2.1.

Characteristics of Elastic Optical Network

Example: As shown in Fig. 2, there are two paths, one from A to D and the other from B to C, each has a bandwidth of $300\mathrm{Gb}/\mathrm{s}$. Here the capacity of fiber is $400\mathrm{Gb}/\mathrm{s}$, so the path from B to C has $100\mathrm{Gb}/\mathrm{s}$ bandwidth available. In case of failure of the link between A and E, the optical path (A to D) is shifted to the recovery path nodes A, B, C, and D. Therefore, SBVT reduces the bandwidth of the path (A to D) from 300 to the $100\mathrm{Gb}/\mathrm{s}$ so that it can accommodate in the available bandwidth in the paths B to C.¹⁹

Fig. 2

Bandwidth squeezed restoration.

The main key aspect of EON is SBVT, which is a BVT of the next-generation network. In conventional BVT, whole transmitter capacity is assigned to a single connection request (source and destination pair). But SBVT is a flexible multiflow transponder, which is capable of routing the data in different destinations simultaneously without increment in the cost and complexity as shown in Fig. 3. Moreover, it also has the ability to tune dynamically according to the optical reach, optical bandwidth by proper adjustment in the parameters such as modulation format, forward error corrections (FEC) coding, gross bit rate, and optical spectrum shaping. These transponders are responsible for reducing the cost and complexity of EON.²⁰

Fig. 3

Multiflow transponder.

2.2.

Superchannel

Superchannel is a set of DWDM wavelengths that are generated from same optical line card and comes into operation in one operation cycle. It passes through the routing devices such as wavelength selective switches and optical add-drop multiplexers (OADMs) as a single entity providing high data rate and capacity. Coherent superchannel forms the foundation of next-generation intelligent optical transport network (OTN) by increasing the capacity of a channel from $100\mathrm{Gb}/\mathrm{s}$ to $1\mathrm{Tb}/\mathrm{s}$ without any compromise in the optical reach. An increment in the spectral efficiency due to coherent detection and higher order modulation helps in the implementation of coherent superchannel. Tera-bit scale superchannel provides the main difference between DWDM and EON.²¹

Problems addressed by the superchannel:

2.2.1.

Concept of superchannel

As mentioned earlier, superchannel brought into operation in a single operational cycle. The main concept of superchannel and the difference between superchannel and the normal channel is described below.²²

In DWDM channel, a guard band is present on the lower and upper edges of the channel. This guard band is necessary for optical switching, multiplexing, and demultiplexing of optical channels. But it consumes a huge amount of spectrum that is not useful for the actual payload transport and causing a reduction in the fiber transport capacity. Therefore, to provide the solution of this problem optical industry has moved toward a technology called superchannel, which is much wider than a conventional DWDM channel and it operates in a single operational cycle.²³

As shown in the example in Fig. 4(a), we consider one DWDM conventional channel of capacity $1.2\mathrm{Tb}/\mathrm{s}$, which is allocated $12*100\mathrm{Gb}/\mathrm{s}$ waves using ITU-T grid, and it is equivalent to a superchannel of $1.2\mathrm{Tb}/\mathrm{s}$. The DWDM channel requires 600 GHz of optical spectrum to carry $1.2\mathrm{Tb}/\mathrm{s}$, whereas the superchannel requires 462.5-GHz optical spectrum for transport as shown in Fig. 4(b). Here, a guard band is needed only at the lower and upper edges and not between each subchannel. Therefore, it is considered to be a single unit replacing 12 individual 50-GHz channels to a single 462.5-GHz channel, which is provisioned in a single operation cycle. But it requires a network that supports a flexible grid channel plan. Fortunately, a flexible grid channel plan is recently standardized by ITU-T under ITU-T G.694.1. Thus, the superchannel saves the optical spectrum by 23% as compared with the conventional DWDM channel.⁴

Fig. 4

(a) Conventional channel and (b) superchannel.

2.2.2.

