Computer Networks

The Internet transports data generated by programs which cause various phenomena in IP flows. By means of machine learning techniques, we can automatically discern between flows generated by different traffic sources and gain a more informed view of the Internet.

In this paper, we optimize Waterfall, a promising architecture for cascade traffic classification. We present a new heuristic approach to optimal design of cascade classifiers. On the example of Waterfall, we show how to determine the order of modules in a cascade so that the classification speed is maximized, while keeping the number of errors and unlabeled flows at minimum. We validate our method experimentally on 4 real traffic datasets, showing significant improvements over random cascades.

This paper examines various techniques for estimating the intensity of Long-Range Dependence (LRD). Trial data sets with LRD are generated using Fractional Gaussian noise and Markov modulated Poisson process. The real data set collected in IITiS PAN is also used.

This paper introduces algorithms for surveillance applications of wireless sensor and actor networks (WSANs) that reduce communication cost by suppressing unnecessary data transfers. The objective of the considered WSAN system is to capture and eliminate distributed targets in the shortest possible time. Computational experiments were performed to evaluate effectiveness of the proposed algorithms. The experimental results show that a considerable reduction of the communication costs together with a performance improvement of the WSAN system can be obtained by using the communication algorithms that are based on spatiotemporal and decision aware suppression methods.

The usage of GPS systems for indoor localization is limited, therefore multiple indirect localization techniques were proposed over the years. One of them is a localization method based on Wi-Fi (802.11) access point (AP) signal strength (RSSI) measurement. In this method, a RSSI map is constructed via Localization Fingerprinting (LF), which allows localizing object on the basis of a pattern similarity. The drawback of LF method is the need to create the RSSI map that is used as a training dataset. Therefore, in this study a Wireless Sensor Network (WSN) is used for this task. The introduced in this paper energy aware localization method allows to acquire the actual RSSI map or broadcast a localization signal, if there is not sufficient information to perform the localization by using nearby APs. To localize objects in a given cell, various classifiers were used and their localization accuracy was analyzed. Simulations were performed to compare the introduced solution with a state-of-the-art approach. The experimental results show that the proposed energy aware method extends the lifetime of WSN and improves the localization accuracy.

Current mobile phones and tablets are equipped with two technologies for accessing the Internet: WiFi and Cellular. Deciding which of these two interfaces provides faster data transfer is often non-trivial, but most of the currently used devices use a simple priority scheme that prefers WiFi to Cellular.

In this paper, we propose a novel system that automatically selects the best link available in the current user location. The system periodically probes the bandwidth available on both links and makes statistical predictions, while avoiding excessive data and battery usage. We experimentally validated our approach using a dedicated application for Android.

The rapid improvement in low power micro-processors, wireless networks and embedded systems has boosted the desire to utilize the very significant resources of mobile devices. The mobile ad hoc network (MANET), an infrastructure-less wireless network, is an emerging technology and is best suited to provide communication between wireless mobile devices. Due to the nature of MANET, the network topology changes frequently, unpredictably and has created the new challenges as traditional routing protocols are not suitable for multi-hop communication in mobile ad hoc environments. There are number of routing protocols proposed for MANET, and development is active in this area, however, there is no single routing protocol which is best suited to address all the basic issues of heterogeneous MANETs. This diverse range of routing protocols have created a new challenge as in general the heterogeneous mobile devices cannot communicate with each other and thus are unable to facilitate the full exploitation of mobile resources. To overcome the above mentioned issues, this paper has proposed an Interoperable Framework for MANET, called an IF-MANET, which hides the complexities of heterogeneous routing protocols and provides a homogeneous layer for seamless communication between these routing protocols. The IF-MANET resides in user space and it will be implemented at application layer to provide runtime interoperability and platform independence.

New threats from the Internet emerging every day need to be analyzed in order to prepare ways of protection against them. Various honeypots combined into honeynets are the most efficient tool how to lure, detect and analyze threats from the Internet. The paper presents recent results in honeynet made of Dionaea (emulating Windows services), Kippo (emulating Linux services) and Glastopf (emulating website services) honeypots. The most important result consists in the fact that the differentiation among honeypots according to their IP address is relatively rough (usually two categories, i.e. academic and commercial networks, are usually distinguished, but the type of services in commercial sites is taken into account, too). Comparisons of results to other similar honeynets confirms the validity of the paper main conclusions.

