Information Technology Reference
In-Depth Information
3.6. Sub-Path Protection Based on Auxiliary Virtual Topology (SPAVT) [26]
In order to solve the problem of the protection switching time, SPAVT has been
proposed, where the protection switching time denotes the time taken from the instant a link
fails to the instant the backup path of a connection traversing the failed link is enabled [26]. In
this method, the virtual topology is formed based on the multiple pairs of working and backup
paths for each node pair in the network using an off-line mode. After receiving a connection
request, the Dijkstra's algorithm is run once within the virtual topology to find a virtual route.
Finally, the optimal pair of sub-paths is selected. The SPVAT algorithm is executed as
follows:
•
The set of primary paths between each node pair (
x
,
y
) is created through the
k
-
shortest path algorithm.
•
The set of backup paths for each node pair (
x
,
y
) is computed by
k
-shortest path
algorithm. Note the backup paths are link-disjoint with the corresponding primary
path.
•
The virtual topology
T
is made. When a connection request
R
(
m
,
n
) arrives, a virtual
route such as
m
-a
-b
-n
is computed. Note that (
m
,
n
) is virtual node pair
corresponding to node pair (
m
,
n
) in
T
. Moreover, a
and b
are virtual nodes. If an
appropriate route is found, the sub-primary paths in the set of primary paths between
each node pair (
m
,
a
), (
a
,
b
) and (
b
,
n
) on the virtual route
m
- a
- b
- n
are
selected, considering least working resources. Moreover, sub-backup paths are
elected similar to the way explained for selecting sub-primary path.
•
The primary path, backup path, and consumed resources are recorded for the
connection request and then the network state is updated. The system then waits for a
new connection request arrival. If no appropriate path is found, the request is blocked
and the network state is updated.
3.7. Fault-Tolerate Ant-Based Survivable Routing with Shared Backup Path
(FT-ABR) [27]
This approach is based on the ant-colony behavior and focuses on dynamic routing and
wavelength assignment (RWA) problem. The Dynamic RWA is performed online when
connection requests arrive. This approach assumes that ant-based mobile agents run in a
separated control plane that is carried out in a packet switching network with the same
topology as optical network, or in the optical domain where control data is transported on a
dedicated wavelength [27]. Moreover, each network node has two tables: a routing table with
a set of feasible protection cycles between source-destination nodes, and a pheromone table
for mobile agents. In order to increase the network performance, low-cost cycles are reported
to each network node through mobile agents.
