Elastic optical networks and flexible-grid DWDM
Optical transport networks carry traffic from data centres, access networks, and mobile systems over fiber. One of the basic techniques used for this purpose is . It allows several independently modulated carriers to share one fiber. Values such as 10 Gbit/s, 40 Gbit/s, and 100 Gbit/s describe equipment profiles. WDM itself does not assign one fixed bit rate to a wavelength.
In a conventional WDM system, the available optical spectrum is divided according to a fixed frequency plan. That plan is simple to operate, but it can leave unused spectrum when the requested service and the available channel profile do not match.
Limitations of fixed-grid WDM
Traditional WDM networks can operate on a fixed grid. ITU-T G.694.1 defines reference frequency plans, including nominal spacings such as 12.5 GHz, 25 GHz, 50 GHz, and 100 GHz, as well as wider options. A deployed system normally supports a selected subset. The channel spacing and equipment profile are fixed for that system, while the bit rate still depends on modulation, symbol rate, coding, and optical reach. A mismatch between the requested service and that profile creates two recurring problems.
1. Underused spectrum
A fixed channel profile has to be reserved even when the requested service uses only part of its supported capacity. The allocation can therefore contain unused capacity or require more than one channel:
- Underutilization: Suppose an equipment profile provides a 100 Gbit/s service in a 50 GHz channel and a request needs 30 Gbit/s. The reserved channel is still the unit of allocation, so part of the supported rate can remain unused. This is one equipment profile. It does not define a universal 50 GHz to 100 Gbit/s mapping.
- Stranded capacity: If a request needs 150 Gbit/s while the selected profile provides 100 Gbit/s per channel, the operator may need two channels. The unused portion and the required channel arrangement depend on the transponder profile, modulation, and routing constraints.
This mismatch between client traffic demands and the optical layer can reduce spectral efficiency. In a loaded network it can also make a new channel or fiber necessary earlier than a more granular allocation would.
2. Electronic aggregation
To use a fixed channel efficiently, an operator can aggregate several client flows in an electronic layer before transport over the optical channel. This function is commonly provided by . Such routers can combine smaller client flows into a larger stream that matches a configured optical service, for example a 100 Gbit/s profile.
This approach adds equipment, power consumption, and processing stages. Depending on the service and architecture, packet handling, buffering, and forwarding in the electronic layer can also add latency. An optical bypass is possible only when the transponders and switching nodes support the required service.
Flexible-grid allocation in EON
Elastic Optical Networks, also known as flexible-grid networks, allocate spectrum with a finer granularity than a fixed-grid plan. The aim is to reserve a frequency range that satisfies the requested signal and the constraints of every link on its route.
In an EON, a frequency slot is identified by a nominal central frequency and a slot width. The flex-grid definition uses a 6.25 GHz granularity for nominal central frequencies around 193.1 THz, while slot widths are multiples of 12.5 GHz. A connection is assigned a contiguous block of slots whose width is supported by the transponder and available on every link of the route. Implementations can support only part of the standardized range.
The allocation is therefore more granular than choosing one fixed channel profile. It does not assign an arbitrary bit rate automatically. The achievable rate and reach remain properties of the selected modulation, symbol rate, coding, transponder, and optical path. The defining feature is the ability to tune the spectrum resource to that supported signal profile.
Advantages of flexible spectrum
- Higher spectral efficiency: Allocating a slot width that matches the selected signal profile can reduce unused spectrum. Guard margins remain equipment and design dependent. More connections can then fit in the same fiber spectrum when the route and node constraints allow it.
- Support for diverse traffic:An EON can combine services with different supported rates on one infrastructure. A high-capacity service can also use several adjacent subcarriers as a "super-channel" when the BVT and switching equipment support that arrangement.
- Potentially fewer electronic stages: When compatible transponders and optical switches are available, an optical path can match a service without electronic aggregation at every intermediate node. The resulting equipment, power, and cooling savings depend on the network design and traffic mix.
- Adaptable resource planning: A flexible grid provides more allocation choices as supported transponders and services change. New profiles still require compatible hardware, signal-quality margins, and control-plane support, but they need not be forced into one legacy channel width.