A multilateral trust between a user and a system is considered. First of all we concentrate on user trust levels associated with the context-oriented CoRBAC model. Consequently, there were computed user profiles on the basis of its implementation in the information processing system “My GUT”. Furthermore, analysis of these profiles and the impact of user trust levels on system security and usability have been discussed.

Honeypots and honeynets are common tools for network security and network forensics. The deployment and usage of these tools is affected by a number of the technical and legal issues. It is very important to consider both issues together. Therefore, paper outlines technical aspects of honeynet and discusses the liability of honeypot’s and honeynet’s administrator. Paper deals with civil and criminal liability. Also paper focuses on cybercrime and liability of attackers.

Balancing security against performance for IT systems is one of the most important issues to be solved. The quality of protection of systems can be achieved on different levels. One can choose factors which have a different impact on the overall system security. Traditionally, security engineers configure IT systems with the strongest possible security mechanisms. Unfortunately, the strongest protection can lead to unreasoned increase of the system load and finally influence system availability. In such a situation the quality of protection models which scale the protection level depending on the specific requirements can be used. In the article, we present the approach which enables balancing security against performance for database systems. The analysis is performed by Automated Quality of Protection Analysis (AQoPA) tool which allow automatic evaluation of system models which are created in the Quality of Protection Modelling Language (QoP-ML).

The paper contains the formulation of the problem of situational safety management of distributed computer networks, mathematical models and algorithms, which can help to identify quickly and eliminate the causes of security breaches of these networks during their operation. The developed software allows to create a methodological basis for the creation of new information security management systems with cognitive functions, the use of which will improve the reliability and efficiency of the enterprises and organizations.

A technique for botnet detection based on a DNS-traffic is developed. Botnets detection based on the property of bots group activity in the DNS-traffic, which appears in a small period of time in the group DNS-queries of hosts during trying to access the C&C-servers, migrations, running commands or downloading the updates of the malware. The method takes into account abnormal behaviors of the hosts’ group, which are similar to botnets: hosts’ group does not honor DNS TTL, carry out the DNS-queries to non-local DNS-servers. Method monitors large number of empty DNS-responses with NXDOMAIN error code. Proposed technique is able to detect botnet with high efficiency.

Distributed control systems are often used in many branches of industry and frequently replace standalone controllers. However, their operation is more complex and include aspects of communication between various devices. To operate correctly, it is crucial to ensure that timeliness of communication is satisfied. In this paper, the approach to modeling of the token passing paradigm, as well as the multi-master communication with token exchange has been presented. The proposed models are based on a few kinds of SysML diagrams, namely Block Definition, Internal Block, State Machine, and Sequence Diagrams. The paper presents a set of dedicating modeling rules together with their detailed explanation.

The paper presents the theoretical and empirical tests which examine the algorithms of automatic scenario selection of cyclic exchange in transmission via two buses. The implemented method is the base for using the second (redundant) bus for data transmission. The main aim of this method is to build two-stream data transmission so that each data stream has similar, or even the same transmission time. The flexibility of the proposed solution results from the fact that it is useful not only for two-bus transmission, but also when the transmission is via only one bus. This method allows estimating the time of the network cycle and also indicates the values of the times for acyclic transmissions. The presented method is only the introduction to designing, developing and controlling the two-stream transmission.

Contemporary industrial and production systems produce huge amounts of data in various models, used for process monitoring, predictive maintenance of the machines, historical analysis and statistics, and more. Apache Hadoop brings a cost-effective opportunity for Big Data analysis, including the data generated in various industries. Integrating Hadoop into industrial environments creates new possibilities, as well as many challenges. The authors of this paper are involved into commercial and scientific projects utilizing Hadoop for industry as predictive analytics platform. In such initiatives the lack of standardization of monitoring of the industrial process in terms of Hadoop cluster utilization is especially perplexing. In this paper, authors propose the methodology of monitoring Hadoop in industrial environments, based on dedicated ontology and widely adopted standards: OPC Unified Architecture and RESTful Web Services.

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.

This paper describes analysis of a real-world case where growing up Platform as a Service (PaaS) provider faced and solved problems with scaling cloud-based infrastructure. Scientific, Petri net-based method was used to asses decisions taken by Heroku – the PaaS manager – and legitimacy of claims of the PaaS clients caused by not satisfying efficiency of the new solutions. Exhaustive information provided in the Internet by both parties of the conflict were used in order to create model of the application and infrastructure corresponding to the real case. The model was then used to perform reliable simulations and show that while the client’s claims were well founded, but growth of the PaaS infrastructure forced and justified changes in the management algorithms.

