Cwdm value of carrier frequencies. What technologies can operators use to enhance the capabilities of existing optical networks? Line quality assessment

Questions often arise about what is the difference between CWDM (Coarse Wavelength Division Multiplexing) and DWDM (Dense Wavelength Division Multiplexing) technologies, besides the different number of channels. The technologies are similar in the principles of organizing communication channels and input-output channels, but have completely different degrees of technological precision, which significantly affects the parameters of the line and the cost of solutions.

Number of wavelengths and channels CWDM and DWDM

CWDM wavelength division multiplexing technology involves the use of 18 wavelengths 1), while precision wavelength division multiplexing DWDM can use 40 wavelengths or more.

CWDM and DWDM frequency grid

Channels in CWDM technology are divided by wavelength, in DWDM - by frequency 2). The wavelength is calculated secondarily from the ratio of the speed of light in a vacuum to the frequency. For CWDM, a wavelength grid with a step of 20 nm is used; for standard DWDM systems, frequency grids are 100 GHz and 50 GHz; for high-density DWDM, 25 and 12.5 GHz grids are used.

CWDM and DWDM wavelengths and frequencies

CWDM technology uses wavelengths from the range 1270 - 1610 nm. Taking into account the tolerances and bandwidth of the filters, the range expands to 1262.5 - 1617.5, which is 355 nm. we get 18 wavelengths.

For DWDM with a 100 GHz grid, the carriers are located in the range from 191.5 (1565.50 nm) THz to 196.1 THz (1528.77 nm), i.e. a range of 4.6 THz or 36.73 nm wide. Total 46 wavelengths for 23 duplex channels.

For DWDM with a 50 GHz grid, the signal frequencies are in the range 192 THz (1561.42 nm) - 196 THz (1529.55 nm), which is 4 THz (31.87 nm). There are 80 wavelengths here.

CWDM and DWDM amplification capability

Wavelength division multiplexing systems based on CWDM technology do not involve amplification of a multi-component signal. This is due to the lack of optical amplifiers operating in such a wide spectrum.

DWDM technology, on the contrary, involves signal amplification. The multi-component signal can be amplified with standard erbium amplifiers (EDFA).

Operating range CWDM and DWDM

CWDM systems are designed to operate on lines of relatively short length, about 50-80 kilometers.

DWDM systems allow data transmission over distances much greater than 100 kilometers. In addition, depending on the type of signal modulation, DWDM channels can operate without regeneration at a distance of more than 1000 kilometers.

Notes

1) At the beginning of 2015, manufacturers of optical modules, including SKEO, introduced CWDM SFP modules with a wavelength of 1625 nm. This wavelength is not specified by ITU G.694.2, but has found use in practice.

2) Frequency grids for CWDM are described in the ITU G.694.2 standard, for DWDM - in the G.694.1 standard (revision 2).

Questions often arise about what is the difference between CWDM (Coarse Wavelength Division Multiplexing) and DWDM (Dense Wavelength Division Multiplexing) technologies, besides the different number of channels. The technologies are similar in the principles of organizing communication channels and input-output channels, but have completely different degrees of technological precision, which significantly affects the parameters of the line and the cost of solutions.

Number of wavelengths and channels CWDM and DWDM

CWDM wavelength division multiplexing technology involves the use of 18 wavelengths 1), while precision wavelength division multiplexing DWDM can use 40 wavelengths or more.

CWDM and DWDM frequency grid

Channels in CWDM technology are divided by wavelength, in DWDM - by frequency 2). The wavelength is calculated secondarily from the ratio of the speed of light in a vacuum to the frequency. For CWDM, a wavelength grid with a step of 20 nm is used; for standard DWDM systems, frequency grids are 100 GHz and 50 GHz; for high-density DWDM, 25 and 12.5 GHz grids are used.

CWDM and DWDM wavelengths and frequencies

CWDM technology uses wavelengths from the range 1270 - 1610 nm. Taking into account the tolerances and bandwidth of the filters, the range expands to 1262.5 - 1617.5, which is 355 nm. we get 18 wavelengths.

For DWDM with a 100 GHz grid, the carriers are located in the range from 191.5 (1565.50 nm) THz to 196.1 THz (1528.77 nm), i.e. a range of 4.6 THz or 36.73 nm wide. Total 46 wavelengths for 23 duplex channels.

For DWDM with a 50 GHz grid, the signal frequencies are in the range 192 THz (1561.42 nm) - 196 THz (1529.55 nm), which is 4 THz (31.87 nm). There are 80 wavelengths here.

CWDM and DWDM amplification capability

Wavelength division multiplexing systems based on CWDM technology do not involve amplification of a multi-component signal. This is due to the lack of optical amplifiers operating in such a wide spectrum.

DWDM technology, on the contrary, involves signal amplification. The multi-component signal can be amplified with standard erbium amplifiers (EDFA).

Operating range CWDM and DWDM

CWDM systems are designed to operate on lines of relatively short length, about 50-80 kilometers.

DWDM systems allow data transmission over distances much greater than 100 kilometers. In addition, depending on the type of signal modulation, DWDM channels can operate without regeneration at a distance of more than 1000 kilometers.

Notes

1) At the beginning of 2015, manufacturers of optical modules, including SKEO, introduced CWDM SFP modules with a wavelength of 1625 nm. This wavelength is not specified by ITU G.694.2, but has found use in practice.

2) Frequency grids for CWDM are described in the ITU G.694.2 standard, for DWDM - in the G.694.1 standard (revision 2).

The technology packed wavelength division multiplexing (Dense Wave Division Multiplexing, DWDM) is designed to create a new generation of optical backbones running at multi-terabit speeds and. Information in fiber-optic communication lines passed at the same time a large number of light waves. DWDM networks operate on the principle of channel switching, each light wave is a single spectral channel and is essential information.

Opportunities of DWDM

The number of channels in a single fiber - 64 light beams in the 1550 nm window transparency. Each light wave transmits information at 40 Gb/s. hardware development is also underway with data rates at speeds of up to 100 Gbit/s and Cisco, are already in progress to develop such technology.

In DWDM technology has predecessor - wavelength division multiplexing technology (Wave Division Multiplexing, WDM), which utilizes four spectral channel transmission windows 1310 nm and 1550 nm, with a carrier spacing of 800-400 GHz. Multiplexing DWDM called "densified" due to the fact that it uses a considerably smaller distance between the wavelengths than the WDM.

