Jiri Slachta

Embedded multiplatform SIP server solution

This paper describes BESIP (Bright Embedded Solution for IP Telephony) project. It was developed by LipTel team supported by CESNET. Primary goal was to implement multiplatform embedded SIP communication server with unified configuration interface. The paper explains and describes the whole concept and individual modules, acquaints with the current state and with the future intents. SIP server is based on OpenWRT project core and, there is Asterisk and Kamailio inside as SIP engines. Kamailio was selected for security and reliability, Asterisk for PBX functions. Next to this, there is Web frontend for managing device, Kamailio and Asterisk. Last, but very important part, is NETCONF server which is part of image for unified management.

Monitoring of Speech Quality in Full-Mesh Networks

The quality of speech in network environment is a growing concern of all the companies and institutions, that migrated or are planning to migrate their voice communications to IP based technologies. This trend together with the increasing amount of Internet traffic in general poses a great challenge for network administration to be able to configure and manage the network so it can carry the multimedia traffic without the excessive delays, which degrade the quality of speech as it is perceived by the end users. Continuous monitoring of key nodes of internal infrastructure as well as external interconnections is the possible way to increase the quality of multimedia transmissions for most users, because it allows the network administrators to be informed as soon as the problem rises and consequently to change the network routing and queuing policies to bypass the connection experiencing the quality issues. The tool for this type of monitoring and its architecture are described in detail in this paper.

Utilizing the Neural Networks for Speech Quality Estimation Based on the Network Characteristics

The paper deals with an issue of the speech quality estimation in Voice over IP technology under packet loss. Packet loss is a major problem for real-time Internet applications, we applied four-state Markov model for modeling the impact of network impairments on speech quality, afterwards, the resilient back propagation (Rprop) algorithm was used to train a neural network. The general and RFC3611-compliant solution, which allows for quick and precise speech quality estimation without the need to analyze or model the voice signal carried by the RTP (Real-time Transport Protocol) packets, is the contribution of this paper. The proposed solution is tested on G.711 A-law and further generalizes the already presented concepts of the speech quality estimation in the IP environment. The proposed approach of speech quality assessment belongs to non-intrusive methods and is based on the back-propagation neural networks.

Modelling speech and video quality in networks based on internet protocol

A development of the tool meant to model and estimate the qualitative parameters of the speech and video transmission is the main aim of this paper. The influence of queuing techniques of individual network elements on the quality of service is explored and a model to predict this influence is created based on this information. By using this model the network administrators can successfully plan the network infrastructure with the emphasis on the communication quality, which is so important these days. The results of the model are presented using the standardised scales of mean opinion score and peak signal-to-noise ratio, which makes the results easy to understand. To complete the picture the data transmissions are explored as well making the presented tool a universal mean to predict quality of service parameters for all kinds of services.

Modelling the impact of network features on speech quality in voice over IP

With Voice and Video over IP spreading to the new areas of mobile communications, the speech quality becomes a concern of high importance. Monitoring systems of various types are deployed massively in the infrastructure of all the providers. These systems provide the knowledge base for the providers to set the infrastructure of theirs properly and efficiently. However, this approach only allows the administrators to react on the situation that already exists with a certain delay. In this paper the possible approach for speech quality estimation using the network fingerprint, data preprocessing and neural networks is presented. The description of the designed system is followed by the experimental measurements that confirm the validity of the approach. This way, the paper attempts to further improve the currently known methods for speech quality estimation.

Speech Quality Measurement in IP Telephony Networks by Using the Modular Probes

The paper presents a system for monitoring the speech quality in the IP telephony infrastructures using modular probes that are placed at key nodes in the network. The system is developed mainly within research of the Czech educational and scientific network, but the tool can generally be used in any IP telephone traffic. The system is capable of using an objective intrusive methods to measure speech quality dynamically and the results are collected on a central server. Information about the speech quality are displayed in the form of automatically generated maps and tables. The article is logically divided into a description of the technology and algorithms, as well as the implementation procedure is presented and measured results were validated in real traffic. Contribution of the work consists of a new system design utilizing modular probes for the measurement of speech quality in IP telephony networks and the result of applied research enables administrators to respond and optimize network traffic efficiently.

Comparison of artificial intelligence classifiers for SIP attack data

Honeypot application is a source of valuable data about attacks on the network. We run several SIP honeypots in various computer networks, which are separated geographically and logically. Each honeypot runs on public IP address and uses standard SIP PBX ports. All information gathered via honeypot is periodically sent to the centralized server. This server classifies all attack data by neural network algorithm. The paper describes optimizations of a neural network classifier, which lower the classification error. The article contains the comparison of two neural network algorithm used for the classification of validation data. The first is the original implementation of the neural network described in recent work; the second neural network uses further optimizations like input normalization or cross-entropy cost function. We also use other implementations of neural networks and machine learning classification algorithms. The comparison test their capabilities on validation data to find the optimal classifier. The article result shows promise for further development of an accurate SIP attack classification engine.

Multi-agent system for monitoring the quality of speech in computer networks

The paper deals with speech quality monitoring system using probes placed on individual network nodes operating VoIP (Voice over Internet Protocol) services. The system is developed mainly within research of the Czech educational and scientific network, but the tool can be deployed generally in any VoIP infrastructure. In order to assess speech quality, an intrusive measurement approach has been applied and results from probes are instantly collected on a central server. Information on speech quality is visualized in a graph respecting topology of probes. In the article, we provide an overall description of the technology and algorithms used in the speech quality monitoring system and results achieved in this applied research were verified in real operation. Contribution of the work lies in the new proposed multi-agent system enabling speech quality monitoring and its verification in Czech academic network.

Management, monitoring and multimedia services on embedded systems

The paper deals with advanced techniques of automated configuring devices based on OpenWrt Linux distribution. In the paper, the overall concept and the architecture of the system is described, one part is devoted to the tool for automatic firmware building and the issues of key modules are also discussed. The solution has been built as a universal open-source modular system within the framework of a BESIP project (Bright Embedded Solution for IP Telephony) since May 2011. In the framework of the project, more than fifty patches were implemented and accepted in OpenWrt and several packages regarding IP telephony were ported into this distribution as well. The contribution of this applied research and experimental development lies in the overall concept of the BESIP and the approach used in the build system which can help networkers to maintan increasing amount of devices that are under their administration.

VoIP attacks detection engine based on neural network

The security is crucial for any system nowadays, especially communications. One of the most successful protocols in the field of communication over IP networks is Session Initiation Protocol. It is an open-source project used by different kinds of applications, both open-source and proprietary. High penetration and text-based principle made SIP number one target in IP telephony infrastructure, so security of SIP server is essential. To keep up with hackers and to detect potential malicious attacks, security administrator needs to monitor and evaluate SIP traffic in the network. But monitoring and following evaluation could easily overwhelm the security administrator in networks, typically in networks with a number of SIP servers, users and logically or geographically separated networks. The proposed solution lies in automatic attack detection systems. The article covers detection of VoIP attacks through a distributed network of nodes. Then the gathered data analyze aggregation server with artificial neural network. Artificial neural network means multilayer perceptron network trained with a set of collected attacks. Attack data could also be preprocessed and verified with a self-organizing map. The source data is detected by distributed network of detection nodes. Each node contains a honeypot application and traffic monitoring mechanism. Aggregation of data from each node creates an input for neural networks. The automatic classification on a centralized server with low false positive detection reduce the cost of attack detection resources. The detection system uses modular design for easy deployment in final infrastructure. The centralized server collects and process detected traffic. It also maintains all detection nodes.