The article describes results of research in the area of software-defined networking applied to multi-cloud infrastructures. We present an innovative load balancer designed in accordance with a unique set of requirements, stemming from taking the cloud tenant’s perspective on cloud service usage effectiveness. We developed a proof of concept implementation of the load balancer and drawn the conclusions regarding the system itself and its applicability to production environments. Our solution is more comprehensive than others available in contemporary SDN environments, and offers more functionality than just service chaining realized through flow reroutes.

For the scalability extension (SVC) of the H.264/AVC coding algorithm, video data belonging to different layers have different priority depending on their importance to the quality of the video and the decoding process. This creates new challenges but also new opportunities for video streaming systems, especially those that use the Internet infrastructure. Therefore, mechanisms of the traffic engineering used in the IP network should as far as possible take into account internal distribution of priorities inside video streams. The paper proposes the Weighted Priority Pre-marking (WPP) algorithm for color-aware SVC video streaming over a DiffServ network. This solution takes into account relative importance of Network Abstraction Layer Units (NALU) and does not require any changes in the DiffServ marker algorithm. Presented results of simulations show that the color-aware transmission system, based on the WPP packet pre-marking, can provide better perceived video quality than the standard (best effort) streaming of multi-layered SVC video as well as single-layered AVC video.

Processing video information is a complex problem requiring a variety of methods to solve. Local computer resources are generally used to process home video. Local computer networks resources are more suitable with larger size video files. Such an approach is expedient in processing an average volume of information and fulfills the medium requirements of processing time. This method has several disadvantages which can be eliminated with use of GRID-systems. However, GRID-calculations also demand the increasing of the efficiency by finding the number of tasks in a job and parameters of processing environment. An outline is given in this paper comparing methods of video processing based upon Grid calculations and cloud computing.

In convergent access networks, access nodes (composed by optical and wireless components) can serve Base Stations (BSs). In the access node, the wireless receivers are one of the components with high energy consumption level. We can save energy by implementing transmission techniques to receive data in some periods of times, and switching off the wireless receivers the rest of time. However, this strategy will result in increased access delay for packets waiting in the BSs until transmission to the access node. In this paper, we aim at developing a framework for modeling the reception of data by the wireless receivers at the access nodes. We present an algorithm, a numerical formulation and an asymptotic approximation. Comparison between simulation and numerical results is presented and the analysis of results is summarized.

We evaluate the effect of automatic updates to the LTE Neighbor Relation Table (NRT) to optimize the efficiency of handovers in the network. We propose an algorithm which monitors the handovers within the LTE network and detects events when the procedure was started to early (handover to early) or too late (handover too late). Basing on the observed statistics in Operation and Maintenance server (OAM) the proposed algorithm updates the NRT to maximize the network throughput and minimize the probability of loss of connectivity. Performed simulations show that correctly forbidden handovers can reduce number of unnecessary signaling, reduce risk of fail during handover and reduce the time required to perform the handovers.

The scope of this paper is to check influence of the size of the speech corpus on the speaker recognition performance. Obtained results for TIMIT corpus are compared with results obtained for smaller database ROBOT. Additionally influence of feature dimensionality and size of the speaker model was tested. Achieved results show that the best results can be obtained for MFCC features. The lowest

EER

for larger TIMIT database are 4 times worse than the best result for ROBOT corpus which confirms that biometric systems should be tested on as large data sets as possible to assure that achieved error rates are statistically significant.

Wireless propagation environment utilised by broadband transmission systems usually has a sparse nature, i.e. only several isolated propagation paths are essential for information transfer. Receiver can recover the parameters of the particular paths using greedy, iterative algorithms that belong to the family of compressed sensing techniques. How to stop the iterative procedure, if no precise knowledge about the order of the channel sparsity is available in the receiver a priori, is a key question regarding a practical implementation of the method. The paper provides stopping criteria analysis of the Orthogonal Matching Pursuit (OMP) algorithm that is used as the core of the channel impulse response estimation method for Time-Domain Synchronous OFDM transmission system. There are investigated the residual error and the difference of successive residual errors of the OMP algorithm, as the possible metrics applied to stop the iteration procedure. Finally, a new heuristic stopping rule based on these two errors is proposed and numerically examined.