Frequency plans

At present, two of the frequency plan (ie a set of frequencies that are separated from each other by a constant value) are defined recommendation G.692 Sector ITU-T:

  • The frequency plan pitch (spacing between adjacent frequency channels) of 100 GHz (0.8 nm = YES), whereby the data transmission wave 41 is applied in the range of 1528.77 (196.1 THz) to 1560.61 nm (192.1 THz);
  • Frequency plan in increments of 50 GHz (YES = 0.4 nm), allowing you to transfer in the same range of 81 wavelengths.
  • Some companies also produced equipment, the highly-called wavelength division multiplexing equipment (High-Dense WDM, HDWDM), capable of operating with a frequency up to 25 GHz increments.

The main problem in the construction of super-dense DWDM systems is that with decreasing frequency step there is an overlapping of the spectrum of adjacent channels and there is blurring of the light beam. That leads to an increase in the number of errors and the inability to transmit information on the system

Frequency plans of DWDM

In the following channel plans are currently being used for various types of DWDM systems, CWDM, HDWDM, WDM.

Frequency plans DWDM

Optical fiber amplifiers

The practical success of DWDM technology in many ways defined the appearance of a fiber-optic amplifiers. Optical devices directly amplify light signals in the 1550 nm band, eliminating the need for intermediate conversion to electrical form, as do the regenerators used in the SDH network. The disadvantage of systems of electric signal regeneration is that they have to take a certain type of coding, which makes them quite expensive. Optical amplifiers, "transparent" transmission information, allow to increase the line speed without the need to upgrade the amplifier units. Length of the section between the optical amplifiers can reach 150 km or more, which provides economical DWDM backbones generated in which multiplex section length is today 600-3000 km with use of 1 to 7, the intermediate optical amplifiers.

Recommendation ITU-T G.692 defined three types of amplifying sections, ie sections between two adjacent multiplexers, DWDM:

  • L (Long)- plot consists of a maximum of 8 spans of fiber-optic communication lines and 7 of optical amplifiers, the maximum distance between the amps - up to 80 km with a maximum total length of the section of 640 km;
  • V (Very long)- plot consists of a maximum of 5 spans of fiber-optic communication lines and 4 optical amplifiers, the maximum distance between the amps - up to 120 km with a maximum total length of 600 km section;
  • U (Ultra long)- plot without repeaters up to 160 km

Restrictions on the amount of coasting and long associated with the degradation of the optical signal in the optical amplification. Although the optical amplifier restores the signal strength, it does not fully compensate for the effect of chromatic dispersion (i.e. propagation of different wavelengths at different rates, due to which the signal at the receiving end is "smeared" fibers) and other nonlinear effects. Therefore, to build a more extensive highways need to be installed between the reinforcing portions DWDM multiplexers performing signal regeneration by converting it into electrical form and back. To reduce non-linear effects in the DWDM signal limitation also applies power systems.

Typical topologies

Ultralong two-point connection on the basis of terminal multiplexers, DWDM

DWDM circuit with input-output in the intermediate nodes

Ring topology

The ring topology provides survivability DWDM network through redundant paths. traffic protection methods used in DWDM, similar to the methods in SDH. To some the connection was secured, two paths are established between its endpoints: main and reserve. Multiplexer endpoint compares the two signals and selects the best signal quality.

Ring DWDM multiplexers

The mesh topology

With the development of DWDM networks are increasingly mesh topology is used, which provides the best performance in terms of flexibility, performance, and resiliency than other topologies. However, to implement a mesh topology, you must have optical cross connects (Optical Cross-Connector, PL), which not only add waves to the overall transit signal and outputting them out, as do the multiplexer input-output, but also support arbitrary switching between optical signals transmitted waves of different lengths.

Mesh DWDM

Optical multiplexers IO

Passive muliplexers used in DWDM networks (without power supply and active conversion) and active multiplexers, demultipleskory.

Passive multiplexers Active multiplexers
The number of light waves output lowThe number of light waves is limited to the applicable frequency plan and a set of light waves
It allows you to display and input signal is a light wave without changing the overall spectrum of the light beamIt does not introduce additional attenuation because it produces a complete demultiplexing of all channels and converting into electrical form
Introduces additional attenuationIt has a high cost
It has a budget cost

Optical cross-connects

In networks with mesh topology is necessary to provide the flexibility to change the route of the wave of connections between network subscribers. Such capabilities provide optical cross-connects, to guide the any of the waves at any output port from each input port signal (of course, provided that no other signal of this port does not use the wave must perform another broadcast wavelength).

There are optical cross-connects two types:

  • Optoelectronic cross connectors with intermediate conversion to electrical form;
  • all-optical cross-connects, or photonic switches.

MicroElectro Mechanical System, MEMS

Factors to be considered in the construction of DWDM systems

Chromatic dispersion

Chromatic dispersion- as a result of its influence, as it propagates through the fiber, the pulses constituting the optical signal become wider. When transmitting signals over long distances pulses can be superimposed on the adjacent, making it difficult for accurate recovery. With increasing speed of the transmission optical fiber length and chromatic dispersion effect increases. To reduce the effect of chromatic dispersion on the transmitted signals, dispersion compensators are applied.

Polarization Mode Dispersion

PMD occurs in an optical fiber due to the difference in the propagation velocities of the two mutually perpendicular polarization mode components, which leads to distortion of the transmitted pulses. The reason for this phenomenon is the heterogeneity of the geometric shape of the optical fiber. Effect of polarization mode dispersion on the transmitted optical signals with increasing rate with increasing number of channels and sealing system with increasing fiber length.

Stimulated backscatter Mandelstam - Brillouin, the essence of this phenomenon is to create an optical signal of periodic domains with varying refractive index - a kind of a virtual diffraction grating, passing through which signals propagate like the acoustic wave. Reflected this virtual grid signals are added and amplified to form a reverse optical signal with the Doppler frequency down. This phenomenon leads to an increase in the noise level and prevents the spread of the optical signal, since a large part of its power is dissipated in the reverse direction. Often mistakenly called this phenomenon reflected acoustic wave.

