Communications and, more recently telecommunications, are needs deeply engrained in human history. These needs have significantly evolved over time enabling today’s content-rich (text, music, images and video, etc), real-time and multi-location exchanges through electrical, optical or, more broadly, electromagnetic signals conveyed by different media. Among the more versatile is optical fiber.
Why is that hair-thin silica pipe at the core of an ongoing telecommunication and media revolution? Why is it now so widely deployed, from inside houses to super-capacity backbone pipes, covering an impressive array of applications ranging from biomedical, telemedicine, and industrial (automation, sensors, etc), to space, automotive and, of course, telecommunications fields (voice, data, television, music, video, etc)? The answer is simple — because it has versatility, immunity to external parasitical radiation, and the capacity to transmit tremendous flows of data.
An optical telecommunication system basically consists of a coding element that transforms content into an electrical signal, which is then converted into an optical signal through a modulator that hashes the light coming from a continuous laser source. This light then follows its optical path down the fibers in optical cables that are deployed in a variety of environments (buried, ducts, sewers, aerial, etc). After traveling through the network to reach its destination, the optical signal is detected, converted back to an electrical signal, processed and, in the end, the content is recovered and can be displayed. All of this happens in a fraction of a second.
There are different types of fiber depending on the part of the network where it is deployed, whether in the core or in the access network, or in the campus or local area network (single mode fibers, bend insensitive fibers, multi-mode fibers of different class depending on the capacity-reach to be targeted).
There are also different optical cable constructions and deployment methods for different environments. Optical fibers are long cylindrical strands (125μm diameter) of silica glass. They are obtained by melting large silica rods, or preforms, made of several concentric lathes providing light guiding capabilities, in huge furnaces on top of tall drawing towers. They essentially consist of a central core surrounded by a cladding that has a smaller refractive index, thus allowing internal reflection throughout the fiber length. Typical values for core-cladding index differences range from ≈5 to 30×10-3 with core diameters from ≈5 to 10μm. Key characteristics of fibers depend upon the refractive index profile.
Single-mode light guiding is a prerequisite for long-distance transmissions (from a few kilometers to thousands of kilometers) as it is not possible to compensate for distortions caused by the presence of different light modes inside the fiber. Cut-off wavelength, λc, is a common characteristic of single-mode fibers; it defines the range of wavelength values for which fiber can operate in the single-mode regime (above λc). Typical transmission fibers guarantee a single-mode regime above 1300nm or 1450nm, enabling transmission inside the suited standardized ITU transmission windows (Figure 1).
Loss, which will dictate the pace for regular re-amplification of the optical signals, is another key parameter for line fibers. It is mainly controlled by the characteristics of the silica. At 1550nm, where most systems operate, typical values are ≈ 0.19dB/km (i.e. the signal power is divided by 80 every 100km), close to the fundamental limit for silica. Loss values ≈ 0.17dB/km can be achieved using pure-silica-core structures.
Bends can constitute another source of loss for optical signals guided inside a fiber. The smaller the bend radius, the higher the loss (light leakage at the bend). Fiber can also exhibit micro-bend loss once the surrounding cable construction is done. Recently fibers exhibiting reduced bending sensitivity compared to standard Single-Mode Fibers (SMFs, described in the ITU-T G.652 recommendation), have been introduced to better account for the specific needs of access networks. They are now covered in a newly-released ITU-T G.657 recommendation. To meet both recommendations (G.652.D & G.657.B), an interesting solution consists of including a depressed layer (or “trench”) in the cladding to improve mode confinement without sacrificing other characteristics. This type of structure also offers the advantage of being fully compatible with mature manufacturing technologies. Introducing holes in the cladding, which have a similar impact to that of a trench, is an interesting alternative. Issues involving splice-ability, connectorization, or strength, however, need further investigation.
With light being confined in the fiber core over extremely long distances, interactions can happen between light and glass, distorting the transmitted signals. These effects are referred to as “non-linear” effects. Such effects can be linked to the optical channel itself (intra-channel effects) or to neighboring channels (inter-channel effects) when multiple channels are launched in parallel in the fiber to increase throughput (Wavelength Division Multiplexing or WDM). The fiber effective area (Aeff) basically describes the size of the light spot inside the fiber, hence the confinement. The smaller the area, the larger the potential for interaction with the glass. The effectiveness of such non-linear interactions is also conditioned by the “reactivity” of the medium, here silica, measured by the nonlinear coefficient called n2. Line fibers have Aeff ranging from 50 to ≈80μm2 and n2 around 2.7×10-20m2/W, which impose design constraints in the implementation of transmission systems when a few dozen WDM channels are heavily packed together (e.g. state of the art 50GHz spacing corresponding to ≈ 0.4μm).
