Passive estimation of TCP round-trip times
ACM SIGCOMM Computer Communication Review
More Netflow Tools for Performance and Security
LISA '04 Proceedings of the 18th USENIX conference on System administration
Fine-grained traffic classification with netflow data
Proceedings of the 6th International Wireless Communications and Mobile Computing Conference
Two samples are enough: opportunistic flow-level latency estimation using netflow
INFOCOM'10 Proceedings of the 29th conference on Information communications
Evaluating IPv6 adoption in the internet
PAM'10 Proceedings of the 11th international conference on Passive and active measurement
Analysis of the impact of sampling on NetFlow traffic classification
Computer Networks: The International Journal of Computer and Telecommunications Networking
Hi-index | 0.00 |
Netflow can be employed for accounting, anomaly detection and network monitoring, and can bring new data source for network management. But most IPv6 routers in CERNET2 backbone network don't provide IPFIX or NetFlow flow record function. Netflow flow records don't have the network performance information, such as RTT and packet loss ratio, so we hardly use the Netflow data to analyze the network performance. In this paper, we designs a NetFlow v9 measurement system (N9MS) which converts IPv6 packet headers into the NetFlow v9 flow records and monitors the link performance with these flow records. The N9MS has two improvements to the traditional Cisco's sampled NetFlow feature. Firstly, the Cisco's sampling strategy is to sample packets, while that in the N9MS is flow sampling which can keep all packets in the sampled flows to infer network performance based on these sampled packets. Secondly, N9MS directly uses these sampled packets to calculate these performance metrics, such as round trip time (RTT) and packet loss ratio. We also define both RTT and packet loss ratio fields in the scalability NetFlow v9 template format. In this paper we also use the N9MS to monitor a 10Gbps backbone link between Nanjing site and the CNGI-CERNET2 backbone, and give some experimental results.