Phase modulation at high power levels of the laser signal modulation of its own phase of the signal can occur. This modulation extends the range and broadens or compresses the signal in time, depending on the sign of the chromatic dispersion. In dense WDM systems, self-modulation signal with an expanded spectrum signals may be superimposed on the adjacent channels. Phase modulation signal is increased with increasing power, increasing the transmission rate and with a negative chromatic dispersion. Influence of phase modulation is reduced at zero or a small positive chromatic dispersion

Cross-phase modulation the phenomenon resulting signal modulates the phase of one channel signals from neighboring channels. Factors affecting the cross-phase modulation, coinciding with the influencing factors the phase modulation. In addition, cross-phase modulation effect depends on the number of channels in the system.

Four-wave mixing, is shown at the threshold power level laser, in which case the non-linear characteristics of the fiber leads to the interaction of three waves and the fourth wave of the new appearance, which may coincide with the frequency of another channel. Such overlay frequency increases the noise level and signal reception difficult

Insertion EDFA amplifier noise, the reason for this phenomenon - the power of the amplified spontaneous emission that occurs due to the design features edfa amplifiers. In the process of passing through the amplifier to the useful component of the optical signal is added to the noise, thereby reducing the ratio of "signal / noise" as a result of the signal can be received in error. This phenomenon limits the amount of in-line amplifiers.

DWDM technology

Dense wavelength-division multiplexing (DWDM) is modern technology transmission of a large number of optical channels over one fiber, which underlies the new generation network technologies. Currently, the telecommunications industry is undergoing unprecedented changes associated with the transition from voice-based systems to data transmission systems, which is a consequence of the rapid development of Internet technologies and a variety of network applications. With large-scale deployment of data networks comes a modification of the network architecture itself. This is why fundamental changes are required in the principles of network design, control and management. The new generation of network technologies is based on multi-wavelength optical networks based on dense wavelength-division multiplexing (DWDM).

Description of technology

The most important parameter in dense wave multiplexing technology is undoubtedly the distance between adjacent channels. Standardization of the spatial arrangement of channels is needed, if only because on its basis it will be possible to begin conducting tests for the mutual compatibility of equipment from different manufacturers. The Telecommunications Standardization Sector of the International Telecommunication Union ITU-T has approved a DWDM frequency plan with an inter-channel spacing of 100 GHz (nm), (Table 1). At the same time, a major debate continues around the adoption of a frequency plan with an even smaller channel spacing of 50 GHz (nm). Without understanding the limitations and benefits of each spectrum plan, carriers and organizations planning to expand their network capacity may face significant challenges and unnecessary investment.

100 GHz grid.

The table on the right shows 100 GHz frequency plan grids with varying degrees of channel sparsity. All grids except one 500/400 have equally spaced channels. Uniform distribution of channels allows you to optimize the operation of wave converters, tunable lasers and other devices of an all-optical network, and also makes it easier to build it up.

The implementation of a particular frequency plan grid largely depends on three main factors:


  • type of optical amplifiers used (silicon or fluorozirconate);

  • transmission speeds per channel - 2.4 Gbit/s (STM-16) or 10 Gbit/s (STM-64);

  • influence of nonlinear effects.
Moreover, all these factors are strongly interconnected.

Standard silicon fiber EDFAs have one drawback - large gain variation in the region below 1540 nm, which leads to lower signal-to-noise ratios and gain nonlinearity in this region. Both very low and very high gain values ​​are equally undesirable. As the bandwidth increases, the minimum signal-to-noise ratio allowed by the standard increases - for example, for the STM-64 channel it is 4-7 dB higher than for STM-16. Thus, the nonlinearity of the silicon EDFA gain more strongly limits the zone size for STM-64 multiplex channels (1540-1560 nm) than for STM-16 channels and lower capacity (where almost the entire silicon EDFA gain zone can be used, despite the nonlinearity) .

50 GHz grid.

A denser, yet unstandardized frequency grid plan with an interval of 50 GHz allows more efficient use of the 1540-1560 nm zone in which standard silicon EDFAs operate. Along with this advantage, this grid has its disadvantages.

In- first, with decreasing interchannel intervals, the influence of the four-wave mixing effect increases, which begins to limit maximum length inter-regeneration line (line based only on optical amplifiers).

In- second,The short inter-channel distance of 0.4 nm may limit the ,possibility of multiplexing STM-64 channels. As can be seen from the figure, multiplexing STM-64 channels with an interval of 50 GHz is not permissible, since then the spectra of adjacent channels overlap. Only if there is a lower transmission rate per channel (STM-4 and below), spectrum overlap does not occur.

IN- third, at 50 GHz, the requirements for tunable lasers, multiplexers and other components become more stringent, which reduces the number of potential equipment manufacturers and also leads to an increase in its cost.

DWDM multiplexers

DWDM multiplexers (unlike more traditional WDM) have two distinctive features:


  • using only one transparency window of 1550 nm, within the region of C-band 1530-1560 nm and L-band 1570-1600 nm;

  • small distance between multiplex channels, 0.8 or 0.4 nm.

In addition, since DWDM multiplexers are designed to work with a large number of channels up to 32 or more, along with DWDM devices in which all channels are multiplexed (demultiplexed) simultaneously, new devices that have no analogues in WDM systems and operate in the addition mode are also allowed or outputting one or more channels to/from a main multiplex stream represented by a large number of other channels. Since the output ports/poles of a demultiplexer are assigned to specific wavelengths, the device is said to perform passive wavelength routing. Due to the short distances between channels and the need to work with a large number of channels simultaneously, the manufacture of DWDM multiplexers requires significantly greater precision compared to WDM multiplexers (usually using transparency windows of 1310 nm, 1550 nm, or additionally the wavelength region in the vicinity of 1650 nm). It is also important to ensure high near-field (directivity) and long-range (isolation) crosstalk performance at the poles of a DWDM device. All this leads to a higher cost of DWDM devices compared to WDM.

Figure "a" shows a typical DWDM multiplexer circuit with a mirror reflective element. Let's consider its operation in demultiplexing mode. The incoming multiplex signal reaches the input port. This signal then passes through the plate waveguide and is distributed over multiple waveguides representing an AWG (arrayed waveguide grating) diffraction structure. As before, the signal in each of the waveguides remains multiplexed, and each channel remains represented in all waveguides. Next, the signals are reflected from the mirror surface and, as a result, the light fluxes are again collected in the waveguide-plate, where they are focused and interfered - spatially separated interference intensity maxima are formed, corresponding to different channels. The geometry of the waveguide-plate, in particular the location of the output poles, and the waveguide lengths of the AWG structure are calculated so that the interference maxima coincide with the output poles. Multiplexing occurs in reverse.

