Information-Theoretic Security in Space-Division Multiplexed Fiber Optic Networks

We study the use of space-division multiplexing to achieve information-theoretically provable physical-layer security in optical networks against fiber tapping attacks. Introduction Today's fiber-optic networks are inherently vulnerable to various types of physical-layer attacks [1] [2], most importantly to fiber tapping. Here, an attacker with physical access to the fiber retrieves a portion of the propagating signals by bending the fiber and detecting the evanescent field at the bend (Fig. 1 (a)). By introducing a sufficiently small amount of bending loss (e.g., by tapping at a point of high optical signal power near an optical amplifier), a fiber-tapping eavesdropper can go unnoticed by the legitimate transmitter and receiver. Both the fact that eavesdropping is possible through fairly simple means and that the presence of an eavesdropper can remain unnoticed are physical-layer security concerns. Quantum key distribution (QKD) addresses both concerns by allowing the exchange of a secure key between a transmitter and a receiver while at the same time providing for intrusion detection [3]. However, the benefits of QKD come with stringent limitations in terms of data rate and reach (e.g., 1 Mb/s over 50 km of fiber [4]) as well as with severe problems arising from optical amplifier noise and from interactions between classical communications and QKD signals on a common optical networking infrastructure [5]. Future fiber-optic networks are likely to exploit space-division multiplexing (SDM) as a cost and energy efficient solution to overcome the looming optical networks capacity crunch [6].