**3.1 Core and metro networks**

The two main categories of network to be considered from the point of view of establishing super-broadband access networks are core and metro networks. In this subsection, the two main challenges facing core and metro networks are discussed. These challenges are realising the bandwidth potential of fiber optic core networks by appropriate wavelength allocation and switching strategies. Therefore in this subsection, the discussion focussed on optical switching paradigms and dynamic wavelength allocation.

The main barrier to the use of most existing core and metro networks for future traffic transportation arises from their active electrical switching and routing systems which delay packets when processing them for switching in the electrical domain. It takes time to convert a signal from the optical to the electrical domain and vice versa. In addition the synchronization and data retiming processing takes time. Indeed, a great part of the research into optical networks is dedicated to transparency in optical networks in order to bypass Optical/Electrical/Optical (O/E/O) conversions in the intermediate nodes of the network. Thus, a number of network protocols such as MPLS (Multi Protocol Label Switching, GMPLS, etc. (Larkin, 2005) together with switching strategies (circuit- burst- or packet-switching) are proposed for data transparency in the network. Among the switching strategies, burst switching is the most compatible with the current optoelectronic technologies in terms of data transparency and switching speed. Packet switching is more efficient for data communication, but due to the limited speed of electrical networks compared to the current optical networks and the insufficient evolution of all-optical signal processing alternatives, packet based optical networks are not a practical solution for transparent optical networks. A comparison of the all optical switching schemes, optical circuit switching (OCS), optical burst switching (OBS) and optical packet switching (OPS) is shown in Fig. 6.

Fig. 6. Comparison of all-optical switching technologies in terms of relative magnitudes of performance measures.

Optical packet switching (OPS) is a viable candidate switching scheme for future networks because it is a purely-connectionless networking solution that is fully compatible with IPcentric data traffic and offers the finest network granularity, the best bandwidth utilization, flexibility, high-speed, and the ability to use the resources available economically.

OPS places more demanding prerequisites on the network than OBS because it processes packets on the fly. The most feasible approach to implementing of OPS involves processing synchronously transmitted packets with fixed lengths. However, in this case the hardware overhead is on the implementation of the packet synchronizer at the input to the switch. Despite their feasibility limitations, OPS demonstrators assisted the development of numerous ultra-fast switching and processing techniques regarding wavelength conversion, header encoding/decoding and processing, label swapping, fast clock extraction, and regeneration.

research into optical networks is dedicated to transparency in optical networks in order to bypass Optical/Electrical/Optical (O/E/O) conversions in the intermediate nodes of the network. Thus, a number of network protocols such as MPLS (Multi Protocol Label Switching, GMPLS, etc. (Larkin, 2005) together with switching strategies (circuit- burst- or packet-switching) are proposed for data transparency in the network. Among the switching strategies, burst switching is the most compatible with the current optoelectronic technologies in terms of data transparency and switching speed. Packet switching is more efficient for data communication, but due to the limited speed of electrical networks compared to the current optical networks and the insufficient evolution of all-optical signal processing alternatives, packet based optical networks are not a practical solution for transparent optical networks. A comparison of the all optical switching schemes, optical circuit switching (OCS), optical burst switching (OBS) and optical packet switching (OPS) is

Fig. 6. Comparison of all-optical switching technologies in terms of relative magnitudes of

Optical packet switching (OPS) is a viable candidate switching scheme for future networks because it is a purely-connectionless networking solution that is fully compatible with IPcentric data traffic and offers the finest network granularity, the best bandwidth utilization,

OPS places more demanding prerequisites on the network than OBS because it processes packets on the fly. The most feasible approach to implementing of OPS involves processing synchronously transmitted packets with fixed lengths. However, in this case the hardware overhead is on the implementation of the packet synchronizer at the input to the switch. Despite their feasibility limitations, OPS demonstrators assisted the development of numerous ultra-fast switching and processing techniques regarding wavelength conversion, header encoding/decoding and processing, label swapping, fast clock extraction, and

flexibility, high-speed, and the ability to use the resources available economically.

shown in Fig. 6.

performance measures.

regeneration.

The main challenges in OPS are the implementation of the optical header processing mechanism, the development of an intelligent switch controller, the realization of ultra fast switching at a nanosecond timescale, and the exploitation on buffering mechanisms to reduce packet blocking (Rodrigo et al, 2009; Raffaelli et al. 2008; Le Rouzic et al., 2005).

Furthermore, the channel allocation and spectral efficiency are other key points for superbroadband network deployment. There are different schemes for channel allocation and multiplexing techniques such as wavelength division multiplexing (WDM), Dense-WDM(DWDM), Highly DWDM, Orthogonal WDM (Goldfarb et al., 2007; Llorente et al., 2005) that are suitable for super-broadband network deployment. The WDM multiplexing based schemes are in addition to multiplexing schemes including time, frequency, and code division multiplexing techniques, which are used in current wired and wireless communication networks and perform well on them. Moreover, cognitive channel and spectrum allocation improves the network's throughput and reduces the cost-over head significantly.