Another method of constructing a multiplexer is based not on one but on a pair of waveguide plates (Fig. b). The principle of operation of such a device is similar to the previous case, except that here an additional plate is used for focusing and interference.

DWDM multiplexers, being passive devices, introduce large attenuation into the signal. For example, losses for a device (Fig. 1a) operating in demultiplexing mode are 4-8 dB, with long-range crosstalk

Transponders and transceivers


To transmit data at wavelengths from a DWDM grid, two types of devices can be used - transceivers and DWDM transponders. DWDM transceivers come in a variety of form factors and can be used in passive DWDM solutions.

Unlike transceivers, transponders allow you to convert the radiation wavelength of the terminal device into a DWDM wavelength for transmission to the multiplexer. The inputs of the optical multiplexer receive optical signals whose parameters comply with the standards defined by the G.692 recommendations. A transponder may have a different number of optical inputs and outputs. But if an optical signal can be supplied to any transponder input, the parameters of which are determined by rec. G.957, then its output signals must correspond in parameters to rec. G.692. Moreover, if m optical signals are compressed, then at the transponder output the wavelength of each channel must correspond to only one of them in accordance with the ITU frequency plan grid.

Application of optical amplifiers

The development of EDFA-based optical amplification technology has greatly changed the design methodology of fiber optic communication systems. Traditional fiber optic systems use regenerator repeaters that increase signal power (Fig. 3a). When the length between remote nodes begins to exceed, in terms of signal attenuation, the maximum permissible flight length between neighboring nodes, additional regenerators are installed at intermediate points that accept weak signal, amplify it in the process of optoelectronic conversion, restore the duty cycle, fronts and time characteristics of the pulse repetition, and after converting it into optical form, they transmit the correct amplified signal, in the same form as it was at the output of the previous regenerator. Although such regeneration systems work well, they are quite expensive and, once installed, cannot increase the capacity of the line.

Based on EDFA, the power loss in the line is overcome by optical amplification (Fig. 3b). Unlike regenerators, this "transparent" gain is not tied to the bit rate of the signal, allowing information to be transmitted at higher rates and increasing throughput until other limiting factors such as chromatic dispersion and polarization mode dispersion come into play . EDFA amplifiers are also capable of amplifying a multi-channel WDM signal, adding another dimension to the bandwidth.

Although the optical signal generated by the original laser transmitter has a well-defined polarization, all other nodes along the optical signal path, including the optical receiver, should exhibit a weak dependence of their parameters on the direction of polarization. In this sense, EDFA optical amplifiers, characterized by a weak polarization dependence of the gain, have a tangible advantage over semiconductor amplifiers.

Unlike regenerators, optical amplifiers introduce additional noise that must be taken into account. Therefore, along with the gain, one of the important parameters of EDFA is the noise figure.

Application of ROADM devices


The use of a reconfigurable optical add/drop multiplexer (ROADM) enables flexible deployment and remote configuration of spectrum channels. At any node in the ROADM network, it is possible to switch the state of the spectrum channel to input/output and end-to-end transmission without interrupting existing services. When working with a tunable laser, ROADM provides flexible control of spectral channels. ROADMs allow you to build networks with multiple rings or mixed networks: based on spectrum selector switching (WSS) technology.

Construction of DWDM networks


Urban DWDM networks, as a rule, are built using a ring architecture, which allows the use of protection mechanisms at the DWDM level with a recovery speed of no more than 50 ms. It is possible to build a network infrastructure on equipment from several suppliers with an additional distribution level based on Metro DWDM equipment. This level is introduced to organize traffic exchange between networks with equipment from different companies.

In DWDM technology, the minimum signal resolution is the optical channel, or wavelength. The use of entire wavelengths with channel capacity of 2.5 or 10 Gbit/s to exchange traffic between subnets is justified for building large transport networks. But transponder-multiplexers allow you to organize traffic exchange between subnetworks at the level of STM-4/STM-1/GE signals. The distribution level can also be built on the basis of SDH technology. But DWDM has a great advantage associated with the transparency of control channels and service channels (for example, service communications). When SDH/ATM/IP signals are packaged into an optical channel, the structure and contents of the packets do not change. DWDM systems only monitor individual bytes to ensure that signals are flowing correctly. Therefore, connecting subnets over a DWDM infrastructure at a single wavelength can be considered as connecting with a pair of optical cables.

When using equipment from different manufacturers, two data transmission subnets from one manufacturer are connected through a DWDM network from another manufacturer. A control system physically connected to one subnet can also control the operation of another subnet. If SDH equipment were used at the distribution level, this would not be possible. Thus, based on DWDM networks, it is possible to combine networks from different manufacturers to transmit heterogeneous traffic.

SFP (WDM, CWDM, DWDM) – WHAT IS IT? WHAT ARE THEY NEEDED FOR?

Wavelength Division Multiplexing (WDM) technologies.

Spectrum multiplexing is based on a method of multiplexing optical channels. Principle this method lies in the fact that each information stream is transmitted over one optical fiber at a different wavelength (at a different carrier frequency), spaced at a distance of 20 nm from each other.

Using special devices - optical multiplexers - the streams are combined into one optical signal, which is introduced into the optical fiber. On the receiving side, the reverse operation is performed - demultiplexing, carried out using optical demultiplexers. This opens up truly inexhaustible possibilities for both increasing line capacity and building complex topological solutions using a single fiber.

When choosing the number of channels, you should pay attention to the type of single-mode fiber used!
For example, G.652B fibers (water-peaked fiber at 1383 nm) have high radiation losses at short wavelengths, so the permissible transmission distance is reduced and the number of spectral channels will be less than required.

In Coarse WDM systems, in accordance with ITU recommendation G.694.2, no more than 18 carriers should be used with a pitch of 20 nm: 1270, 1290, 1310 ... 1570, 1590, 1610, i.e. if the total required wavelength range does not exceed 340 nm. It should be taken into account that at the edges of such a wide range the attenuation is quite large, especially in the short wavelength region. The number of channels was increased to 18 using so-called zero water peak fibers (ZWPF, Zero Water Peak Fiber; LWPF, Low Water Peak Fiber), the parameters of which are determined by the ITU-T recommendation G.652.C/D. In fibers of this type The absorption peak at a wavelength of 1383 nm has been eliminated and the attenuation value at this wavelength is about 0.31 dB/km.