Superchannel can only be possible by photonic integrated circuits technology

As mentioned earlier, superchannel brought into operation in a single photonic integrated circuits (PIC) technology is mandatory for the superchannel implementation and is shown in Fig. 5. Consider a superchannel consisting of 12 subcarriers (subchannel). In this case, the implementation of superchannel using discrete optical components is highly complex. To implement this, we need 12 sets of optical components per line card, which is unrealistic. The solution to this problem can be obtained from the PIC technology. In this case, we require two PICs, one for transmission and the other for reception. In PIC, all 12 carriers are fused into a single line card forming a superchannel that operates in one operational cycle reducing power and hardware complexity by 12 times in comparison with 12 discrete components.²² It is a compact, efficient, and reliable technique. Beyond $100\mathrm{Gb}/\mathrm{s}$ PIC technology is mandatory, without it the implementation of superchannel is not possible.²⁴

Fig. 5

(a) Superchannel with PIC technology and (b) superchannel with discreet component.

2.3.

Limitations of the Flexible Grid/Gridless Optical Networks

The challenges or limitations of the flexible grid/gridless optical networks are the cost and complexity of hardware and software network control systems. To provide the flexible (elastic) bandwidth at high speed, advance control planes are required, which support the flexible transceivers and network elements. On the other hand, flexible grid optical network increases the hardware complexity due to the deployment of flexible grid equipment, such as SBVT, superchannels, and flexible grid reconfigurable optical add-drop multiplexer (FlexROADM). In a flexible grid optical network, conventional channels have been replaced with superchannels. The managing of the flexible-grid superchannels is a big challenge as they occupy variable spectrum on the ITU-T grid based on the modulation format used (PM-BPSK, PM-QPSK, PM-16QAM, etc.). This leads to the evolution of FlexROADM, which can flexibly switch variable optical spectrum.⁴ However, the cost and complexity of FlexROADM equipment including spectrum selective switches, tunable laser, tunable filter, etc increase while reaching to much finer spectrum granularity due to the switching flexibility issues. Furthermore, flexible grid network affects the cost of the power amplifiers due to an increment in the number of channels in the fiber. More numbers of channels inject more power on the link, thereby exceeding the power handling capacity of the amplifier that degrades the link performance unless it is replaced by the high-power amplifiers. Therefore, the requirement of advance control planes and deployment of FlexROADMs, SBVT, superchannels, and high-power amplifiers make flexible grid optical network a quite expensive approach, despite being of providing large capacity.²⁵

3.1.

Modulation

Modulation is the process of imposing a digital signal onto a carrier signal in the analog domain, whereas it is imposed on the beam of light in the optical domain. The simplest and effective form of modulation that has been utilized in the past decades by DWDM network is intensity modulation direction detection (IMDD)—which is also called on/off keying (OOK). In IMDD modulation technique, the signal encodes a single bit (a 1 or 0) in each symbol and each symbol is representing one cycle of a clock. Its implementation is simple and cost-effective with less hardware complexity. But the spectral efficiency is poor along with the limitation for 10 over $2.5\mathrm{Gb}/\mathrm{s}$. It provides a wider transport spectrum for 10 over $2.5\mathrm{Gb}/\mathrm{s}$, but the period of the bit rate is small leading to chromatic dispersion. Therefore, to overcome the effect of chromatic dispersion, dispersion compensation is deployed. It is also found that during an increase in the channel capacity or channel data rate ($100\mathrm{Gb}/\mathrm{s}$), the effect of some more factors arises to affect the performance that includes polarization mode dispersion. Hence, it is necessary to consider all such factors while designing a DWDM or EON.²²

Now it is found that IMDD or OOK technique is not suitable for higher channel capacity such as beyond $10\mathrm{Gb}/\mathrm{s}$. For higher channel capacity, the optical industry prefers to go for higher order modulation techniques (M-ary modulation techniques) that can carry more bits in each symbol. In this M symbols are transmitted where each symbol is comprised with at least $N(N=\log 2M)$ bits. This technique also supports a higher number of bits per symbol supporting high data rate thereby reducing the symbol rate. The effect of dispersion is also reduced in this technique. Therefore, it is seen that higher order modulation techniques have good spectral efficiency with improved information capacity of the fiber.¹ Many higher modulation techniques are available, which supports higher data rate. Bits per symbol of different modulation techniques is shown in Table 2.

