Telfor Journal Vol.9 No.1 (2017)

Content

Speeding Up IP Lookup Procedure in Software Routers by Means of Parallelization

M. Vesović, A. Smiljanić, and M. Tomašević
Topic:
Telecommunications Networks
Abstract
Performance of software routers is limited by the speed of the operating system network protocol stack. A faster network protocol stack can be implemented in user space by utilizing parallelization and different optimization techniques. In this work we demonstrate an efficient implementation of the IP lookup algorithm in user space with multibit trie structures. Afterwards, we demonstrate the improvements achieved through parallelization. We evaluate pthreads and OpenMP parallelization methods and compare their performance.
Keywords
10 GbE, high speed packet I/O, IP lookup, parallelization, software routers.
Full Text
HyperLink Speeding Up IP Lookup Procedure in Software Routers by Means of Parallelization
Page(s)
2-7
Doi
10.5937/telfor1701002V

4K Video Traffic Prediction using Seasonal Autoregressive Modeling

D. R. Marković, A. M. Gavrovska, and I. S. Reljin
Topic:
Telecommunications Networks
Abstract
From the perspective of average viewer, high definition video streams such as HD (High Definition) and UHD (Ultra HD) are increasing their internet presence year over year. This is not surprising, having in mind expansion of HD streaming services, such as YouTube, Netflix etc. Therefore, high definition video streams are starting to challenge network resource allocation with their bandwidth requirements and statistical characteristics. Need for analysis and modeling of this demanding video traffic has essential importance for better quality of service and experience support. In this paper we use an easy-to-apply statistical model for prediction of 4K video traffic. Namely, seasonal autoregressive modeling is applied in prediction of 4K video traffic, encoded with HEVC (High Efficiency Video Coding). Analysis and modeling were performed within R programming environment using over 17.000 high definition video frames. It is shown that the proposed methodology provides good accuracy in high definition video traffic modeling.
Keywords
4K Video traffic, HEVC, autoregressive model, seasonal technique, video tracing, prediction.
Full Text
HyperLink 4K Video Traffic Prediction using Seasonal Autoregressive Modeling
Page(s)
8-13
Doi
10.5937/telfor1701008M

Birkhoff-von Neumann Switch with Deflection Based Load Balancing

S. Durković and Z. Čiča
Topic:
Telecommunications Networks
Abstract
Load balanced Birkhoff-von Neumann (LB-BvN) packet switches have low hardware complexity while achieving high performance. We propose a novel LB-BvN based switch that achieves 100% throughput for any admissible traffic scenario. The proposed switch uses the deflection mechanism to decrease overall hardware complexity of the switch. The delay and buffer bounds of the proposed switch are derived and analyzed using the network calculus theory. The proposed switch is compared to other LB-BvN based solutions.
Keywords
Birkhoff-von Neumann switches, load balancing, network calculus, packet switching.
Full Text
HyperLink Birkhoff-von Neumann Switch with Deflection Based Load Balancing
Page(s)
14-19
Doi
10.5937/telfor1701014D

A Cognitive Radio based Solution to Coexistence of FH and OFDM Signals Implemented on USRP N210 Platform

M. Janjić and M. Erić
Topic:
Radio Communications
Abstract
A new concept development and practical implementation of an OFDM based secondary cognitive link are presented in this paper. Coexistence of a secondary user employing Orthogonal Frequency Division Multiplexing (OFDM) and a primary user employing Frequency Hopping (FH) is achieved. Secondary and primary links are realized using Universal Software Radio Peripheral (USRP) N210 platforms. Cognitive features of spectrum sensing and changing transmission parameters are implemented. Some experimental results are presented.
Keywords
Cognitive radio, FH, OFDM, secondary link, spectrum sensing, USRP platform.
Full Text
HyperLink A Cognitive Radio based Solution to Coexistence of FH and OFDM Signals Implemented on USRP N210 Platform
Page(s)
20-25
Doi
10.5937/telfor1701020J

Model for Detection and Classification of DDoS Traffic Based on Artificial Neural Network

D. Peraković, M. Periša, I. Cvitić, and S. Husnjak
Topic:
Communication Systems
Abstract
Detection of DDoS (Distributed Denial of Service) traffic is of great importance for the availability protection of services and other information and communication resources. The research presented in this paper shows the application of artificial neural networks in the development of detection and classification model for three types of DDoS attacks and legitimate network traffic. Simulation results of developed model showed accuracy of 95.6% in classification of pre-defined classes of traffic.
Keywords
ANN, DDoS, network traffic, network security.
Full Text
HyperLink Model for Detection and Classification of DDoS Traffic Based on Artificial Neural Network
Page(s)
26-31
Doi
10.5937/telfor1701026P

A Review of Serbian Parametric Speech Synthesis Based on Deep Neural Networks

T. Delić, M. Sečujski, and S. Suzić
Topic:
Signal Processing
Abstract
In this paper the research related to the development of a deep neural network based speech synthesizer for the Serbian language, trained on recorded utterances of a single female voice talent, is described. Two separate networks are used for prediction of acoustic features and phonetic segment durations. Through a set of experiments the optimal values of the hyper-parameters of the neural networks are established, and then the influence of the amount of training data on the quality of synthesized speech is examined. The quality is evaluated through objective measures as well as appropriate listening tests. It has been confirmed that 4-layer deep neural networks with 512 units per hidden layer, trained on 3 hours of data, produce speech of very good quality. The results also suggest that a further increase in the amount of training data may contribute to further improvement in quality.
Keywords
DNN, HMM, speech synthesis.
Full Text
HyperLink A Review of Serbian Parametric Speech Synthesis Based on Deep Neural Networks
Page(s)
32-37
Doi
10.5937/telfor1701032D