G.653 fiber proved unsuitable for the new, rapidly evolving WDM wavelength division multiplexing technology due to its zero dispersion at 1550 nm, which led to a sharp increase in signal distortion from four-wave mixing in these systems. The most suitable optical fiber for dense and high-density WDM (DWDM and HDWDM) was G.655, and the recently standardized G.656 optical fiber for sparse WDM.
The creation of fibers without a “water peak” made it possible to use all waves in the range from 1260 to 1625 nm in communication systems, i.e. where quartz optical fiber has the greatest transparency.

BASIC EQUIPMENT

Multiplexers/demultiplexers (MUX/DEMUX); allow you to sum and separate optical signals.

allow you to select and add a signal to the fiber at certain carrier frequencies.

Depending on the task at hand, the configuration of the multiplexer/demultiplexer (Mux/Demux) is determined by the following characteristics:

Dual fiber multiplexer (2 fiber)
Single fiber multiplexer(1 fiber (single fiber) or bidirectional)
4 or 8 channel multiplexer(8 or 16 wavelengths), operating on one fiber
8 or 16 channel, operating on two fibers
multiplexer with two "common"(COMMON) conclusions to implement a “ring” topology
For “Point-to-Point” or “Ring” topologies, a “pairwise” (Tx–Rx ports) set of multiplexers is required - Mux/Demux Type I, Mux/Demux Type II
Connectors – FC,SC,LC,ST,FA,SA

Multiplexers can be supplied in the following versions:
Rack 19" 1RU
In a plastic case(for wall or box mounting)
By type of connector– LC, SC, etc.


SFP (Small Form Factor Pluggable) transceivers (SFP, SFP+, X2, XFP) –
generate and receive optical signals (certain wavelengths) in a CWDM system; convert a signal from electrical to optical and vice versa. SFP module combines both a transmitter and a receiver. Therefore, it supports simultaneous transmission and reception of data over two links within a single channel. Since the days of radio, such devices have been called transceivers. This is why SFP modules are called transceivers.

Each SFP transceiver operates on two fibers and, unlike standard two-fiber 1000Base LX transceivers, operates on two different wavelengths - broadband receiver works with one wavelength and the transmitter with another.
To form a data channel in an SFP system, transceivers are configured in pairs.

Transceivers also differ in signal strength (mileage), i.e. they operate over different distances.

For stronger optical signal compression, “color” SFP modules are used that operate in a certain wavelength range (CWDM). Such SFP transceivers are designed to generate “main carrier” optical signals from 1270 to 1610 nm (20 nm step).

SFP modules are available that operate over both one and two fibers with a throughput of 1.25, 2.5 and 4.25 Gbps. These modules can be installed directly into switching equipment from virtually any manufacturer, allowing seamless integration of CWDM into existing infrastructure. The same module can serve as a Gigabit Ethernet, Fiber Channel or SDH interface, which significantly adds flexibility to the solution.

It is also possible to install CWDM SFP modules in media converter chassis. Using a chassis is the most flexible solution, completely eliminating problems of equipment incompatibility. Using the chassis, you get standard 1000BASE-T Gigabit Ethernet ports, eliminating the need for expensive switches with SFP ports.

Special attention should be paid to the compaction of 10 Gbit/s channels. Just three years ago, there were no transceivers operating at speeds of 10 Gbit/s and supporting wavelengths of the frequency grid of sparse spectrum multiplexing systems; now such modules have appeared, however, their use imposes significant restrictions on the capabilities of the system, compared to channel multiplexing 1 .25 Gbit/s and 2.5 Gbit/s.

There are currently no 10 Gbps lasers operating in the 1350-1450 nm wavelength range, so the maximum number of 10 Gbps multiplexed channels cannot exceed 12 when using two G.652D fibers. In addition, when using 10 Gbit/s channels, it is necessary to take into account that the maximum optical budget of such modules is currently no more than 28 dBm, which corresponds to an operating range of approximately 80 kilometers over single-mode fiber. In cases where it is necessary to compress and transmit more than 12 10 Gbit/s channels, incl. over distances of more than 80 kilometers, DWDM equipment is used.

OADM modules - input/output multiplexers; allow you to select and add a signal to the fiber for certain carriers.

Basic properties:
Single channel input/output
Passive optics
Low insertion loss for backhaul links
Dedicated wavelength to end user

Fundamentally, single-channel and dual-channel OADM modules are distinguished. Their difference lies in the ability to receive and receive an optical signal from one or two multiplexers and is physically due to the presence of one or two transceiver units. Accordingly, a single-channel OADM module has one transceiver unit and is capable of working with only one multiplexer in one direction. The two-channel OADM module has two transceiver units and is capable of working “in two directions” with two multiplexers/demultiplexers.

The transceiver unit of the single-channel OADM module has four interfaces:

Com port – receives a signal from the multiplexer
Express port – passes the signal to other elements of the system
Add port – adds a channel at a certain wavelength to the line,
Drop port – extracts a channel at a certain wavelength from the line.

Such devices have no restrictions on protocols or bandwidth.
Accordingly, the two-channel OADM module has two additional Add and Drop ports.
If a dual-fiber system is used, Com2 and Express2 ports are also added.
A single-channel OADM module works in tandem with one SFP transceiver, a two-channel OADM - with two

OADM terminal transit module ( drop/pass module) takes one channel from the trunk and routes it to the local port. The remaining channels are passed directly to other network nodes.

The single-channel OADM multiplexing module (drop/add module) has two local interfaces. The first one takes one channel from the trunk and directs it to the local port, the second one adds this channel back to the trunk in the opposite direction. Such a module is necessary when constructing a “ring” topology network.

OADM modules can be supplied in the following versions:
Rackmount 19” 1RU
In a plastic case (for mounting on a wall or in a sleeve)
Connectors – LC, SC, etc.

The main wavelength division multiplexing systems are:

- WDM (Wavelength Division Multiplexing)

- CWDM (Coarse Wavelength Division Multiplexing)

So what is WDM?