Table 2

Bits per symbol in different modulation techniques.

Further, symbol rate reduction and spectrally efficiency are improved using polarizing multiplexing (PM). The spectral efficiency and the bits-carrying capacity of the symbol are increased twice in this technique. Bits per symbol in polarized modulation multiplexing techniques is shown in Table 3.

Table 3

Bits per symbol in polarized multiplexed modulation techniques.

Optimum spectrum and resource utilization are achieved in higher modulation technique. It is also noted that an increase in the order of the modulation technique decreases the optical reach. So, we find that all modulation techniques are not suitable for all optical reach. Every modulation technique has its own spectral efficiency for the corresponding optical reach as shown in Table 4.²⁶

Table 4

Normalized reach versus total capacity of different modulation techniques.22

3.2.

Transmission Techniques

Currently, deployed conventional optical network is using wavelength-division multiplexing (WDM) with incoherent technologies. It can support a maximum data rate of $40\mathrm{Gb}/\mathrm{s}$. Hence to fulfill the increasing demand for data transmission capacity, it is required to have the better exploration of existing optical network in the C-band along with good transmission techniques that can provide high spectral efficiency.

SBVT supports transmission techniques that include Nyquist wavelength-division multiplexing (NWDM), orthogonal frequency division multiplexing (O-OFDM), and time-frequency packing (TFP). Each transmission technique is spectrally efficient for a particular scenario.²⁷

3.2.1.

Nyquist wavelength-division multiplexing

The main idea of Nyquist WDM transmission technique is to employ a digital pulse-shaping filter at the transmitter section that limits the bandwidth of the signal within the Nyquist frequency (i.e., half of the symbol rate) or in the other words, it limits the channel spacing equal to the baud rate. This technique improves the spectral efficiency by placing Nyquist filtered WDM channels closer to each other as shown in Fig. 6.²⁷ Here, a root raised cosine acts as a matched filter to limit the bandwidth. Due to lesser channel spacing, this technique has intersymbol interference (ISI) free transmission, minimum interchannel crosstalk along with good tolerance against the distortion effects such as dispersion and fiber nonlinearity.²⁸ NWDM-based superchannel provides negligible changes in the current single channel receiver by enabling signal demodulation DSP. It provides high spectral efficiency, and it is suitable for long-distance transmission. Nowadays, the NWDM technique provides an integrated solution and it also provides commercial products that are available on the market.²⁹

Fig. 6

Spectrum saving in OFDM and NWDM transmission technique.

4.1.

General Requirements of Sliceable Bandwidth Variable Transponder

4.2.

Architectures of Sliceable Bandwidth Variable Transponder

In this section, three architectures have been proposed in Refs. 3537.–38, as follows.

First architecture of SBVT is as follows.

The above-mentioned architecture shown in Fig. 8 consists of the electronic processing block, the PIC, and mux/coupler. Initially, data are passed through an electronic processing unit (which performs filtering, encoding, and pulse shaping) and then fed into PIC for modulating the signal. PIC is capable of generating two modulation techniques (PM-QPSK and PM-16QAM). The modulated subcarriers are multiplexed by mux/coupler to form a superchannel then directed into the specific media channel.

Fig. 8

SBVT architecture.

The architecture section shown in Fig. 9 describes the information rate of $400\mathrm{Gb}/\mathrm{s}$ or we can say modulation and transmission of $400\mathrm{Gb}/\mathrm{s}$ data.³⁷

Fig. 9

(a) SBVT architecture enables information rate of 400Gb/s and (b) internal functioning of the dotted block.

