17-03-2014, 06:09 PM
INTRODUCTION
In WDM optical networks, each wavelength channel
provides the transmission rate of several gigabits per second
to satisfy the explosion requirements of related services, such
that the failures may lead to a lot of traffic blocked.
Therefore, survivability has become a key problem in WDM
optical networks. The aim of survivability is to ensure the
networks with services continuity by preplanning the backup
resources for working resources. In previous work [1], the
authors studied the shared-path protection (SPP) method in
which each connection request is equipped with a working
path and a failure-disjoint backup path. Once the working
path is unavailable by failures, the working traffic can be
switched to the backup path such that the service can be
protected. Since the single-link failure is dominant in current
optical networks, this paper will focus on the protection
design for this failure scenario [1].
In SPP, although the resources utilization ratio may be
good, the restoration time may be long and the protection
switching procedure may be complicated since the failures
may lead to many messages and signals to notify the repaired
source node. Therefore, the effective management in SPP is
a challenge. To overcome the drawback, previous papers
proposed the protection method based on p-cycles [2, 3],
which can achieve the efficient resources utilization as the
shared mesh protection method meanwhile perform the fast
restoration time as the ring networks. Generally, in order to
provide effective protection with p-cycles, the network may
be divided into many local areas that may lead to
complicated cooperation of different cycles. As a special
case of p-cycles, Hamiltonian Cycle Protection (HCP)
method has been proposed which can achieve fast restoration
and simple management [4-7]. In order to achieve HCP,
there must be at least a Hamiltonian cycle in the network.
Fortunately, we can find the Hamiltonian cycles in most
current optical backbone networks, e.g., US National, China
CERNET, NJLATA, ECNET networks, etc. In the following
Fig. 1, we explain the basic idea of HCP.
A B
D
C
Hamiltonian Cycle
A B
D
C
(a)
(b)
Figure 1. Illustration of Hamiltonian cycle protection
In Fig. 1, there exists a Hamiltonian cycle A-B-C-D,
where the fiber links are divided into two categories: oncycle
links and straddling links. Obviously, these links on
Hamiltonian cycle are on-cycle links (thick links in Fig. 1),
and other links are straddling links (thin links in Fig. 1). In
Fig. 1(a), any single failure of on-cycle link can be protected
by the residual available routes on Hamiltonian cycle; for
example, the failure of on-cycle link A-B can be protected by
the route B-C-D-A. Therefore, the backup wavelengths on
Hamiltonian cycle should be enough to protect the working
wavelengths on the failed link; for example, if there are four
working wavelengths on link A-B, there will need four
backup wavelengths on each link on Hamiltonian cycle.
Another case is in Fig. 1(b), where any single failure of
straddling link can be protected by two available routes on
Hamiltonian cycle; for example, the failure of straddling link
A-C can be protected by two routes A-B-C and A-D-C.
Therefore, the backup wavelengths on Hamiltonian cycle
should be enough to protect the half of working wavelengths
on the failed link; for example, if there are six working
wavelengths on link A-C, there will need three backup
wavelengths on each link on Hamiltonian cycle.
In WDM optical networks, each wavelength channel
provides the transmission rate of several gigabits per second
to satisfy the explosion requirements of related services, such
that the failures may lead to a lot of traffic blocked.
Therefore, survivability has become a key problem in WDM
optical networks. The aim of survivability is to ensure the
networks with services continuity by preplanning the backup
resources for working resources. In previous work [1], the
authors studied the shared-path protection (SPP) method in
which each connection request is equipped with a working
path and a failure-disjoint backup path. Once the working
path is unavailable by failures, the working traffic can be
switched to the backup path such that the service can be
protected. Since the single-link failure is dominant in current
optical networks, this paper will focus on the protection
design for this failure scenario [1].
In SPP, although the resources utilization ratio may be
good, the restoration time may be long and the protection
switching procedure may be complicated since the failures
may lead to many messages and signals to notify the repaired
source node. Therefore, the effective management in SPP is
a challenge. To overcome the drawback, previous papers
proposed the protection method based on p-cycles [2, 3],
which can achieve the efficient resources utilization as the
shared mesh protection method meanwhile perform the fast
restoration time as the ring networks. Generally, in order to
provide effective protection with p-cycles, the network may
be divided into many local areas that may lead to
complicated cooperation of different cycles. As a special
case of p-cycles, Hamiltonian Cycle Protection (HCP)
method has been proposed which can achieve fast restoration
and simple management [4-7]. In order to achieve HCP,
there must be at least a Hamiltonian cycle in the network.
Fortunately, we can find the Hamiltonian cycles in most
current optical backbone networks, e.g., US National, China
CERNET, NJLATA, ECNET networks, etc. In the following
Fig. 1, we explain the basic idea of HCP.
A B
D
C
Hamiltonian Cycle
A B
D
C
(a)
(b)
Figure 1. Illustration of Hamiltonian cycle protection
In Fig. 1, there exists a Hamiltonian cycle A-B-C-D,
where the fiber links are divided into two categories: oncycle
links and straddling links. Obviously, these links on
Hamiltonian cycle are on-cycle links (thick links in Fig. 1),
and other links are straddling links (thin links in Fig. 1). In
Fig. 1(a), any single failure of on-cycle link can be protected
by the residual available routes on Hamiltonian cycle; for
example, the failure of on-cycle link A-B can be protected by
the route B-C-D-A. Therefore, the backup wavelengths on
Hamiltonian cycle should be enough to protect the working
wavelengths on the failed link; for example, if there are four
working wavelengths on link A-B, there will need four
backup wavelengths on each link on Hamiltonian cycle.
Another case is in Fig. 1(b), where any single failure of
straddling link can be protected by two available routes on
Hamiltonian cycle; for example, the failure of straddling link
A-C can be protected by two routes A-B-C and A-D-C.
Therefore, the backup wavelengths on Hamiltonian cycle
should be enough to protect the half of working wavelengths
on the failed link; for example, if there are six working
wavelengths on link A-C, there will need three backup
wavelengths on each link on Hamiltonian cycle.