We consider dynamic delay guarantee and bandwidth allocation in communications networks. Our scenario includes linear pricing scheme for both Quality of Service parameters. The goal is (i) to maximize the revenue and (ii) guarantee fair resource allocation for connections. On the contrary to the traditional Lagrangian approach, we approach the problem by modified one, where the sum of the weights of the scheduler acts as the penalty term. This modified approach yields closed form approximate algorithm for updating the scheduler weights, being very fast and real-time implementable. We compare the algorithm with the brute-force method, which optimizes weights in the large grid - optimal brute-force method has exponential complexity. The revenue obtained by the closed form method is about 99.9 % of the optimal, computationally expensive approach, thus being tempting both from the point of view of the service provider and the customers. NS-2 simulator is used in the experiments.

IETF RFC 2002 originally introduced the wireless Mobile-IP protocol to support portable IP addresses for mobile devices that often change their network access points with the Internet. The inefficiency of registration process, mainly within the handoff management of this protocol, produces large end-to-end packet delays, since all registrations take place after roaming the mobile node into a foreign subnet, and further degrades the system efficiency due to packet losses between subnets. The criterion to initiate a simple and fast full-duplex connection between the home agent and foreign agent is a very important issue to be considered by a work. 
In this paper, a workaround was developed, mapping the roaming behavior to Mobile-IP registration process to present the feasibility of designing and modeling a CIA: communication inter-agents procedure, as an extension to the basic Mobile-IP. It supposes an early registration of mobile node to a subnet predicted to roam into, to reduce the roaming duration and eliminate packet losses. Analytical results of corresponding state-transition diagram, as a function of network parameters, are presented. The heuristic of configuration file during practical setup session for registration parameters, on Cisco platform Router-1760 using IOS 12.3 (15)T is created. Finally, stand-alone performance simulations results on Simulink Matlab, within each subnet and also between subnets, are illustrated for reporting better end-to-end packet delays. Results verified the effectiveness of our Mathcad analytical manipulation and experimental implementation. It showed much lower values of end-to-end packet delay for Mobile-IP using CIA procedure. Furthermore, it reported packets flow between subnets, i.e., no more gaps for data losses.

The present document aims to explore the possibilities of introduction of the IPFIX (IP Flow Information eXport) and PSAMP (Packet Sampling) standards at implementation of a non-intrusive device for monitoring the QoS parameters in high speed networks. Since no implementation of such device, fully conform with both standards, currently exists, this paper provides the basic points for its development. It analyzes in detail all requirements related to documents released by the task groups that deal wi th the proposals of these standards. Furthermore, it suggests possible solutions of problems related to their implementation into the measuring device itself, as well as possible modifications of these standards in future versions, with aim to simplify the development of further IPFIX and PSAMP based applications. The requirements of these standards are analyzed step-by-step, in the same way as they are found in the documents of their respective task groups.
The majority of solutions presented in the paper is, or will be, included in the measuring device which is being developed in the Computer Network Laboratory of the Department of Computers and Informatics, at the Technical University, Kosice. At the same time, the majority of important aspects of these standards is assessed in connection to this tool, and conclusions related to their implementation and further development are drawn.

In this paper we present a Mobile IPv6 analysis related to its performance in heterogeneous multi-access environment. The used access technologies include IEEE 802.3 Ethernet, IEEE 802.11b Wireless LAN and Bluetooth. We also discuss the capability of Mobile IPv6 to provide Always-Best-Connected access in multi-access environment to the user and applications. The analysis is made according to real-life tests in Mobile IPv6 for Linux (MIPL) test network. The Mobile IPv6 protocol by itself cannot handle all the tested scenarios in the best possible way even though it provides a good ground for the All-IP mobility management. We claim that a cross-layer approach would enable intelligent and proactive handovers, and thus improve the end- user experience.

Current trends in communication networks point to an aggregation of all kind of traffic (data, voice, etc) in the same transport technology. The emergence of real-time applications and the widespread use of wireless and mobile devices have generated the need to provide quality of service (QoS) support in wireless and mobile networking environments. QoS present a key problem of today's mobile networks and plays an important role in QoS provisioning for mobile networks. Supporting appropriate QoS for mobile networks is a complex and difficult issue because of the dynamic nature of the network topology and generally imprecise network state information, and has become an intensely active area of research in the last few years. Although lots of research has been done on supporting QoS in the Internet and other networks but they are not suitable for mobile networks and still QoS support for such networks remains an open problem. In this article we introduce basic definition of QoS in mobile networks and discuss trends in QoS for next generation mobile networks.

Delay is one of the crucial factors which influence QoS. This factor depends not only on the network throughput. In case of single channel networks, delay depends on the access mechanism as well. This article reflects single channel network with random access and how the volume of network state information can influence the throughput. Markov model was used to describe and model this type of network. Simulations were done and results were compared with the analytical model. The first simulation results are presented in this paper.