Chromatic dispersion (group velocity dispersion between all frequency components of an optical signal) is another important parameter because of its double impact on light pulse broadening — through different sub-frequencies traveling at different speeds — and conversion of nonlinear effects to pulse distortions. Fiber optimization for long-haul dense WDM networks implies trade-offs on dispersion values to minimize inter-channel nonlinear effects — which requires high local dispersion value — while also ensuring proper correction of intra-channel non-linear effects. Moreover, fiber dispersion has to be low enough to reduce the amount of Dispersion-Compensating Fibers (DCFs) used inside compensating modules. Most available transmission fibers have positive chromatic dispersions, ranging from 17ps/nm/km (standard SMF) to 4ps/nm/km (firstly-introduced Non-Zero Dispersion-Shifted Fibers). More recently a class of Medium-Dispersion Fibers — with values ranging from 6 to 11ps/nm/km — was also developed targeting dense WDM operations from 1460 to 1625nm (from S- to L-band).
As stated above, DCFs are widely used to periodically compensate for accumulated chromatic dispersion, which is mandatory at high bit-rates (≥10Gbps). Τhey exhibit strong negative chromatic dispersion values (<-100ps/nm/km at 1550nm) and slope (as a function of wavelength) to compensate for the positive dispersion and slope of line fibers. Because of their small and high-index cores, DCFs suffer from high loss and high nonlinearity compared to line fibers, that is Aeff ≈ 20μm2 and n2 ≈ 3×10-20m2/W. From a system point of view, it is important to quantify how much loss and nonlinearity is added and how it impacts performance.
Another source of dispersion arises in single-mode fiber when circular symmetry is broken, yielding a slight birefringence and different group velocity for orthogonal polarization modes. It is referred to as Polarization Mode Dispersion (PMD). PMD of recent transmission fibers is now well mastered and is well below 0.10ps/√km, enabling it to travel ultra-long distances at high channel rates.
Besides single-mode fibers, used for the telecom environment (long distance — metro — FTTx), multimode fibers (MMF) are used in the datacom world (short distance in-building enterprise networks). The large core diameter of such fibers (50 μm or 62.5 μm) offers much more relaxed tolerances in the connectorization between fibers and to sources. Specifically the source to fiber coupling results in cheaper transmitters (by reduced assembly costs), a prerequisite in datacom applications with point-to-point network constructions where (high) costs of a source cannot be shared over multiple users.
The great majority of MMF based datacom systems have been installed in the past for intended bit-rates below 1 Gbit/s. The development of cost-effective, high-speed 850 nm Vertical Cavity Surface Emitting Lasers (VCSELs) combined with laser-optimized graded-index multimode fibers (LO-MMF) has enabled the standardization and use of 1Gbit/s Ethernet over a distance of 550 meters (1000BASE-SX; 1998) and 10Gb/s Ethernet over a distance of 300 meters (10GBASE-SX; 2002) for data-com systems in enterprise networks.
Unlike single mode fibers, multimode fibers carry more than one mode, traveling along the fiber with different velocities (Figure 2).
This introduces a capacity limitation, called intermodal bandwidth, which also depends on the launch conditions of the multimode fiber. These conditions determine the amount of power carried over the different mode (groups).
Around 2002, together with the development of 10GBASE-SX, a new class of 50 μm core diameter multimode fibers was developed, optimized for laser launch. Such fibers are characterized by means of Differential Mode Delay (DMD) and are optimized for a very precise core refractive index design, leading to differences in modal delay times below 0.3 ps/m. The popular acronym for such fibers is OM3.
IEEE 802.3 has started new discussions about the next generation Ethernet; a new task force (IEEE P802.3ba) was approved in December 2007 to prepare two system speeds: 40GbE for the server environment and 100GbE for core net- work and data centre applications, both to be completed by 2010. Again, MMF versions are foreseen using OM3 fibers to distances up to at least 100 meters. Two primary solutions for MMF are under investigation: the parallel solution (10 fibers parallel, e.g. in ribbon) and Coarse Wavelength Division Multiplexing (CWDM), in both cases with 10Gb/s per channel. It is possible that a combination of both will be used.
Cable Construction and Rights of Way Solutions
Whether single-mode or multi-mode fibers, it is necessary to encase the fibers in some form of dedicated cable construction, depending on the targeted cable capacity (from a few fiber pairs to several hundred) and the targeted environment: outdoor, indoor, or some combination of the two (indoor/outdoor). Cable construction should be suited to each particular application, such as: installations in ducts, micro-ducts, directly buried, aerial, attached to walls (inside or outside buildings), or installed inside buildings (subscriber cables, risers). Environmental constraints should also be taken into consideration, such as exposure to extreme temperatures (very hot or deep freezing, depending on location), humidity, water, rodents, etc. Other installation techniques also exist, some of which may require Right of Way negotiations with the utility that owns the conduits (sewers, gas pipes, water pipes, etc.). Often referred to as “Right of Way solutions,” these installations impose specific requirements on cable design. Indeed, such a wide array of cable construction techniques is mandatory to deploy optical fiber everywhere it is needed for proper coverage without distortion.
Content-rich services such as music, images and video have had a significant impact on communications and telecommunications. One of the most versatile media used to convey these signals is optical fiber. There are different types of fiber depending on the part of the network where it is deployed — whether in the core or in the access network, in the campus or local area network (single mode fibers, bend insensitive fibers, multimode fibers) — and the capacity-reach that needs to be targeted. The versatility of fiber optic cable, its immunity to external parasitical radiation, and its capacity to transmit tremendous amounts of data will continue to make it a popular choice for meeting the challenges of a wide variety of environments.