Abstract :
The traditional single path routing can cause imbalanced link utilization and is not efficient for all traffic types such as long lived large flows. Moreover, it can lead to low network throughput and high network latency. Traffic engineering TE is a key solution to solve these problems of single path. The main purpose of TE is to optimize the network resource utilization and improve network performance by measuring and controlling network traffic. One of the TE approach for large flows is multipath routing which distribute traffic load among available multiple paths. However, most of multipath solutions do not classify traffic flows for example elephant or mice and do not concern the existing delays of routes. Therefore, to be intelligent multipath routing based on traffic types, we proposed three main folds 1 large flow detection approach by using sFlow analyzer in real time, 2 measuring end to end delays of available paths between source node and destination node where large flow occurred and 3 reroute the large flow to minimum round trip time delay path in order to improve network performance. Through experimental results, our proposed method gains over 30 77 throughput improvement over reactive forwarding application which is implemented in ONOS controller. Aung Htein Maw "Traffic Engineering in Software-Defined Networking (SDN)" Published in International Journal of Trend in Scientific Research and Development (ijtsrd), ISSN: 2456-6470, Volume-3 | Issue-5 , August 2019, URL: https://www.ijtsrd.com/papers/ijtsrd26618.pdf Paper URL: https://www.ijtsrd.com/computer-science/computer-network/26618/traffic-engineering-in-software-defined-networking-sdn/aung-htein-maw
Keyword :
Computer Network, multipath, sdn, traffic engineering