Computer networks are one of the major slices of the global energy consumption. Since 2009, a couple of standards have been developed for energy conservation in passive optical networks. These standards and most of the researches are based on improvements on Optical Network Unit side. In this study, a novel energy efficiency algorithm, which is based on coupling two Optical Line Terminal to reduce energy consumption in central office, is proposed. Our design employs optical switches and amplifiers to create a switch-box, which is under control of both Optical Line Terminal pairs.

The phenomenon of quantum entanglement is an important component of many protocols in the field of quantum computing. In this paper a level of entanglement in short qudit chains is evaluated. The evaluation is carried out with use of the CCNR criterion and the concurrence measure. There are also some explicit formulae describing the values of CCNR criterion and concurrence for exemplary short spin chains. Mentioned formulae may be used as an invariant for a protocol which describes a transfer of information in XY-like spin chains.

In the paper a multicast architecture for programmable networks based on separation of group management and network control tasks is proposed. Thanks to this separation, services which want to make use of multicast communications no longer have to implement low-level network functionalities and their operation is greatly simplified. Abstracting service’s view of the network into a fully connected cloud enables us to transparently implement within the network complex features such as host mobility without any disturbances to the operation of the service or it’s clients. Aspects of the multicast architecture such as signaling, addressing, form of data delivery, security and cooperation with legacy applications are discussed; precise requirements for interfaces specified.

Number of mobile devices and applications is growing rapidly in recent years. Capabilities and performance of these devices can be tremendously extended with the integration of cloud computing. However, multiple challenges regarding implementation of these type of mobile applications are known, like differences in architecture, optimization and operating system support. This paper summarizes issues with mobile cloud computing and analyzes existing solutions in this field. A new innovative approach consisting of an application model and a mobile offloading framework are considered and adopted for practical applications.

A multi-server queueing system with an infinite buffer operating in the Markovian random environment is analyzed. Under the fixed state of the random environment, the arrival flow is described by the Markovian arrival process. Customers in the buffer may be impatient and leave the system. The number of available servers, the arrival process, the rate of customers’ service, and the impatience intensity depend on the state of the random environment. Behavior of the system is described by the multi-dimensional asymptotically quasi-Toeplitz Markov chain. The ergodicity condition is derived. Algorithm for computation of the stationary distribution is provided. The main performance measures of the system are calculated.

The paper, Maciej presents a new model of Erlang’s Ideal Grading with multi-service traffic sources. Traffic sources of this kind correspond better with actual devices utilized by network users. As the proposed model is an approximate one, the results obtained with its help have been compared to the results of a computer simulation. The comparison has confirmed the correctness of the model’s theoretical assumptions. The new model may successfully be used for modeling modern communication systems.

In this paper, we define a class of asymptotically work – conserving service disciplines which, in particular, are used in some existing wireless systems. Then we describe a class of state-dependent queueing systems with focus on the workload-dependent and queue- dependent systems with asymptotically work-conserving discipline. We outline the regenerative approach to stability analysis of state- dependent systems and formulate corresponding sufficient stability conditions. In the last section these conditions are verified by simulation a system with workload-dependent and queue-dependent input rates and a wireless system with Markov-modulated transmission rates.

In this paper, we analyze a queueing system with so called “cold” redundancy. The system consists of an infinite buffer, the main unreliable server (server 1) and the absolutely reliable reserve server (server 2). The input flow is a

BMAP

(Batch Markovian Arrival Process). Breakdowns arrive to the server 1 according to a

MAP

(Markovian Arrival Process). If the server 1 is fault-free, it serves a customer, if any. After breakdown occurrence the server 1 fails and the repair period starts immediately. The customer, whose service is interrupted by a breakdown, goes to the second server, where its service is restarted. When the repair period ends, the customer whose service on the server 2 has not yet completed goes back to the server 1 and its service begins anew. We assume that the switching from one server to another takes time. Switching times as well as service times and repair time have

PH

(Phase type) distribution. The queue under consideration can be applied for modeling of a hybrid communication system consisting of the FSO – Free Space Optics channel (server 1) and the radio-wave channel (server 2). We derive a condition for stable operation of the system, calculate its stationary distribution and base performance measures and derive an expression for the Laplace-Stieltjes transform of the sojourn time distribution.

We investigate processor sharing systems with demands having some random space requirements (volumes) and demands lengths depending on their volumes. For such systems we determine non-stationary total demands capacity distribution in terms of Laplace and Laplace-Stieltjes transforms.

