Traffic analysis is a compound of strategies intended to find relationships, patterns, anomalies, and misconfigurations, among others things, in Internet traffic. In particular, traffic classification is a subgroup of strategies in this field that aims at identifying the application’s name or type of Internet traffic. Nowadays, traffic classification has become a challenging task due to the rise of new technologies, such as traffic encryption and encapsulation, which decrease the performance of classical traffic classification strategies. Machine learning (ML) gains interest as a new direction in this field, showing signs of future success, such as knowledge extraction from encrypted traffic, and more accurate Quality of Service management. ML is fast becoming a key tool to build traffic classification solutions in real network traffic scenarios; in this sense, the purpose of this investigation is to explore the elements that allow this technique to work in the traffic classification field. Therefore, a systematic review is introduced based on the steps to achieve traffic classification by using ML techniques. The main aim is to understand and to identify the procedures followed by the existing works to achieve their goals. As a result, this survey paper finds a set of trends derived from the analysis performed on this domain; in this manner, the authors expect to outline future directions for ML-based traffic classification.

/pdf/towards-the-deployment-of-machine-learning-solutions-in-3xcu79u8f4.pdf

Towards the Deployment of Machine Learning Solutions in Network Traffic Classification: A Systematic Survey

In recent decades, power line communication has attracted considerable attention from the research community and industry, as well as from regulatory and standardization bodies. In this article we provide an overview of both narrowband and broadband systems, covering potential applications, regulatory and standardization efforts and recent research advancements in channel characterization, physical layer performance, medium access and higher layer specifications and evaluations. We also identify areas of current and further study that will enable the continued success of power line communication technology.

/pdf/state-of-the-art-in-power-line-communications-from-the-47r4u55osx.pdf

State-of-the-art in Power Line Communications: from the Applications to the Medium

Based on an analysis of the statistical nature of IP traffic and the way this impacts the performance of voice, video, and data services, we question the appropriateness of commonly proposed quality-of-service mechanisms. This paper presents the main points of this analysis. We also discuss pricing issues and argue that many proposed schemes are overly concerned with congestion control to the detriment of the primary pricing function of return on investment. Finally, we propose an alternative flow-aware networking architecture based on a novel router design called cross-protect. In this architecture, performance requirements are satisfied without explicit service differentiation, creating a particularly simple platform for the converged network.

Internet traffic, QoS, and pricing

In this paper we present the architectural implications of flow-aware networking, or FAN, a new approach for realizing QoS guarantees. FAN performs traffic control based on user-defined flows using implicit admission control and per-flow scheduling. This allows adequate performance guarantees for streaming and elastic flows without requiring class of service distinction or relying on signalled traffic specifications. We argue that FAN is a new direction for Internet quality of service that is arguably necessary as an alternative to flawed classical architectures and sufficient to meet user performance requirements in a cost-effective way.

A new direction for quality of service: flow-aware networking

The content-centric networking (CCN) paradigm proposed by PARC holds considerable promise as the architecture for the future Internet but remains incomplete. In this paper we argue that it is necessary to supplement CCN with mechanisms enabling controlled sharing of network bandwidth by concurrent flows. Traffic control is necessary to ensure low latency for conversational and streaming flows and to realize satisfactory bandwidth sharing between elastic flows. These objectives can be realized using “network neutral” traffic controls based on per-flow fair bandwidth sharing. The paper describes necessary CCN mechanisms and discusses the strategies to be implemented by end systems and routers. A simulation-based performance evaluation demonstrates the effectiveness of the proposed mechanisms and and the impact of possible strategies. We describe our experience in implementing some of the mechanisms using the CCNx prototype software.

