An approach for efficient, accurate, and timely estimation of traffic matrices

Abstract

Network operators desire to obtain accurate evaluations of the traffic matrices of their networks because they are critical inputs to many network functions such as traffic engineering, capacity provisioning and anomaly detection. Under the current Internet architecture, however, it is extremely challenging to precisely measure the traffic between an ingress and egress node pair. In this paper, we argue that a future Internet should make it easy for network operators to be aware of the accurate traffic matrices of their networks in an efficient and timely manner. In particular, we present the requirements for TM estimation and the corresponding implications on the future Internet architecture. Based on these implications, we then present a future Internet architecture that makes it easy to accurately, efficiently, and timely estimate traffic matrices. We also present numerical results to demonstrate the performance of the architecture in estimating traffic matrices.

Improving Network Security by Dynamically Changing Path Identifiers in Future Internet

“…The preliminary results show that CoLoR satisfies many requirements of the future Internet including being information-centric, encouraging innovation, and providing efficient support for mobility, multicast, multi-homing, and middleboxes [10]. In addition, CoLoR makes it easy to accurately, efficiently and timely estimate the traffic matrices of networks [13], and to efficiently integrate information-centric networking and software-defined networking [14]. Furthermore, the work in [12] analyzes CoLoR's inherent advantages on security and illustrates how it defenses against traditional attacks (without using the mechanism proposed in this paper).…”