4.BACKUP PATH PROTECTION ALGORITHM:-
Problem Formulation:
The multiple next-hops found by MNP can be utilized to detour around failed
links. However, there is no guarantee that such backup next-hops can always be
found, and thus the protection capability of MNP is limited (so do the other
cooperationfree mechanisms). To further improve the network availability, we
propose backup path protection (BPP) to find backup paths when MNP fails to
do so. Since forwarding along such paths needs additional cooperation/signaling
mechanisms, we want to utilize them as little as possible. Given a network
G(V,E) and the set of next-hops computed by MNP, our objective is to find a
set of links and the corresponding backup paths, such that the network
availability requirement (A(G) ≥ R, where R is a parameter chosen by network
designers or operators) can be satisfied, while the number of such links is
minimized.
Algorithm:
Algorithm 3 illustrates the framework of BPP.
5.CONCLUSION
Traditional intra-domain routing protocols react to network failures by globally
exchanging link state advertisements and recalculating routing table. This
mechanism will greatly increase the convergence time, resulting in large
number of packets are discarded when network failures occur. Therefore, lots of
routing protection schemes have been proposed to enhance network availability
to support mission-critical and real-time applications in the internet. However,
all of the aforementioned algorithms cannot strike a good balance between the
implementation of efficiency and network availability. Unlike the above works,
however, our main concerns are computational efficiency and network
availability, as these are critical for the algorithm. Therefore, we presented HLP
as a novel link protection scheme to achieve high network availability for link-
state routing networks. We described how HLP improves network availability
by combining MNP-e and BPP. HLP first computes multiple nexthops for
source-destination pairs, then selects a minimum number of links to protect, so
that it can meet network availability requirement without inducing significant
overhead. Evaluation results on real and synthetic networks show that HLP can
provide high network availability with low overhead. We believe networks can
be made more efficient and reliable by adopting this mechanism
References: -
1.)J. Zheng, H. Xu, X. Zhu, G. Chen, and Y. Geng, “Sentinel: Failure recovery
in centralized traffic engineering,” IEEE/ACM Trans. Netw., vol. 27, no. 5, pp.
1859–1872, 2019.
2.) S. Wang, H. Xu, L. Huang, X. Yang, and J. Liu, “Fast recovery for single
link failure with segment routing in SDNs,” in Proc. IEEE
HPCC/SmartCity/DSS, Aug. 2019, pp. 2013–2018.
3
.) https://ieeexplore.ieee.org/document/9042897
4.)
https://ieeexplore.ieee.org/search/searchresult.jsp?newsearch=true&que
ryText=computer%20network