TY - GEN
T1 - Fast link-disjoint path algorithm on parallel reconfigurable processor DAPDNA-2
AU - Kihara, Taku
AU - Shimizu, Sho
AU - Arakawa, Yutaka
AU - Yamanaka, Naoaki
AU - Shiba, Kosuke
PY - 2008/12/1
Y1 - 2008/12/1
N2 - This paper proposes fast parallel link-disjoint path algorithm using dynamically reconfigurable processor and implements it on DAPDNA-2 (IPFlex Inc) which is newly structured. The conventional k-shortest path algorithm finds multiple link-disjoint paths between the source node and the destination node. When the network scale is large, the calculation time of k-shortest path algorithm increases rapidly. Moreover, in the worst case, k-shortest path algorithm can not find optimum link-disjoint path pair because this algorithm always finds the shortest path at first and removes those links from network. Our proposed algorithm collects all path information in the network and calculates optimum linkdisjoint path pair (i.e. minimum cost link-disjoint path pair) at high speed by using parallel operation. Additionally, our proposed algorithm finds optimum link-disjoint path pair at a high rate in a limited of calculation time. The evaluation shows our proposed algorithm can decrease the calculation clock about 90%. copyright
AB - This paper proposes fast parallel link-disjoint path algorithm using dynamically reconfigurable processor and implements it on DAPDNA-2 (IPFlex Inc) which is newly structured. The conventional k-shortest path algorithm finds multiple link-disjoint paths between the source node and the destination node. When the network scale is large, the calculation time of k-shortest path algorithm increases rapidly. Moreover, in the worst case, k-shortest path algorithm can not find optimum link-disjoint path pair because this algorithm always finds the shortest path at first and removes those links from network. Our proposed algorithm collects all path information in the network and calculates optimum linkdisjoint path pair (i.e. minimum cost link-disjoint path pair) at high speed by using parallel operation. Additionally, our proposed algorithm finds optimum link-disjoint path pair at a high rate in a limited of calculation time. The evaluation shows our proposed algorithm can decrease the calculation clock about 90%. copyright
UR - http://www.scopus.com/inward/record.url?scp=66149153350&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=66149153350&partnerID=8YFLogxK
M3 - Conference contribution
AN - SCOPUS:66149153350
SN - 4885522323
SN - 9784885522321
T3 - 2008 14th Asia-Pacific Conference on Communications, APCC 2008
BT - 2008 14th Asia-Pacific Conference on Communications, APCC 2008
T2 - 2008 14th Asia-Pacific Conference on Communications, APCC 2008
Y2 - 14 October 2008 through 16 October 2008
ER -