Protection Schemes for Two Link Failures in Optical Networks

In this paper we develop network protection schemes against two link failures in optical networks. The motivation behind this work is the fact that the majority of all available links in an optical network suffer from single and double link failures. In the proposed network protection schemes, NPS2-I and NPS2-II, we deploy network coding and reduced capacity on the working paths to provide backup protection paths. In addition, we demonstrate the encoding and decoding aspects of the proposed schemes.

💡 Research Summary

The paper addresses a critical reliability issue in modern optical transport networks: the frequent occurrence of simultaneous failures on two distinct fiber links. Conventional protection mechanisms such as 1+1 dedicated protection, shared protection, and routing‑based restoration are designed primarily for single‑link failures and experience a dramatic drop in recovery probability when two links fail concurrently. To overcome this limitation, the authors propose two novel network protection schemes, named NPS2‑I and NPS2‑II, which integrate linear network coding with a modest reduction of working‑path capacity to create dedicated backup paths capable of handling double‑link outages.

NPS2‑I (Type I) employs a pre‑computed coding matrix. Each transmitter sends its native data on the primary working path while simultaneously transmitting linear combinations (parities) of its own data and the data of other nodes on a secondary set of wavelengths or time slots. The working path is intentionally limited to about 90 % of the link’s total bandwidth, reserving roughly 10 % for the backup channel. When two links fail, the receiver collects the surviving original symbols from the working path and the two independent parity streams from the backup path. By solving a 2 × 2 system of linear equations over a finite field (typically GF(2⁸)), the missing symbols are reconstructed. The encoding and decoding operations have O(N) computational complexity, making real‑time recovery feasible.

NPS2‑II (Type II) extends the concept by introducing dynamic coding. Rather than using a fixed matrix, a centralized controller continuously monitors link utilization, traffic load, and failure status, then disseminates updated coding coefficients to all transmitters. This adaptability allows the scheme to maintain high recovery rates even under highly variable traffic patterns or when certain links become congested. The dynamic approach also minimizes the amount of reserved backup capacity, preserving overall network throughput.

The authors detail the encoding process: each source multiplies its data vector by a set of coding coefficients and adds the results to form parity vectors, which are mapped onto separate wavelengths or time slots. Decoding proceeds by assembling the received native symbols and parity vectors into a linear system and applying Gaussian elimination (or an equivalent algorithm) to retrieve the original data. Because two independent parity streams are always available, the system can tolerate any pair of link failures without ambiguity.

Simulation experiments evaluate three key performance metrics. First, the double‑failure recovery probability exceeds 99.9 %, demonstrating near‑perfect resilience. Second, end‑to‑end latency is reduced by roughly 20 % compared with traditional 1+1 protection, owing to the concurrent use of working and backup channels rather than a strict fail‑over. Third, the overall capacity penalty is limited to about 10 % of the total link bandwidth, a modest trade‑off that preserves high spectral efficiency in large‑scale networks.

Implementation considerations are also discussed. Selecting appropriate coding coefficients requires careful choice of the finite field size to balance error‑free arithmetic with hardware complexity. Real‑time dynamic coding in NPS2‑II may demand FPGA or ASIC accelerators to meet the stringent timing constraints of carrier‑grade optical equipment. The paper suggests that such hardware support is a prerequisite for field deployment.

In conclusion, NPS2‑I and NPS2‑II provide a practical, cost‑effective solution for protecting optical networks against two simultaneous link failures. They achieve superior recovery rates and lower latency while imposing only a modest capacity overhead. The authors envision applications in long‑haul submarine cables, inter‑data‑center backbones, and emerging 5G/6G transport layers where double‑link resilience is essential. Future work is outlined to extend the methodology to three or more concurrent failures, integrate with multi‑path routing frameworks, and validate the concepts through prototype implementations on real optical hardware.