Technology for adding optical signals with different wavelengths, transmitted simultaneously along one fiber, 2 or more signals separated at the far end by wavelength. The most typical (2-channel WDM) combine wavelengths 1310 nm and 1550 nm in a single fiber.

Two-channel WDM (and three-channel) can be used to quickly and easily add an additional (or two additional) wavelengths. It is very easy to install and connect and very inexpensive. In most cases, WDM is the most cost-effective solution for cable shortages, providing a 2 to 1 or 3 to 1 fiber gain by combining the 1310 nm, 1550 nm and 1490 nm wavelengths into a single fiber.

In cases where more channels are required to expand the existing fiber optic infrastructure, CWDM provides an effective solution for short optical spans (up to 80 km). CWDM can easily and quickly add up to 18 additional wavelengths at ITU standardized frequencies. It is ideal for moderately sized networks with cross-sectional dimensions up to 100 km. Since the wavelength spacing is 20 nm, less expensive lasers can be used, resulting in very low cost. CWDM systems, although multichannel, do not have any optical amplification mechanisms and range limitations are determined by the channel with maximum attenuation. Moreover, channels from the 1360nm to 1440nm region may experience the greatest attenuation (1 to 2 dB/km) due to the water peak in this region for some types of optical cable.

Where high capacity or long distance transmission is required, solutions DWDM is the preferred method for increasing fiber capacity. With its high-precision lasers optimized to operate in the 1550 nm window (to reduce loss), DWDM systems are ideal solution for more demanding networks. DWDM systems can use EDFA to amplify all wavelengths in the DWDM window and extend transmission lengths up to 500 km.

DWDM systems are typically limited in range to 4-5 amplification sections due to Amplified Spontaneous Emissions (ASE) noise in the EDFA. Simulation tools are available to determine exactly how many EDFAs can be installed. On long sections (>120 km) dispersion can be a problem, requiring the installation of dispersion compensation modules. The DWDM band is limited to wavelengths ranging from 1530 nm to 1565 nm by the EDFA gain range.

Types of solutions:

1. Point - point.

Adding a point-to-point spectral system to an optical system is a simple and cost-effective solution to the fiber shortage problem.
Systems with a similar topology are typical for solving problems of simultaneous transmission of a large number of data streams to increase the number of services provided (video, voice, etc.). In this case, fibers from an already existing optical transport network are used. In this mode of operation, information is transmitted through channels between two points. To successfully transmit data over a distance of up to 50-80 km, multiplexers/demultiplexers are needed in those nodes where information flows will be combined and subsequently separated.

Branch connection

This architecture implements the transfer of information from one node to another with intermediate nodes along this path, where individual channels can be input and output using OADM modules. The maximum number of branches is determined by the number of duplex transmission channels (for example, 4 or and the optical budget of the line. When calculating, you need to remember that each OADM module introduces attenuation, as a result of which the total length of the path is correspondingly reduced. An optical channel can be extracted at any point in the path.

In this case, OADM modules (dual-channel) are installed between two multiplexers / demultiplexers.
In this case, each two-channel OADM module must be equipped with two SFP transceivers.

Point with branches.

The fundamental difference from the first option is the absence of a second multiplexer/demultiplexer. Thus, the exchange of signals occurs between the central communication center and the end equipment on different sections of the line. This architecture seems promising from an economic point of view, because in fact, it allows you to eliminate the aggregation layer switch from the network with significant savings in fiber. In this case, the distance from the OADM module (single-channel) to the location of the final equipment (switch, router, media converter) is limited only by the signal power in the line and the insertion losses from the multiplexing equipment.

Advantages
Saving optical fiber - the spectrum multiplexing system allows you to transmit up to 8 channels over one fiber with a throughput of up to 2.5 Gb/s per channel
Independence from power supply - power is required only for active equipment
No problems with crashes, reboots, etc.
There is no need to organize permanent access to the locations of system elements - there are OADM modules designed for placement in optical couplings
Reduced level of influence of the “human factor” - the absence of active components that require configuration, management, etc.
Significant reduction in cost of ownership - lower operating costs
Relatively low cost, possibility of eliminating aggregation level equipment
The maximum operating range is 80 kilometers or more
Independence from client protocols – transmission of up to 18 independent services over two pairs of optical fibers; transparency for all data transfer protocols
Availability various types equipment for installation in various conditions: in a rack, in a coupling, on a wall.

Surely everyone has heard about transmitting information over fiber optic networks, and also that this method provides the highest speeds to date. It is the latter that provides a good reason for the development of data transmission technologies over optical fiber. Already today, throughput can reach the order of terabits (1000 gigabits) per second.

When compared with other methods of information transmission, the order of magnitude TB/s is simply unattainable. Another advantage of such technologies is transmission reliability. Fiber optic transmission does not have the disadvantages of electrical or radio signal transmission. There is no interference that can damage the signal, and there is no need to license the use of the radio frequency. However, not many people imagine how information is transmitted over optical fiber in general, and even less are familiar with specific implementations of technologies. In this article we will look at one of them - DWDM (dense wavelength-division multiplexing) technology.

First, let's look at how information is transmitted over optical fiber in general. An optical fiber is a waveguide that carries electromagnetic waves with a wavelength of the order of a thousand nanometers (10-9 m). This is a region of infrared radiation that is not visible to the human eye. And the main idea is that with a certain selection of the fiber material and its diameter, a situation arises when for some wavelengths this medium becomes almost transparent and even when it hits the boundary between the fiber and the external environment, most of the energy is reflected back into the fiber. This ensures that radiation passes through the fiber without much loss, and the main task is to receive this radiation at the other end of the fiber. Of course, such a brief description hides the enormous and difficult work of many people. Do not think that such material is easy to create or that this effect is obvious. On the contrary, it should be treated as a great discovery, since it now provides a better way of transmitting information. You need to understand that the waveguide material is a unique development and the quality of data transmission and the level of interference depend on its properties; The waveguide insulation is designed to ensure that the outward energy output is minimal. Specifically speaking about a technology called “multiplexing,” this means that you transmit multiple wavelengths at the same time. They do not interact with each other, and when receiving or transmitting information, interference effects (superposition of one wave on another) are insignificant, since they manifest themselves most strongly at multiple wavelengths. Right here we're talking about about using close frequencies (frequency is inversely proportional to wavelength, so it doesn’t matter what you talk about). A device called a multiplexer is a device for encoding or decoding information into waveforms and back. After this short introduction, let's move on to a specific description of DWDM technology.

