common source of randomness, through channels that are less noisy than the channel of the hacker. We exploit communication channels such as a multimode optical fiber with distributed mode coupling that is inherently random, but deterministically symmetric as a result of reciprocity, for simultaneous key generation and distribution. In RCC, both Alice and Bob send a continuous-wave single-mode laser through an arbitrary degree of freedom in space into a random, spatially-, spectrally- and temporally-varying multidimensional channel, such as a multimode fiber, and both receive the time-varying intensities in the same degree of freedom in space. A common key can be established between Alice and Bob from the measured intensities, which are correlated with each other because the CW lights traverse the reciprocal paths. Neither Alice nor Bob needs to generate a key. Instead, the secure key is generated in a distributed fashion along the multidimensional channel and becomes simultaneously available to Alice and Bob. Security of RCC is enhanced as a consequence of hacking asymmetry. In RCC, Alice and Bob only need to make a small number of measurements whereas, in order to break the key, the eavesdropper must make M simultaneous measurements, where M is the number of fiber modes, which could be on the order of several hundred. Thus, if measurements performed by Alice and Bob represent the state-of-the-art, measurements required of the eavesdropper will be several orders of magnitude beyond the state-of-the-art. We believe that the key rate and distance of RCC can be 10Gb/s and 300 km, respectively, using off-the-shelf components. We propose to conduct research to:- determine the performance limits of RCC, - design methods to reach those limits, and further - prove the security of RCC against general passive and active attacks. In conclusion, RCC has the potential to become a secure, high-capacity key-distribution method for commercial applications.This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.