Here, we have chosen $400\mathrm{Gb}/\mathrm{s}$ PM-QPSK typically, and we realize it as four $100\mathrm{Gb}/\mathrm{s}$ interfaces are considered instead of one $400\mathrm{Gb}/\mathrm{s}$ interface. One $400\mathrm{Gb}/\mathrm{s}$ interface requires a baud rate of around 100 Gbaud, which is practically impossible in the current scenario. Hence, we have used four $100\mathrm{Gb}/\mathrm{s}$ interfaces with a baud rate of 25 Gbaud, which are handled by four PICs. Each PIC will modulate $100\mathrm{Gb}/\mathrm{s}$. In Fig. 9(a), traffic from clients initially enters into the OTN, which is an interface between client and SBVT or it can also be defined as an electrical layer between IP layer and the optical layer. Later it enters into electronic processing module, where processing is done such as encoding, pulse shaping, filtering, and lastly it is directed to proper PIC for modulation.

Figure 9(b) shows the internal functioning of a dotted block of Fig. 9(a). Here, we consider two cases for the modulation of $200\mathrm{Gb}/\mathrm{s}$ using PIC: initially by PM-16QAM, which is followed by PM-QPSK.

4.3.

Photonic Integrated Circuit

PIC is essential for the superchannel generation. In this architecture, PIC is modulated the subcarriers by two techniques PM-16QAM and PM-QPSK as shown in Fig. 10.

Fig. 10

Photonic integrated circuit.

In this, PM-16QAM signal is generated if clients are applied as an input to all the eight ports, the first four inputs generate a 16-QAM signal then polarizer beam combiner generates a polarization multiplexed signal (PM-16QAM). Next PM-QPSK is generated by applying input to ports 1, 3, 5, and 7 keeping the ports 2, 4, 6, and 8 in a switched OFF condition.³⁷ But in the third SBVT architecture, PM-BPSK is also generated by applying clients as input to ports 1 and 5.³⁸

The second architecture of SBVT is as follows:

The above Fig. 11 shows outlook of the second SBVT architecture. The functioning of SBVT architecture can be explained by three sections. First is OTN interface, next is a multiflow optical module and lastly optical cross-connects (OXC) as shown in Fig. 11.

Fig. 11

SBVT architecture.

OTN is an electrical layer interface between the IP layer and client’s layer. It is an efficient and cost-effective technique for the operators to form a systematic and well-organized optical network. The International Telecommunication Union-Telecommunication Standardization Sector (ITU-T) organization develops and designs the standards and recommendations for this. These recommendations provide the information to meet the future network needs supporting higher data clients. These standards and recommendations change according to the needs and trends in the industry. It performs various functions such as monitoring, client mapping, grooming, and multiplexing making a network able to support all types of present and future client protocols.³⁹ Another important characteristic of OTN is FEC, which is added in the transmitter and decoded in the receiver. A decoder that is employed at the receiver performs many functions, such as decoding, detecting, and correcting the corrupted. It allows the transmission of the data up to thousands of kilometers with high performance.^40,41 Multiplexing capability of OTN helps to supports all types of traffic, such as Ethernet, SONET/SDH. Apart from this, it improves network efficiency by managing the network functions and performances. It mainly contains two units, an optical channel data unit (ODU) followed by an optical transport unit (OTU). ITU-T recommends the OTU4 and ODU4 for $100\mathrm{Gb}/\mathrm{s}$ and for the data formats beyond $100\mathrm{Gb}/\mathrm{s}$ ITU-T G.709 version 3 recommends ODUCn/OTUCn, where C is $100\mathrm{Gb}/\mathrm{s}$ granularity and n is an integer number handled by optical network.^42,43

The OTN layer also divides higher data streams into the lower data stream, in case it is unable to accommodate the entire data stream into a superchannel due to unavailability of the contiguous spectrum as shown in Fig. 12. Here, $1\mathrm{Tb}/\mathrm{s}$ stream is split into two streams of 600 and $400\mathrm{Gb}/\mathrm{s}$, respectively. Next, the streams enter into the multiflow optical module.³⁵

Fig. 12

OTN frames and flexible association with media channel.

