Project topics and materials
Get Complete Project Material File(s) Now! »
Sub-wavelength switching solutions in the spectral domain
In WDM transport networks, a technology to increase the channel capacity carried by a wavelength is to increase the bandwidth it occupies. But this solution is limited by a lot of hardware constraints: the bandwidth of Digital to Analog Converters (DACs) and Analog to Digital Converters (ADCs), the speed of DSP, the nonlinear impairments, and so on. In order to improve the capacity of the WDM network, the super-channel solution has been proposed. A super-channel is a large optical spectrum structure that is larger than a single wavelength optical channel. In this structure, multiple optical bands are assembled together with minimal or even no guard-band, which allows saving spectral bandwidth. Moreover, this scheme releases the constraints, such as DAC, ADC, and DSP speed, at transmitters and receivers side, since it is easier to deal with a data flow with less bandwidth and smaller bit rate compared to the previous solution. The entire structure of a super-channel can cover a bandwidth of several hundreds GHz and transport huge bit rates. Inside the super-channel, the sub-band allocation is governed by the G.694.1 flex-grid ITU-T standard [5]. The super-channel technology now is always deployed in the network based on the Optical Circuit Switching (OCS) technology. In OCS, the network is configured to establish circuits, from one node toanother node, by adjusting the optical cross-connect circuits in the core routers. The super-channel is considered as a single spectral entity by optical switching devices at the network and is distributed to a node pair will be kept for a long duration. Recently, a new generation of elastic switching solutions based on high spectral resolution optical filters and Wavelength Selective Switches (WSSs) [6][7] has been developed. These solutions offer dynamic all-optical traffic aggregation/grooming at the sub-band level inside the super-channel. Thus, super-channel technology with the filtering and switching elements permits the sub-wavelength switching.
Modulation formats adapted to sub-wavelength switching
In order to compose the super-channel, it is necessary to generate sub-bands signal having a rectangular profile, which allows reducing the guard band between two adjacent sub-bands. At the same time, the high spectral efficiency feature is also a key point to be considered. With these two criteria, there are two promising technologies that have been widely studied in the last decade: Nyquist-WDM [9][10] and Multi-Band Orthogonal Frequency Division Multiplexing (MB-OFDM) [11][12] Both of these two candidates are compatible with direct detection and/or coherent detection technologies. The architectures of these two technologies are shown in Figure 1-3. The left architecture presents the direct detection technology. In direct detection, a photo-detector converts the optical signal in the electrical domain by producing a current proportional to the square of the received optical field amplitude. The right architecture presents the coherent detection technology. In coherent detection, it requires a Local Oscillator (LO), hybrid 90° and 4 balanced photo-detectors in order to detect the real part and imaginary part of the signal. When compared to direct detection, coherent detection is proportional to the optical field and allows recovering both amplitude and phase of the signal, opening the way to simple and efficient digital signal processing. Direct detection is cheaper (requiring only one photodiode), having a lower power consumption and lower cost-per-bit than coherent detection. But the direct detection offers a lower spectral efficiency [13][14]. With coherent detection, the linear impairments in transmission of the fiber, such as Chromatic Dispersion (CD) and Polarization Mode Dispersion (PMD), can be mitigated easily [15]. Coherent detection technology is thus more suitable for metro and long-haul transmission system applications than direct detection which is better suited for low spectral efficiency short-reach applications. Some experimental demonstrations using direct detection and coherent detection technologies are compared and shown in the table 1-1. As we can see that the direct detection technology is generally used in the short-reach transmission systems. Considering our specific specifications and requirements in this thesis, we chose coherent detection technology as the technology used in our test-bed.
Sub-wavelength switching solutions in the time domain
Besides the sub-wavelength switching in the spectral domain, the sub-wavelength switching can also be performed in the time domain. Indeed, time domain sub-wavelength switching solutions divide the wavelength into time entities, different end-users can access the same wavelength while using the wavelength for only a fraction of time (several time slots), which permit to share the same wavelength resource. Sub-Lambda Photonically Switched Networks (SLPSN) [60] has been normalized by the ITU-T in 2012 to describe the characteristics and common points of all time domain sub-wavelength switching solutions. In SLPSN, according to the length of the optical burst or optical packet, the sub-wavelength switching solution in the time domain is called Optical Packet Switching (OPS) [61] or Optical Burst Switching (OBS) [62]. Due to the lack of key optical components, such as big size fast optical memories and commercial fast optical switches, the realization of OPS in the real scale network faces significant challenges. Thus, OBS has been introduced as an alternative to OPS. Indeed, in OBS, optical bursts are an assembly of several optical packets that are sent to the same destination. Therefore, OBS manages time entities in the range of 1 μs to 10 ms which releases the constraints on the switching components. Consequently, we will focus on discussion on OBS solutions in this subsection.
Table of contents :
List of abbreviations
Introduction
Chapter 1: State of the art: Sub-wavelength switching solutions
1.1 Sub-wavelength switching solutions in the spectral domain
1.1.1 Super-channel concept
1.1.2 Flex grid standard.
1.1.3 Modulation formats adapted to sub-wavelength switching
1.1.4 ROADM Principles
1.1.5 High spectral resolution filters
1.2 Sub-wavelength switching solutions in the time domain
1.2.1 Principles and interests
1.2.2 POADM solution
1.2.3 OPST solution
1.2.4 TWIN solution
1.2.5 Comparison between the lossless solutions
1.3 Conclusion
References
Chapter 2: Time and Spectral Optical Aggregation solution
2.1 Data plane
2.2 Control plane
2.3 Flexibility of TISA solution
2.4 Main building blocks of TISA
2.4.1 The data plane
2.4.2 The control plane
2.5 Conclusion
References
Chapter 3: The burst mode CO-MB-OFDM transmitter/receiver
3.1 Fast tunable laser with ultra-narrow linewidth
3.1.1 Overview of tunable laser solutions
3.1.2 Fast tunable laser for TISA
3.2 OFDM transmitter set up
3.2.1 Experimental transmitter set up
3.2.2 Transmitter DSP procedures
3.2.3 Dimensioning
3.3 OFDM receiver set up
3.3.1 Experimental receiver set up
3.3.2 Receiver DSP procedures
3.3.3 Specific DSP for burst mode
3.4 Experimental back-to-back validation
3.4.1 Settings of the transmitter and the receiver
3.4.2 Experimental validation using DP-QPSK format (in burst mode)
3.4.3 Experimental validation using DP-16QAM format (in burst mode)
3.5 Conclusion
References
Chapter 4: Control plane for TISA solution
4.1 Control plane architecture
4.2 Grants calculation and distribution
4.2.1 Grants calculation and generation
4.2.2 Grants distribution
4.2.3 Grants storage and update
4.3 Burst generation according to Grants
4.3.1 SOAG control signal
4.3.2 External clock realization
4.3.3 Synchronization
4.4 Bursts alignment at the Multi-Band OFDM combiners
4.5 Conclusion
Reference
Chapter 5: TISA network realization and evaluation
5.1 Burst routing through the TISA network
5.1.1 MB-OFDM combiner
5.1.2 Core Node
5.1.3 MB-OFDM separator
5.1.4 Performance of bursts through the TISA network
5.2 Sensitivity to high resolution filter detuning
5.3 Flexibility evaluation
5.4 Novel training sequence strategy
5.4.1. Experimental setup and system characterization
5.4.2. Proposed solution and results
5.5 Conclusion
Reference