The authors present a new approach to find stochastic bounds for a Markov chain – namely with the use of the GPU for computing the bounds. A known algorithm [

1

,

2

] is used and it is rewritten to suit the GPU architecture with the cooperation of the CPU. The authors do some experiments with matrices from various models as well as some random matrices. The tests are analyzed and some future considerations are given.

The paper defines a method to extract a part of a graph, which corresponding to a network (Internet) topology covering a certain area. The method refers to data gathered within Opte Project and describing the Internet topology. The extracted parts of graphs have been analyzed in terms of the number of neighbours, longest path length and existence of cycles. Then the resulting topologies are used to model transient behaviour of wide area networks with the use of fluid-flow approximation.

The newest versions of the most ubiquitous USB interface are 3.0/3.1. USB 3.0 improves utilization of bus throughput through the changing of system architecture and introduction a new scheduling algorithm. In some applications using USB ports (e.g. vision system), the quality of service (QoS) is required. This paper presents proposal of model of USB system based on theory of scheduling and definition of QoS for communication interfaces (in particular USB 3.0). Some experiments (using prepared USB 3.0 application) were performed to verify compliance with QoS and their results are shown in this article.

The concern about the data security is still growing. The applications we use still lack basic security. The security can be obtained easily without heightening the awareness of the users by simply implementing methods that allow end-to-end security. In the paper we present the new method of public key distribution using Social Media Networks. We use groups or circles that users can create there and publish images in the gallery with proper public or private access. After establishing secure session user can switch into different medium of communication, e.g. different messenger or text-SMS to share their data in secure way. Then there are described several applications of the presented method Secure Data Exchange (

SDEx

), e.g. secure SMS texting, Social Media Networking and data sharing. We propose frame format of the published QRcode as a public key and data message frame format.

Creating of speaker recognition application requires advanced speech processing techniques realized by specialized speech processing software. It is very possible to improve the speaker recognition research by using speech processing platform based on open source software. The article presents the example of using open source speech processing software to perform speaker verification experiments designed to test various speaker recognition models based on different scenarios. Speaker verification efficiency was evaluated for each scenario using TIMIT speech corpus distributed by Linguistic Data Consortium. The experiment results allowed to compare and select the best scenario to build speaker model for speaker verification application.

Paper presents Savitzky-Golay digital filter implementation using the R software environment in private Hadoop Data Mining System. The idea of the research was the implementation of the digital filter to handle big data came from measurements. The study focused on proper filter implementation and comparison of the results with National Instruments Diadem – enterprise class system for data management and mining.

Online social networks (OSNs) can be considered as huge success. However, this success costs users their privacy and loosing ownership of their own data; Sometimes the operators of social networking sites, have some business incentives adverse to users’ expectations of privacy. These sort of privacy breaches have inspired research toward privacy-preserving alternatives for social networking in a decentralized fashion. Yet almost all alternatives lack proper feasibility and efficiency, which is because of a huge mismatch between aforementioned goal and today’s network’s means of achieving it. Current Internet architecture is showing signs of age. Among a variety of proposed directions for a new Internet architecture is Named Data Networking (NDN), focused on retrieving content by name, which names packets rather than end-hosts. NDN characteristics greatly facilitate development of applications tailored for today’s needs. In this paper a decentralized architecture for social networking is proposed that provides strong privacy guarantees while preserving the main functionalities of OSNs, in a content-based paradigm. The simulation results show that not only it is feasible to have decentralized social networking over content centric networks, but also it is significantly more efficient from a global network point of view.

The USB device designers, as well as the advanced users often need a software utility to capture and analyse USB data exchanged between a device and an application in host. The software USB analysers are designed for this purpose.

Software USB analyser solution designed for Windows XP, Vista, 7 and 8 is presented in the paper. The solution consists of USBPcap responsible for USB data capture and Wireshark responsible for data analysis. The solution was used to record and analyse messages exchanged between DigiTech RP250 guitar multi-effect processor and host-side software package X-Edit supplied by vendor (DigiTech). For this purpose Wireshark was extended with two dissectors: USB Audio and X MIDI SysEx.

The main objective of the paper was to test whether the devices compatible with Bluetooth Low Energy are reliable for indoor localization system. To determine the reliability of this technology several tests were performed to check if measured distance between Bluetooth transmitter and mobile device is close to the real value. Distance measurement focused on Bluetooth technology based mainly on received signal strength indicator (RSSI), which is used to calculate the distance between a transmitter and a receiver. As the research results show, the Bluetooth LE signal power cannot be the only reliable source of information for precise indoor localization.