The main parts of a multiflow optical module are flow distributor, subcarrier generator, and flex subcarrier module. In multiflow module, modulation, transmission, and reception are performed. Flow distributor helps in directing the data stream from OTN to the flex subcarrier module as shown in Fig. 13. The flex subcarrier module contains two sections, such as transmitting section and receiving section. ³⁵

Fig. 13

Flex subcarrier module.

In this section, the input data are first encoded in the encoder followed by the pulse shaping using the pulse shaping circuit and finally fed into the predistortion module. Later, it is fed into the PM I/Q polarized multiplexed modulator for modulation. The modulation format is dependent on bit rate and the optical reach. Now, all the modulated subcarriers are combined together to form a superchannel, which is guided into a proper media channel for transmission. Later media channels are coupled into the OADM/OXC through a mux/coupler.

Receiving section consists of two units, namely coherent detection and advanced DSP. In coherent detection unit, a polarized beam splitter divides the signal into two orthogonal polarized components. They are combined with the local oscillator (LS) output in a 90-degree optical hybrid mixer providing in-phase and quadrature-phase components. They are then fed into balanced photodiodes for detection. The detected components are amplified by radio frequency amplifier (RFA), followed by a high-speed coherent ADC. Lastly, data enter into an advanced DSP unit consisting of modules such as filtering, clock, and carrier frequency recovering, equalization, carrier phase recovery, and lastly decision, demapping and decoding module, where finally we get the decoded data.³⁵

Here, we describe the third SBVT architecture consisting of transmitter and receiver section shown in Fig. 14. Transmitter section consists of various units, such as distance module (DM), which is followed by modulation and transmission module (M&TM). The receiver section consists of demodulator supporting (PM-16QAM/PM-QPSK/PM-BPSK) and DSP unit.³⁸

Fig. 14

SBVT transmission section.

Distance module is a programmable module, which dynamically allocates the modulation technique to a data-stream/main-stream according to the optical reach. Data-stream confines the direction and route, where one or more data-stream can have the same direction and route with different dropping points along the route. These data-streams are combined together into a main-stream. Main stream/data stream is carried by a superchannel. While the data stream exceeds the capacity of a superchannel, it gets split forming a new main-stream, which is controlled by a new superchannel. A superchannel is made of multiple subchannels. A data stream/main stream consists of one and more substreams, which depend on the dropping points along the path. Distance module assigns any one modulation technique to its substreams depending on dropping distance. Every distance module consists of a demux, which demultiplexes the main stream/data-stream into individual substreams according to the dropping points in the path and amount of data dropped at each dropping point. A modulation scheme is assigned to the demultiplexed output depending on the optical reach of each substream as shown in Fig. 15. Here, three main modulation techniques are used, namely PM-QPSK, PM-16QAM, and PM-BPSK. Each technique is best suited for the given capacity and reach. For short haul distance, PM-16QAM is preferred and PM-QPSK is preferred for the long haul. For extra-long haul and submarines, PM-BPSK is preferred for efficient utilization of spectrum. Each data stream/main stream is carried by a superchannel, whereas substreams are carried by one or more subchannels (subcarriers).³⁸

Fig. 15

Distance module.

In each of the distance modules, modulation techniques are assigned to data streams. However, in this section, both modulation and transmission are performed. In modulation and transmission section, data-stream/main-stream initially enters into the data processing unit where encoding, pulse shaping, and filtering of data stream/main-stream are performed. Later, it is passed into the demux and switching matrix, where the data stream/main-stream gets split into multiple substreams. They are then directed into suitable planar light wave circuit (PLC) for modulation of the subcarrier according to the distance module. Lastly, modulated subcarriers are coupled by multimode interference (MMI) coupler for the superchannel generation as shown in Fig. 16.

Fig. 16

Modulation and transmission module.

This section consists of two subsections, namely a decision-making block, which is followed by demodulator block.

