Information Technology Reference
In-Depth Information
3.2. Partial SRLG-Disjoint Shared Path Protection with Differentiated
Reliability (Pd-SPP)[23]
In order to improve the network resources utilization, [23] uses shared risk link group
(SRLG) -disjoint, the k -shortest paths algorithm and differentiated reliability. The shared risk
link group is a group of network links in which a common physical resource (such as fiber,
cable) is shared among them, and any link failure within the group will cause the failure of all
links in that group. The information on SRLGs is manually attained based on the knowledge
of the physical fiber plant of the network by the network operator. Note the reliability is the
probability that a system or connection will operate correctly in a period of time [23]. For the
source node s , destination node d , bandwidth request b , and requested reliability r , the
procedure of Pd-SPP is executed as follows:
When a connection request arrives, a temporary graph is first created using the links that
their free bandwidths are more than b . Second, the k -shortest path algorithm is executed until
k candidate working paths are found. Then, for each candidate working path, all the links and
nodes belonging to candidate working paths are added to the temporary graph and the
bandwidth of all added links are recorded. The Dijkstra's algorithm is run to compute the
shortest path. This process is executed until the path pair is found. In order to satisfy the
reliability r , Eq . (1) should be verified:
where F b/w is the conditional failure probability of the backup path when the working path
( P wp ) fails, R represents the requested reliability of the connection, l i and r i are the wavelength
link i and the reliability of l i , respectively.
After determining the path pair, bandwidth resources are assigned to the path, where R
free bandwidths are allocated to each link on a working path. Moreover, for each link
belonging to a backup path, the number of free bandwidths that can be assigned to the backup
path is determined based on the shareable backup bandwidth. Finally, the connection is
established. If no candidate path pair is found, the connection request is blocked. In order to
release the working and backup paths, a reversed process of the call setup is executed.
3.3. Two-Segment (TS) Protection [15]
In order to prevent “traps” in survivable WDM mesh networks, the new segment
protection is suggested in [15]. The trap problem occurs when an SRLG-disjoint backup path
cannot be found for an active path [15]. The main idea of TS is that an active path should be
protected through two backup segments if there is a trap problem for the given active path.
The procedure of the TS algorithm is as follows:
For a connection request and its path ( p ) from source node s to destination node d
and intermediate nodes n i (where i =1,…, k ), first the links belonging to p as well as
other links that are not SRLG-diverse with p should be removed (i.e., the links that
use the same SRLG with p ).
Search WWH ::

Custom Search