Shortest path first with emergency exits
SIGCOMM '90 Proceedings of the ACM symposium on Communications architectures & protocols
Routing in the Internet
QoS routing using alternate paths
Journal of High Speed Networks - Special issue on quality of service routing and signaling
Computer Networks
Computers and Intractability: A Guide to the Theory of NP-Completeness
Computers and Intractability: A Guide to the Theory of NP-Completeness
Algebra and algorithms for QoS path computation and hop-by-hop routing in the internet
IEEE/ACM Transactions on Networking (TON)
Combinatorial algorithms for feedback problems in directed graphs
Information Processing Letters
On path selection for traffic with bandwidth guarantees
ICNP '97 Proceedings of the 1997 International Conference on Network Protocols (ICNP '97)
Hi-index | 0.00 |
In an Open Shortest Path First (OSPF) based best effort network, the OSPF shortest path can become the bottleneck when congestion occurs. OSPF cannot forward packets though less congested alternate paths. Hence, OSPF cannot be used to provide Quality of Service. Earlier, we reported a Load Sensitive Routing (LSR) algorithm which finds alternate path based on OSPF property. In the earlier work, the LSR used global coefficients i.e. all the nodes in the network use the same coefficient for a given destination. But assigning network-wide global coefficient may lead to uneven distribution of alternate paths. That is, some nodes may have many alternate paths whereas others may have few or none. The use of global coefficient was thought to be necessary to make the protocol loop free. In this study, we allow nodes to choose LSR coefficients locally (we call the coefficient L-LSR coefficient) while still retaining the loopfree property. This leads to nodes having more number of alternate paths than the case where they had to use global coefficient. But allowing local coefficients makes the process of calculating the local coefficients complex. Since our protocol has to be loop free, the local coefficients have to be calculated such that the loop free OSPF property is still satisfied. This paper presents detailed algorithm for calculating L-LSR coefficients. Using simulation, we show that L-LSR algorithm not only performs better than OSPF, but also has very significant performance improvement over the other LSR family of algorithms.