Fig. 17

Receiving section.

Table 6 shows the comparison of the first, second, and third architectures of SBVT by considering some of their characteristics such as programmable distance module, the concept of subchannel, modulation format, add and drop network, the flexibility of receiving section design, and capability of PLC.

Table 6

Comparison between the first,37 the second,35 and third architecture of SBVT.38

The sliceable property of SBVT makes it to effectively manage the bandwidths that are dynamically changeable. It serves different destinations with different data rates on demand as shown in Fig. 17. As shown in Fig. 18(a), $400\mathrm{Gb}/\mathrm{s}$ can send the data to four different destinations with each of $100\text{-}\mathrm{Gb}/\mathrm{s}$ bitrate and $37.5\text{-}\mathrm{GHz}$ spacing. Similarly, it can send the data to two different destinations with $200\text{-}\mathrm{Gb}/\mathrm{s}$ bitrate with 50-GHz spacing each as shown in Fig. 18(b). Finally, it is also seen that it can send data with $400\mathrm{Gb}/\mathrm{s}$ to a single destination with 75-GHz spacing and is shown in Fig. 18(c). This shows the flexible grid spacing and sliceablity property of SBVT, which helps in dynamically changeable multiple direction optical flow.³⁵

Fig. 18

(a) Data flow directed in four different destinations, (b) data flow directed in two different destinations, and (c) data flow directed in a single destination.

SBVT provides a solution to the optical industry for getting migrated toward higher data rates.³⁵

SBVT helps in the restoration of link or fiber failure without the need for additional transponder due to its sliceablity property.

SBVT helps in the restoration of link or fiber failure without any additional transponder due to its sliceable property. For example, consider a super channel of 400 Gbs with a bandwidth of 75 GHz and a link is failed in a path as shown in Fig. 19. The alternate paths are available with 50 Hz, which is less than the bandwidth of the failed link. Now in this situation, the SBVT divides the super channel into two media channels of 50-GHz bandwidth to restore the failure situation. Now, it is seen that under this failure condition, the data are streamlined in two different channels with lower bandwidth thereby delivering the data without loss.³⁵

Fig. 19

Restoration of failure.

5.1.

Array of Lasers

In this method, an array of lasers is used for the generation of multiple subcarriers.

5.2.

Optical Frequency Comb

OFC is a multicarrier optical source that simplifies the structure of DWDM and EON involving many individual lasers. In this subcarrier generation technique, a single laser source is capable of generating multiple subcarriers. It is also called a multiwavelength source. OFC refers to an optical spectrum with evenly spaced optical frequency components with substantial variation in the intensity of comb lines. Usually, this kind of optical spectrum has a regular train of optical pulses associated with it. It has a fixed rate of pulse repetition, which is used for determining the inverse line spacing in the spectrum.⁴⁵

OFC can be represented in the frequency domain by $S_n$:

OFC provides a large number of uniform optical lines from a single device, which is required for broadband and high-speed detection. In recent days, OFCs are most preferred for measuring frequency due to its higher accuracy, spectral purity, and broad spectral coverage.⁴⁶ OFCs, which were initially developed to establish a connection between the optical and radio frequency domain, are now being used in many fields of research that includes optical code division multiple access, arbitrary waveform generation, WDM, elastically optical network attosecond science, remote sensing, microwave synthesis, optical communications, laser cooling, optical frequency metrology, astronomical spectrograph calibration, synthesis of ultrapurity optical and RF frequencies, time-frequency transfer based on optical clocks, exoplanet searches, medical diagnostics, molecular fingerprinting, and astrophysics. Now OFC systems are well established in the visible and near-infrared (IR) spectral regions. It is further extending into the mid-IR, along with the terahertz, and extreme-ultraviolet region.⁴⁷

The cost, simplicity, and capability of reducing the effects of Kerr-nonlinearity present in the optical fibers are important factors that determine the appropriateness of this technology, especially in SBVT.⁴⁸