Flow-aware traffic control for a content-centric network

In this paper we present Cross-protect, a combination of router mechanisms allowing quality of service differentiation while maintaining the simple user-network interface of the best effort Internet. By associating implicit flow level admission control and per-flow fair queuing in a router it is possible to distinguish streaming and elastic flows and meet their respective quality requirements without requiring specific packet marking. We describe the implied mechanisms and justify the claimed performance and scalability properties by means of simulation and analysis.

http://ieeexplore.ieee.org/iel5/9133/28952/01303427.pdf

Cross-protect: implicit service differentiation and admission control

Despite its well-known advantages, per-flow fair queueing has not been deployed in the Internet mainly because of the common belief that such scheduling is not scalable. The objective of the present paper is to demonstrate using trace simulations and analytical evaluations that this belief is misguided. We show that although the number of flows in progress increases with link speed, the number that needs scheduling at any moment is largely independent of this rate. The number of such active flows is a random process typically measured in hundreds even though there may be tens of thousands of flows in progress. The simulations are performed using traces from commercial and research networks with quite different traffic characteristics. Analysis is based on models for balanced fair statistical bandwidth sharing and applies properties of queue busy periods to explain the observed behaviour.

Evaluating the number of active flows in a scheduler realizing fair statistical bandwidth sharing

Traffic classification (TC) is a fundamental task of network management and monitoring operations. Previous works relying on selected packet header fields (e.g. port numbers) or application layer protocol decoding techniques are becoming increasingly difficult and inefficient when facing encrypted traffic and peer-to-peer flows. In this paper, we address the problem of flow based TC using machine learning (ML) algorithms. Our work considers a supervised approach, namely eXtreme Gradient Boosting (XGBoost) algorithm, which has never been investigated for TC. Performance evaluation results show that we obtain 99.5% accuracy on a dataset containing real flows. Additionally, compared to other ML algorithms, XGBoost is the most accurate one.

On using eXtreme Gradient Boosting (XGBoost) Machine Learning algorithm for Home Network Traffic Classification

An enhanced flow-aware Internet is arguably a more effective means of ensuring adequate performance than implementing the complex standardized QoS architectures. This flow-aware network would provide flow-level performance guarantees for real time and data applications by implementing per-flow fair queueing and by limiting the impact of overload through flow level admission control. The paper discusses the feasibility of the implied router mechanisms and proposes original solutions that minimize the necessary overhead with respect to the current best effort network. Preferred solutions significantly reduce requirements for flow state by employing directly addressed bitmaps to record flow status, as necessary for scheduling and admission control, respectively.

Minimizing the overhead in implementing flow-aware networking

The time-varying capacity links, such as wireless links and power line communication (PLC) links, which are integrated to emergent connectivity devices of Home Networks (HNs) require QoS mechanisms to protect sensitive and critical HN flows. In this paper, we couple the “Iperf in lightweight TCP mode” probing technique (developed in a previous work to estimate available bandwidth) with the path selection procedure of the “Inter-MAC” software (which was derived from the European OMEGA project, to handle heterogeneous HN technologies). As we have shown in our previous work, the lightweight TCP Iperf probing is accurate over time-varying capacity links. On the contrary, the Inter-MAC default implementation biases the measurements, because it computes the available bandwidth by subtracting the bit counters of local home network interfaces from their correspondent physical capacities, which are supposed constants. We show that coupling our lightweight TCP Iperf technique with the Inter-MAC software is really convenient to avoid performance blind spots on HN paths. Additionally, we show a realistic use case on test bed with HN extenders based on WiFi/PLC dual-links. We show the possibility of efficiently protecting sensitive IPTV flows and enhancing the link utilization, when the PLC/WiFi links capacities are strongly degraded.

Abdesselem Kortebi

Papers

Cross-protect: implicit service differentiation and admission control

Evaluating the number of active flows in a scheduler realizing fair statistical bandwidth sharing

On using eXtreme Gradient Boosting (XGBoost) Machine Learning algorithm for Home Network Traffic Classification

Minimizing the overhead in implementing flow-aware networking

Available Bandwidth Probing for path selection in heterogeneous home Networks