Analysis of ADSL traffic on an IP backbone link

Abstract

Measurements from an Internet backbone link carrying TCP traffic towards different ADSL areas are analyzed in this paper. For traffic analysis, we adopt a flow based approach and the popular mice/elephants dichotomy. The originality of the experimental data reported in this paper, when compared with previous measurements from very high speed backbone links, is in that commercial traffic comprises a significant part generated by peer-to-peer applications. This kind of traffic exhibits some remarkable properties in terms of mice and elephants, which are described in this paper. It turns out that by adopting a suitable level of aggregation, the bit rate of mice can be described by means of a Gaussian process. The bit rate of elephants is smoother than that of mice and can also be well approximated by a Gaussian process.

Prevention Mechanism on DDOS Attacks by using Multilevel Filtering of Distributed Firewalls

“…It permits traffic from the trusted location to untrusted locations and it does not need any external configuration. The main purpose of firewall is maintaining the untrusted or non-requested users from accessing our computers [6]. A firewall may be a software or hardware and it is normally placed at the network with the range and safeguards the incoming and outgoing connections.…”

Evaluation of 'user-oriented' and 'black-box' traffic models for link provisioning

“…Often r plays the role of the access rate, see for instance Ben Fredj et al [5]; supported by extensive measurements D is often assumed to have a power-law tail, see for instance Crovella and Bestavros [2]. We also refer to the detailed study by Ben Azzouna et al [7] and the references therein.…”

“…Note that we have focused on traffic from and to the access network (that is, to and from the core network (Internet), respectively), i.e., traffic that is sent over the 'uplink'. LAN traffic as well as backbone traffic, has been subject of various other studies, see e.g., [1], [7].…”

“…Such an approach was pursued in recent studies on flow-level modeling, e.g., [7]. There it was found that it is possible to describe real traffic using an M/G/∞ model, but the accuracy of the fit is at the expense of the number of sub-streams, and the corresponding tuning parameters (for instant the threshold that distinguishes the mice from the elephants).…”

“…Recently attention shifted somewhat to user-oriented models, such as the flow-level models of Ben Fredj et al [5] and Barakat et al [6]. We also mention a recent study by Ben Azzouna et al [7] who succeed in finding a detailed flow-level description of ADSL traffic by decomposing the aggregate stream into several classes, of which the parameters are estimated separately. As indicated above, the thresholds characterizing the different 'flow groups' are parameters, and need to be tuned; it is also noted that some components are estimated by black-box models (i.e., models in which the user behavior cannot be recognized).…”