The main characteristics of DWDM multiplexers, which distinguish them from just WDM multiplexers:

  • use of only one transparency window of 1550 nm, within the EDFA amplification region of 1530-1560 nm (EDFA - optical amplification system);
  • short distances between multiplex channels - 3.2/1.6/0.8 or 0.4 nm.

For reference, let's say that the wavelength of visible light is 400-800 nm. In addition, since the name itself speaks of dense transmission of channels, the number of channels is greater than in conventional WDM schemes and reaches several dozen. Because of this, there is a need to create devices that are able to add a channel or remove it, as opposed to conventional schemes where all channels are encoded or decoded at once. The concept of passive wavelength routing is associated with such devices, which operate on one channel out of many. It is also clear that working with a large number of channels requires greater accuracy of signal encoding and decoding devices and places higher demands on line quality. Hence the obvious increase in the cost of devices - while simultaneously reducing the price for transmitting a unit of information due to the fact that it can now be transmitted in a larger volume.

This is how a demultiplexer with a mirror works (diagram in Fig. 1a). The incoming multiplex signal reaches the input port. This signal then passes through the waveguide plate and is distributed over many waveguides, which are an AWG (arrayed waveguide grating) diffraction structure. As before, the signal in each of the waveguides remains multiplexed, and each channel remains represented in all waveguides, that is, so far only parallelization has occurred. Next, the signals are reflected from the mirror surface, and as a result, the light fluxes are again collected in the waveguide-plate, where they are focused and interfered with. This leads to the formation of an interference pattern with spatially separated maxima, and usually the geometry of the plate and mirror is calculated so that these maxima coincide with the output poles. Multiplexing occurs in reverse.

Another method of constructing a multiplexer is based not on one, but on a pair of waveguide plates (Fig. 1b). The principle of operation of such a device is similar to the previous case, except that here an additional plate is used for focusing and interference.

DWDM multiplexers, being purely passive devices, introduce large attenuation into the signal. For example, losses for a device (see Fig. 1a) operating in demultiplexing mode are 10-12 dB, with long-range crosstalk interference less than –20 dB and a half-width of the signal spectrum of 1 nm (based on materials from Oki Electric Industry). Due to large losses, it is often necessary to install an optical amplifier before and/or after the DWDM multiplexer.

The most important parameter in dense wave multiplexing technology is undoubtedly the distance between adjacent channels. Standardization of the spatial arrangement of channels is needed if only because on its basis it will be possible to begin conducting tests for the mutual compatibility of equipment from different manufacturers. The telecommunications standardization sector of the International Telecommunication Union (ITU-T) has approved a DWDM frequency plan with an inter-channel spacing of 100 GHz, which corresponds to a wavelength difference of 0.8 nm. The issue of transmitting information with a difference in wavelengths of 0.4 nm is also being discussed. It would seem that the difference can be made even smaller, thereby achieving greater throughput, but in this case purely technological difficulties arise associated with the manufacture of lasers that generate a strictly monochromatic signal (constant frequency without interference) and diffraction gratings that separate the maxima in space , corresponding to different wavelengths. When using 100 GHz separation, all channels evenly fill the usable band, which is convenient when setting up equipment and reconfiguring it. The choice of separation interval is determined by the required bandwidth, the type of laser and the degree of interference on the line. However, it must be taken into account that when operating even in such a narrow range (1530-1560 nm), the influence of nonlinear interference at the boundaries of this region is very significant. This explains the fact that as the number of channels increases, it is necessary to increase the laser power, but this, in turn, leads to a decrease in the signal-to-noise ratio. As a result, the use of a stiffer seal is not yet standardized and is under development. Another obvious disadvantage of increasing density is the reduction in the distance over which the signal can be transmitted without amplification or regeneration (this will be discussed in more detail below).

Note that the nonlinearity problem mentioned above is inherent in silicon-based amplification systems. More reliable fluorine-zirconate systems are now being developed that provide greater linearity (in the entire region of 1530-1560 nm) of the gain. As the EDFA operating area increases, it becomes possible to multiplex 40 STM-64 channels at 100 GHz intervals with a total capacity of 400 GHz per fiber (Fig. 2).

The table shows specifications one of the powerful multiplex systems using the 100/50 GHz frequency plan, manufactured by Ciena Corp.

Let's take a closer look at the optical amplification system. What is the problem? Initially, the signal is generated by a laser and sent to the fiber. It spreads along the fiber, undergoing changes. The main change to deal with is signal scattering (dispersion). It is associated with nonlinear effects that arise when a wave packet passes through a medium and is obviously explained by the resistance of the medium. This raises the problem of long-distance transmission. Large - in the sense of hundreds or even thousands of kilometers. This is 12 orders of magnitude longer than the wavelength, so it is not surprising that even if the nonlinear effects are small, then in total at such a distance they must be taken into account. Plus, there may be nonlinearity in the laser itself. There are two ways to achieve reliable signal transmission. The first is the installation of regenerators that will receive a signal, decode it, generate a new signal, completely identical to the one that arrived, and send it further. This method is effective, but such devices are quite expensive, and increasing their capacity or adding new channels that they must handle involves difficulties in reconfiguring the system. The second method is simply optical amplification of the signal, completely similar to sound amplification in a music center. This amplification is based on EDFA technology. The signal is not decoded, but only its amplitude is increased. This allows you to get rid of speed losses in the amplification nodes, and also removes the problem of adding new channels, since the amplifier amplifies everything in a given range.

Based on EDFA, line power loss is overcome by optical amplification (Fig. 3). Unlike regenerators, this transparent gain is not tied to the bit rate of the signal, allowing information to be transmitted at higher rates and increasing throughput until other limiting factors such as chromatic dispersion and polarization mode dispersion come into play. EDFA amplifiers are also capable of amplifying a multi-channel WDM signal, adding another dimension to the bandwidth.

Although the optical signal generated by the original laser transmitter has a well-defined polarization, all other nodes along the path of the optical signal, including the optical receiver, should exhibit a weak dependence of their parameters on the direction of polarization. In this sense, EDFA optical amplifiers, characterized by a weak polarization dependence of the gain, have a noticeable advantage over semiconductor amplifiers. In Fig. Figure 3 shows the operation diagrams of both methods.

