Belén Carro Martínez

Ensemble network traffic classification: Algorithm comparison and novel ensemble scheme proposal

Abstract Network Traffic Classification (NTC) is a key piece for network monitoring, Quality-of-Service management and network security. Machine Learning algorithms have drawn the attention of many researchers during the last few years as a promising solution for network traffic classification. In Machine Learning, ensemble algorithms are classifiers formed by a set of base estimators that cooperate to build more complex models according to given training and classification strategies. Resulting models normally exhibit significant accuracy improvements compared to single estimators, but also extra time cost, which may obstruct the application of these methods to online NTC. This paper studies and compares the performance of seven popular ensemble algorithms based on Decision Trees, focusing on model accuracy, byte accuracy, and latency to determine whether ensemble learning can be properly applied to this modeling task. We show that some of the studied algorithms overcome single Decision Tree in terms of model accuracy and byte accuracy. However, the notable latency increase hinders the application of these methods in real time contexts. Additionally, we introduce a novel ensemble classifier that exploits the imbalanced populations presented in traffic networks datasets to achieve faster classifications. The experimental results show that our scheme retains the accuracy improvements of ensemble methods but with low latency punishment, enhancing the prospect of ensembles methods for online network traffic classification.

Automatic UMTS system resource dimensioning based on service traffic analysis

Mobile network operators base their TDM network capacity dimensioning on Erlang B models. This approach was valid in legacy GSM networks. However, current Universal Mobile Telecommunications System networks deal with different resource consumption services such as voice, video call or data, and different limiting resources such as baseband processing capacity, transmission link capacity to the RNC, or spreading code tree. Operators need models to decide which resource must be upgraded, according to the demand of the services, in order to achieve expected overall service accessibility (i.e., the complementary of blocking probability). Network operation requires detecting when degradation is due to a lack of resources or to a hardware malfunction. Also, when operators need to prevent blockage in a high-capacity demanding event (for which they only have traffic predictions for each service) it is far from trivial to dimension resources. We have implemented a Kaufman Roberts approach to characterize the multiservice resource demand. Using real reported traffic Key Performance Indicators to calibrate the model, an estimated accessibility is obtained at a per-resource level and combined to find global estimated accessibility. The proposed model is intended to assist network operation, estimating individual resource shortage, differentiating congestion from hardware failures, and predicting the necessary resources to be deployed to tackle a high-capacity demanding event.

Integrated Multi-purposed Testbed to Characterize the Performance of Internet Access over Hybrid Fiber Coaxial Access Networks

This paper presents an experimental testbed to study the noise effect on the performance of the transport layer over Hybrid Fiber Coaxial (HFC) networks. We have successfully designed and implemented an integrated complex testbed, which is suitable not only for laboratory environments but also reusable for real-world networks. The main purpose of the testbed is modeling the residential broadband data network using hardware simulation under several noise conditions and observing the effects on the performance of popular Internet applications as well as native TCP/UDP performances. A large number of public domain Internet measurement tools have been evaluated, from which several selective software tools have been used. In addition, new software has been developed to combine all software and hardware devices. Based on the testbed, we were able to study several issues of TCP/UDP over HFC networks by making a large number of automatic measurements and analysis. The testbed infrastructure would be very useful for cable operator and end users for monitoring and troubleshooting HFC networks, and can be effectively reused for related studies in similar environments such as wireless and DSL.

Flexible COM-based Software Solution for HFC Network Performance Monitoring

Emerging HFC (hybrid fiber coaxial) networks with activated return path for data communications requires careful performance monitoring and preventative maintenance over the entire network. Return path is especially problematic due to the noise funneling effects that can strongly affect digital services. Besides, the HFC networks which base on existing CATV infrastructures require many considerations. The special characteristics of HFC network lead to the needs of taking into account many issues related to software, hardware, measurements, etc. This article focuses on the designing issues of HFC network monitoring software to deal with the noise problem in the physical layer, in respect to the trends of measurement software, and describes a performance monitoring software framework, which applies the Component Object Model (COM) approach. The software has been developed in the CEDETEL Lab — ESTI, University of Valladolid, in collaboration with RETECAL — Spain. The article is organized as follows: the first section addresses the general problems of HFC performances. The second section presents the typical issues of monitoring the performance of HFC physical layer. In the third section, the key issues of designing software for HFC network monitoring, and the COM approach will be discussed. In the fourth section, an HFC performance monitoring application, which has been developed in respect to the issues mentioned on the previous sections will be described. Finally, the last section presents conclusions and future works.