The VoIP quality of service can be either estimated by analyzing the voice data flow characteristics, or by recording a transmitted voice probe over the whole VoIP transport chain and comparing it to the original. When the quality estimation by the voice data flow characteristics is deployed, the three most prominent degradation factors are observed - the packet loss, the packet delay and the delay variance (also called jitter). Depending on the estimation model, the values of these degradation factors are used to compute the resulting quality rating estimation. The end-to-end quality measurement provides more dependable results, however, it is also more technically and computationally demanding. We describe our experiences with both methods of quality evaluation and measurement. An experimental VPN for VoIP QoS measurements was created and a voice probe was sent through this network while end-to-end quality and degradation factors were observed. For end-to-end measurements we used our own developed equipment that allowed us to synchronize the original and recorded voice probe and obtain the timing characteristics. The obtained results are also discussed in this article.

In this paper we present the traffic management functions in Asynchronous transfer mode (ATM) networks. In the first part of the paper commonly used traffic models for ATM traffic are discussed. Traffic management and its functions in the network are presented in the next part of the paper. Then connection admission control methods (CAC) as preventive traffic management mechanism are then presented. Overview of two CAC methods proposed in the literature from the groups of statistical based, measurement based and artificial intelligence based methods are presented in the final part of the paper.

The ADSL2+ lines are gradually replacing the first generation of ADSL in homes and small offices connecting to the Internet. Since there is a need to estimate the transmission speed, we have used MATLAB Web Server to design a simulator of xDSL lines that is available at our web pages. This paper compares simulations with measurement results and proposes a methodology for evaluation of measurements based on the suggestions from Slovak Telecom in connection with the practical introduction of ADSL2+ technology in 2006.
The original analogue telephone lines use xDSL digital systems to a lesser or greater extent. Let us focus on the trend in the metallic access networks field and the potential possibilities for further development. Apart from the evolution of ADSL (Asymmetric Digital Subscriber Line) standards the transfer methods for the very high speed VDSL (Very-high-speed Digital Subscriber Line) are being optimized. An ADSL line transmits digital signals at various transfer speeds in opposing directions of transfer. For the connection of larger companies, especially those, which operate their own application and web servers, ADSL connection is not suitable although the SHDSL (Single-pair High-bit-rate Digital Subscriber Line) connection can be used.
Allocation of bands for transfer directions downstream and upstream and also the total width of the band is dependent on the ADSL variant given by the ITU-T standard. According to the recommendation of the ITU-T G.992.1 today's systems use a frequency band of up to 1.104 MHz. The lower part of the spectrum is of course used for telephone channels and for this reason it counts on the coexistence of ADSL in one conduit with the original analogue telephone connection (ADSL/POTS), which is preserved thanks to separation filters (adapter, splitter). Apart from this, ADSL can also coexist with the basic ISDN-BA connection (ADSL/ISDN).

Congestion management (avoidance and control) is a set of mechanisms used to prevent network congestion and assist in the recovery from it. The constant growth of the Internet many congestion control mechanisms have been proposed and some of them implemented on the global Net. Most modifications to the original TCP protocol have faired rather well and have preserved the scalability and the stability of the Internet. The emergence of bandwidth intensive multimedia applications and the usage of UDP have proved that the existing congestion control mechanisms are not sufficient to prevent the network from exhaustion of its resources and possible congestion collapse. Multimedia applications often generate traffic flows that are unresponsive to network congestion. Consequently, the packets are not delivered to their destinations since they are being dropped at the intermediate nodes. In other words, transmitting unresponsive flows through a network is one of the main reasons behind possible network collapse, since the potential of the network is fully utilized, while too few packets reach the intended recipients. It is clear that the congestion management of unresponsive flows in presence of both responsive flows and unresponsive flows is an open problem. The paper describes the study into the possibility to use the mobile agent paradigm for management of mostly unresponsive flows and integrating this mechanism with the classical congestion control protocols in a pure TCP environment. To illustrate the main points of our position, we introduce a new model termed as CM4CC (Combined Model for Congestion Control).

The complexity of discrete systems developments process shows, that it is not realistic to expect, that the only universal formal method will exist, which will cover well all the aspects of this process. We are working towards formal methods integration in order to obtain deeper understanding of system under development/analysis. We chose Petri nets and process algebra, notations with complementary properties for integration. The research activity is very high in this area and the work was inspired by [9,10,11] and other papers published in past years. ACP2PETRI is a software tool for formal method transformations based on results of research performed in past years [1].

The goal of the article is to presents intrusion detections systems and design architecture of intrusion detection based on neural network self organizing map. In the report is described base problematic of neural network and intrusion detection system. The article further deals with specific design of intrusion detection architecture based on user anomaly behavior. A core of the designed architecture represents neural network SOM, which classifies monitored user behavior and determines possible intrusion of monitored computer system. Result of the designed architecture is simulations in real conditions. Acquired results of simulation assign expediencies of using neural network SOM in the intrusion detection systems.