TY - JOUR
T1 - Inter-autonomous system provisioning for end-to-end bandwidth guarantees
AU - Ho, Kin-Hon
AU - Howarth, Michael
AU - Wang, Ning
AU - Pavlou, George
AU - Georgoulas, Stylianos
PY - 2007/12/10
Y1 - 2007/12/10
N2 - This paper addresses the issue of provisioning end-to-end bandwidth guarantees across multiple Autonomous Systems (ASes). We first review a cascaded model for negotiating and establishing service level agreements for end-to-end bandwidth guarantees between ASes. We then present a network dimensioning system that uses traffic engineering mechanisms for the provisioning of end-to-end bandwidth guarantees. The network dimensioning system solves two problems: (1) the economic problem of how to determine the optimum amount of bandwidth that needs to be purchased from adjacent downstream ASes at a minimum total cost; (2) given the available bandwidth resources within and beyond the AS as a result of (1), the engineering problem of how to assign bandwidth guaranteed routes to the predicted traffic while optimizing the network resource utilization. We formulate both as integer-programming problems and prove them to be NP-hard. An efficient genetic algorithm and an efficient greedy-penalty heuristic are, respectively, used to solve the two problems and we show that these perform significantly better than simple heuristic and random approaches.
AB - This paper addresses the issue of provisioning end-to-end bandwidth guarantees across multiple Autonomous Systems (ASes). We first review a cascaded model for negotiating and establishing service level agreements for end-to-end bandwidth guarantees between ASes. We then present a network dimensioning system that uses traffic engineering mechanisms for the provisioning of end-to-end bandwidth guarantees. The network dimensioning system solves two problems: (1) the economic problem of how to determine the optimum amount of bandwidth that needs to be purchased from adjacent downstream ASes at a minimum total cost; (2) given the available bandwidth resources within and beyond the AS as a result of (1), the engineering problem of how to assign bandwidth guaranteed routes to the predicted traffic while optimizing the network resource utilization. We formulate both as integer-programming problems and prove them to be NP-hard. An efficient genetic algorithm and an efficient greedy-penalty heuristic are, respectively, used to solve the two problems and we show that these perform significantly better than simple heuristic and random approaches.
KW - End-to-end bandwidth guarantees
KW - Bandwidth provisioning
KW - Traffic engineering
UR - http://www.scopus.com/inward/record.url?scp=36549086717&partnerID=8YFLogxK
UR - https://www.scopus.com/record/pubmetrics.uri?eid=2-s2.0-36549086717&origin=recordpage
U2 - 10.1016/j.comcom.2007.09.008
DO - 10.1016/j.comcom.2007.09.008
M3 - RGC 21 - Publication in refereed journal
SN - 0140-3664
VL - 30
SP - 3757
EP - 3777
JO - Computer Communications
JF - Computer Communications
IS - 18
ER -