Chapter 3.9 - Metropolitan Area Packet Switched Network

3.9 METROPOLITAN AREA PACKET SWITCHED NETWORKS

In an end-to-end connectivity picture, the current network consists of three major
segments: the access network, which carries the data to and from individual users,
the metropolitan area network (MAN), and the backbone wide area network
(WAN). Metropolitan area networks, or metro networks, for short, interconnect
the backbone networks with the local access networks. Most existing metro
networks are based on a synchronous optical network with synchronous digital
hierarchy (SONET/SDH) technology, a circuit-switched networking technology
designed for symmetric traffic. These types of networks cannot carry bursty asymmetric
IP traffic efficiently, resulting in a bandwidth bottleneck at the metro level,
often referred to as the “metro gap.” This performance mismatch does not allow
high-speed clients and service providers in local access networks to tap into the
vast amounts of bandwidth available in the backbone networks [2, 63–70].

The MAN needs to provide a large variety of service qualities, such as
best-effort-based connectionless datagram delivery and/or quality-of-service-based
connection-oriented virtual circuits. It should also offer scalability to accommodate
the rapidly growing access networks. Metro networks are characterized by high
dynamism of traffic patterns, relatively high aggregate bandwidths, and relatively
short covered distances. Technical solutions for metro architectures are far from
being consolidated, and range from classical circuit-switched SONET/SDH rings,
to extensions of traditional high-speed LANs, such as the Resilient Packet Ring
IEEE 802.17, to switched (multi) Gigabit Ethernets, to Broadband Passive Optical
Networks ITU-T G.983, to more innovative optical packet switching proposals
[71]. The latter are considered by many researchers the only approach capable of
withstanding, in the long term, the continuous growth of aggregate capacities.

Metropolitan area networks are one of the best arenas for an early penetration
of advanced optical technologies. According to several studies, the provision of
low-cost broadband access in metropolitan areas has the potential for fast returns
on investments, and can foster the development of new bandwidth-hungry applications.
Indeed, their large traffic dynamism requires packet switching to efficiently
use the available resources, their high-capacity requirements justify WDM use,
and their limited geographical distances lower the impact of fiber transmission
impairments. Besides the robustness and better resource utilization that are typical
of packet switching, all-optical packet switching eliminates queuing delays at
intermediate nodes and provides bit rate and protocol transparencies [2].

The future MAN will be a very dynamic network, consisting of a large variety
of users with different service requirements. Flexibility and efficiency are two key
requirements for metro networks. Flexibility is required as metro networks
have to support a wide range of heterogeneous protocols, such as ATM, Frame
Relay, SONET/SDH, and IP. Efficiency is required because metro networks are
highly cost-sensitive [2]. The MAN should also offer scalability to accommodate
the rapidly growing access networks. Another issue MANs have to cope with is
the time dependence of traffic patterns due to their proximity to the end user.

The optimization parameters required for metropolitan area networks are [2, 70]:

Flexible upgrade. The transition to the new solutions that will enhance the
performance of a MAN and add to its scalability must be implemented
smoothly and with a controlled investment cost.
Optimized resource utilization. In order to provide bandwidth-flexible services
at reasonable costs, the next-generation metro solutions will need to propose
resource allocation management schemes that will take into account the
increased burstiness of the traffic. Data traffic burstiness is obviously higher
than in the backbone as a result of less efficient statistical multiplexing in a
network much closer to the access and usually with simpler topologies.
Cost per transferred bit. The associated cost is an important factor in metropolitan
area networks that has to be traded off with increases in capacity and
flexibility. Higher utilization of available resources will drive the cost down.
Transparency. In order to protect past networks investments, the new solutions
must operate transparently in relation to the existing variety of protocols.
The introduction of optical technology can assist in this.

A general architecture of a metropolitan area packet switched network with a ring
topology is shown in Figure 3.14. This architecture is suitable for short-term
implementation as it avoids the use of technologically immature optical

FIGURE 3.14 A generic architecture of a metro network with a ring topology.

components. Such a network eliminates the need for buffering and regeneration
within the optical add–drop nodes. Buffering is performed electronically in the
access nodes and packet transmission is guaranteed to be collision-free thanks to
a medium access control (MAC) protocol. The application of a MAC protocol
and the lack of buffering capabilities inside the optical core obviously put some
limit on the topology complexity and dimension of the network, but a variety of
more or less sophisticated protocols are compatible with a metropolitan range.
Regeneration is avoided internally as much as possible due to the limited transmission
distances in metro networks [2].

Ring topologies are often preferred for metropolitan area networks because they
allow the optical packet add–drop node to remain simple (an optical channel only
gets through a 2 × 2 switching function), resulting in decreased cost and a limited
physical impact on the signal quality. On the other hand, ring topologies are
not optimal when traffic experiences a complex connectivity pattern, because
transmissions typically have to traverse multiple network nodes to reach their
destinations [2].

The design and implementation of the MAC protocol has a significant impact
on the network performance and complexity. Medium access control protocols
can be classified into protocols that perform channel inspection, those that omit
this step, and those that use a separate control channel for signaling purposes.
Channel inspection is the process of determining the status of a slot (empty or occupied)
and is part of collision avoidance or detection schemes. Two different channel
inspection strategies can be distinguished: If the packet that will transmit in the
upcoming slot is selected prior to the inspection of the channel the strategy is
referred to as “a priori,” in contrast to the case where the packet selection is performed
after the channel inspection has been completed, referred to as “a posteriori”
strategy. A priori access relaxes the timing constraints for packet selection, but may
result in a packet transmission failure if the corresponding wavelength turns out
to be occupied. In this case, the bandwidth for the specific slot will be wasted as
it is not possible to select a different packet that can transmit on a free wavelength.
The reverse holds true for a posteriori access. The timing constraints under which
packet selection must be performed are tight, but a collision-free transmission is
guaranteed.

Three research projects focused on packet switched metro networks are presented
in the following sections, namely HORNET, DAVID, and RingO.

< Previous Excerpt Next Excerpt >

Purchase This Book

TABLE OF CONTENTS

Chapter 3.9 - Metropolitan Area Packet Switched Network

Contact Preferences

This is embarrasing...

Customize Your GlobalSpec Experience

Select Your Free Newsletters

Industry Newsletters

Select Your Free Product Alerts

This is embarrasing...