[This article belongs to Volume - 55, Issue - 02, 2023]
Gongcheng Kexue Yu Jishu/Advanced Engineering Science
Journal ID : AES-30-11-2023-36

Title : BOOSTED DECISION TREE-BASED AGGREGATION METHOD FOR NETWORK ANOMALY DETECTION
S.Pradeep1, Dr.A.Geetha2

Abstract :

Software-defined networking (SDN) is an evolutionary networking paradigm that is being adopted by a large community and cloud providers. Network administrators must oversee all network devices and establish appropriate policies to respond to various network events. A hybrid SDN (hSDN) solution that shifts between centralized and distributed operational modes in response to network conditions is proposed in this work, which also examines how unreliable controller-to-node communication channels impact network performance. Finally, the global SDN controller splits traffic flow while the SDNs test the multi-path routing under a variety of network conditions (such as link impairments or packet loss ratios) to determine its operational limits. When control channel packet loss ratios increased, the proposed method significantly increased aggregated throughput while only marginally increasing average latency (for instance, 29% throughput improvement for 20% control packet losses). Because of this, networks can function in challenging conditions that they would not be able to under conventional centralized SDN control.