Unlike regenerators, optical amplifiers introduce additional noise that must be taken into account. Therefore, along with gain, one of the important parameters of EDFA is noise figure. EDFA technology is cheaper, for this reason it is more often used in real practice.

Since EDFA, at least in terms of price, looks more attractive, let's look at the main characteristics of this system. This is the saturation power characterizing output power amplifier (it can reach and even exceed 4 W); gain, defined as the ratio of the powers of the input and output signals; the power of amplified spontaneous emission determines noise level, which the amplifier itself creates. Here it is appropriate to give an example of a music center, where one can trace analogies in all these parameters. The third (noise level) is especially important, and it is desirable that it be as low as possible. Using an analogy, you could try to include music Center, without starting any disc, but at the same time turn the volume knob to maximum. In most cases you will hear some noise. This noise is created by amplification systems simply because they are powered. Similarly, in our case, spontaneous emission occurs, but since the amplifier is designed to emit waves in a certain range, photons of this particular range will be more likely to be emitted into the line. This will create (in our case) light noise. This imposes a limitation on the maximum length of the line and the number of optical amplifiers in it. The gain is usually selected so as to restore the original signal level. In Fig. Figure 4 shows the comparative spectra of the output signal in the presence and absence of a signal at the input.

Another parameter that is convenient to use when characterizing an amplifier is the noise factor - this is the ratio of the signal-to-noise parameters at the input and output of the amplifier. In an ideal amplifier, this parameter should be equal to unity.

There are three applications for EDFA amplifiers: preamplifiers, line amplifiers and power amplifiers. The first ones are installed directly in front of the receiver. This is done to increase the signal-to-noise ratio, which allows the use of simpler receivers and can reduce the price of the equipment. Linear amplifiers are intended to simply amplify the signal in long lines or in the case of branching of such lines. Power amplifiers are used to amplify the output signal directly after the laser. This is due to the fact that laser power is also limited and sometimes it is easier to simply install an optical amplifier than to install a more powerful laser. In Fig. Figure 5 schematically shows all three ways of using EDFA.

In addition to the direct optical amplification described above, an amplification device using the Raman amplification effect and developed at Bell Labs is currently preparing to enter the market. The essence of the effect is that a laser beam of a certain wavelength is sent from the receiving point towards the signal, which rocks the crystal lattice of the waveguide in such a way that it begins to emit photons in a wide range of frequencies. Thus, the overall level of the useful signal rises, which allows you to slightly increase the maximum distance. Today this distance is 160-180 km, compared to 70-80 km without Raman enhancement. These devices, manufactured by Lucent Technologies, will hit the market in early 2001.

What was described above is technology. Now a few words about implementations that already exist and are actively used in practice. First, we note that the use of fiber optic networks is not only the Internet and, perhaps, not so much the Internet. Fiber optic networks can carry voice and TV channels. Secondly, let's say that there are several different types networks. We are interested in long-distance backbone networks, as well as localized networks, for example within one city (the so-called metro solutions). At the same time, for trunk communication channels, where the rule “the thicker the pipe, the better” works perfectly, DWDM technology is the optimal and reasonable solution. A different situation arises in urban networks, in which the demands for traffic transmission are not as great as those of trunk channels. Here, operators use good old SDH/SONET based transport operating in the 1310 nm wavelength range. In this case, to solve the problem of insufficient bandwidth, which, by the way, is not yet very acute for urban networks, you can use the new SWDM technology, which is a kind of compromise between SDH/SONET and DWDM (read more about SWDM technology on our CD-ROM ). With this technology, the same fiber ring nodes support both single-channel data transmission at 1310 nm and wavelength division multiplexing at 1550 nm. Savings are achieved by “switching on” an additional wavelength, which requires adding a module to the corresponding device.

DWDM and traffic

One of important points When using DWDM technology, this is the transmitted traffic. The fact is that most equipment that currently exists supports the transmission of only one type of traffic on one wavelength. As a result, a situation often arises where traffic does not completely fill the fiber. Thus, less “dense” traffic is transmitted over a channel with a formal throughput equivalent to, for example, STM-16.

Currently, equipment is appearing that realizes full loading of wavelengths. In this case, one wavelength can be “filled” with heterogeneous traffic, say, TDM, ATM, IP. An example is the Chromatis family of equipment from Lucent Technologies, which can transmit all types of traffic supported by I/O interfaces on a single wavelength. This is achieved through the built-in TDM cross-switch and ATM switch. Moreover, the additional ATM switch is not price determining. In other words, additional functionality of the equipment is achieved at almost the same cost. This allows us to predict that the future lies in universal devices capable of transmitting any traffic with optimal use of bandwidth.

DWDM tomorrow

Smoothly moving on to the development trends of this technology, we certainly will not discover America if we say that DWDM is the most promising optical data transmission technology. This can be attributed to a greater extent to the rapid growth of Internet traffic, the growth rates of which are approaching thousands of percent. The main starting points in development will be an increase in the maximum transmission length without optical signal amplification and the implementation of a larger number of channels (wavelengths) in one fiber. Today's systems provide transmission of 40 wavelengths, corresponding to a 100-gigahertz frequency grid. Devices with a 50-GHz network supporting up to 80 channels are next in line to enter the market, which corresponds to the transmission of terabit streams over a single fiber. And today you can already hear statements from laboratories of development companies such as Lucent Technologies or Nortel Networks about the imminent creation of 25-GHz systems.

However, despite such rapid development of engineering and research, market indicators make their own adjustments. The past year has been marked by a serious decline in the optical market, as evidenced by the significant drop in Nortel Networks' share price (29% in one day of trading) after it announced difficulties in selling its products. Other manufacturers found themselves in a similar situation.

At the same time, while Western markets are experiencing some saturation, Eastern markets are just beginning to unfold. The most striking example is the Chinese market, where a dozen national-scale operators are racing to build backbone networks. And if “they” have practically resolved the issues of building backbone networks, then in our country, sad as it may be, there is simply no need for thick channels for transmitting our own traffic. Nevertheless, the exhibition “Departmental and corporate networks Communications" revealed the huge interest of domestic telecom operators in new technologies, including DWDM. And if such monsters as Transtelecom or Rostelecom already have state-scale transport networks, then the current energy sector is just beginning to build them. So, despite all the troubles, optics are the future. And DWDM will play a significant role here.

ComputerPress 1"2001




Top