What Are Elastic Optical Networks?

Elastic optical networks (EONs) are fiber-optic transport systems that allocate spectral bandwidth in variable-sized slots rather than in the fixed 50 GHz or 100 GHz channels of conventional dense wavelength-division multiplexing (DWDM) systems. By subdividing the available optical spectrum into narrow frequency slots, typically 12.5 GHz wide as specified in ITU-T G.694.1, EONs allow each lightpath to occupy exactly the number of slots its data rate requires. A high-capacity 400 Gbps trunk claim a wide swath of spectrum while a low-rate 10 Gbps connection uses only a narrow portion of the same fiber, without occupying the fixed-grid channel it would occupy in a conventional DWDM network. The result is substantially better spectral efficiency, particularly in networks carrying a diverse mix of traffic types and rates.

EONs emerged from research in the late 2000s and early 2010s as a response to the bandwidth inefficiency of fixed-grid optical systems. They rely on transponders that support flexible modulation formats and variable symbol rates, combined with flexible optical switches at network nodes that can route spectral slices of arbitrary width. The architecture draws on advances in digital signal processing, coherent optical modulation, and software-defined networking to make runtime adaptation of capacity assignments practical.

Flexible Grid and Spectrum Allocation

The defining mechanism of an EON is the flexible frequency grid defined in ITU-T G.694.1, which replaces the rigid fixed-grid channel plan with a continuous spectrum that network control planes can partition at 12.5 GHz granularity. A lightpath is established by assigning a contiguous set of frequency slots on every link along its route, a problem known as routing and spectrum assignment (RSA). Because adjacent lightpaths share boundaries in the spectral domain, the RSA algorithm must also enforce a guard-band constraint to prevent interchannel interference. Arxiv survey papers on optical network routing and spectrum allocation document a wide range of RSA algorithms including integer linear programming formulations, graph-coloring heuristics, and machine-learning-guided search methods developed to solve the RSA problem efficiently at scale.

Adaptive Modulation and Transponders

EON transponders are engineered to vary both symbol rate and modulation order in response to the optical signal quality of each path. A lightpath traversing a short, low-loss fiber span can use a high-order format such as 64-QAM, packing more bits per symbol and occupying fewer spectrum slots per unit of data capacity. A lightpath over a long transoceanic span must use a more power-efficient format such as QPSK or 8-QAM to maintain an adequate signal-to-noise ratio. IET Communications research on dynamic spectrum allocation in flexible-grid networks demonstrates how adaptive modulation combined with dynamic slot assignment allows EONs to respond to changing traffic patterns and fiber conditions without manual reconfiguration by network operators.

Software-Defined Network Control

EONs are typically governed by software-defined networking (SDN) controllers that maintain a global view of spectrum occupancy and signal quality across the network. When a new connection request arrives or an existing lightpath degrades, the SDN controller computes a new RSA solution and pushes configuration updates to optical switches and transponders through standardized control interfaces. This architecture separates the data plane, where photons travel through fiber, from the control plane, where routing decisions are made, allowing centralized optimization of spectrum resources across hundreds of network nodes. ScienceDirect's overview of elastic optical network spectrum management in software-defined networks summarizes how SDN-based control planes have become the standard approach for managing the complexity of elastic spectrum assignment in large-scale deployments.

Applications

Elastic optical networks have applications in a range of fields, including:

• Internet backbone transport, where traffic grows continuously and spectrum efficiency directly affects infrastructure capital expenditure
• Data center interconnect, where high-capacity, variable-rate connections link geographically distributed compute clusters
• 5G fronthaul and backhaul networks, which must carry heterogeneous traffic from many base station types with different capacity requirements
• Scientific research networks, where large data transfers between laboratory sites benefit from on-demand provisioning of high-capacity lightpaths