Performance of 2D SOVA Along and Across Track in Shingled Magnetic Recording Media

M. B. Abdulrazaq, M. Z. Ahmed, and P. Davey
Topic:
Signal Processing
Abstract
Serial Concatenation of Two Dimensional Soft Output Viterbi Algorithm (2D-SOVA) and regular Viterbi Algorithm (VA) for 2D equalisation and detection of Shingled Magnetic Recording (SMR) media provides excellent performance as compared to the use of 1 Dimensional (1D) maximum likelihood detector. In this paper, we implement and evaluate the performances of two versions of it. The first version performs 2D SOVA along the tracks to eliminate the effect of inter-symbol interference (ISI) and then the Viterbi detector across the tracks to remove inter-track interference (ITI). The second version carries out 2D-SOVA across the tracks and VA along the tracks. The results for high ITI and ISI show a better performance when using 2D-SOVA across the track with a small difference in computational complexity in favour of 2D-SOVA across the tracks.
Keywords
2D detection, Maximum Likelihood Detection, Shingled Magnetic Recording, SMR, Soft Output Viterbi Algorithm, SOVA.
Full Text
HyperLink Performance of 2D SOVA Along and Across Track in Shingled Magnetic Recording Media
Page(s)
38-42
Doi
10.5937/telfor1701038A

Thresholding Approach to Radiography Image Processing Acceleration

V. S. Ostojić, Đ. S. Starčević, and V. S. Petrović
Topic:
Signal Processing
Abstract
An approach for detecting low‑value Laplacian pyramid coefficients which can be used to accelerate radiography image processing is presented.The acceleration is achieved through modifying Laplacian pyramid construction algorithm. Detection of low-value coefficients in lower pyramid layers is based on corresponding low-value higher layer coefficients which are lower than the threshold. Detected coefficients are then omitted in pyramid calculation and by the radiography image enhancement algorithm. Clinical radiography images are used to evaluate the proposed approach. Structural similarity index (SSIM) is employed to measure similarity between images processed with and without proposed acceleration. Threshold value analysis shows that a high image quality with a mean SSIM of 0.995 can be achieved with mean processing time reduction of 4.62 %.
Keywords
digital radiography, image processing, Laplacian pyramid, multi-scale methods.
Full Text
HyperLink Thresholding Approach to Radiography Image Processing Acceleration
Page(s)
43-48
Doi
10.5937/telfor1701043O

Simple and Precise Commercial Camera based Eye Tracking Methodology

V. Bobić and S. Graovac
Topic:
Signal Processing
Abstract
In this paper, an eye tracking methodology is presented. It represents a commercial camera based system for detection of eye pupil and its trajectory and classification of eye movements that led to the formation of assessed eye path. Methodology was examined on 10 subjects, who performed specified series of eye movements in two recording conditions: with and without simultaneous head motions. Image segmentation was performed and coordinates of eye pupil were found for each consecutive video frame. Modified velocity threshold method for eye movement classification was performed. Classification accuracy was evaluated with two parameters: Sensitivity (S) and Positive predictive value (PP), and compared for both recording conditions. In the absence of head motions, a high accuracy result was obtained (S=96.25%; PP=100%). A less accurate result was obtained in the second case (S=87.50%; PP=88.61%). However, it was shown that proposed methodology can provide good results for commercial eye tracking regardless of the recording conditions.
Keywords
eye tracking, image gradient, pupil detection, velocity threshold.
Full Text
HyperLink Simple and Precise Commercial Camera based Eye Tracking Methodology
Page(s)
49-54
Doi
10.5937/telfor1701049B

The Excess Attenuation of Electric Field in the Presence of Deciduous Trees

J. G. Perovic and D. I. Olcan
Topic:
Applied Electromagnetics
Abstract
Excess attenuation of electric field of an electromagnetic wave propagating through deciduous forest is approximated with two novel analytical formulas. The coefficients of those formulas were obtained through a computationally intensive optimization process by fitting the simulation results obtained from stochastic 3D and 2D models of forests. The electric field was either vertically or horizontally polarized, at frequencies 30 MHz, 60 MHz and 100 MHz. According to numerical experiments, these formulas predicted the excess attenuation within several decibels, up to at least 100 MHz. Nevertheless, the proposed formulas matched very well the measured data from literature below 1 GHz, and neatly followed the measurements even up to 10 GHz. Therefore, the frequency range considered in the numerical experiments can be extrapolated at least tenfold.
Keywords
Attenuation, approximation formula, electromagnetic waves, modeling, full‑wave electromagnetic simulation, trees.
Full Text
HyperLink The Excess Attenuation of Electric Field in the Presence of Deciduous Trees
Page(s)
55-60
Doi
10.5937/telfor1701055P

Comparison Elements on STG DICE cell for Content-Addressable Memory and Simulation of Single-Event Transients

V. Ya. Stenin and A. V. Antonyuk
Topic:
Applied Electronics
Abstract
Comparison elements on base the STG DICE cell and the logical element “Exclusive OR” for a content-addressable memory were designed and simulated. The comparison element contains two identical joint groups of transistors that are spaced on the chip by the distance of four micrometers, so the loss of data in STG DICE cell practically excluded. On the characteristics of the new 65-nm CMOS comparison element, we predict the hardness of these item to single event rate (SER) more to hundred times compared to elements on 6-transistors cells and the standard DICE cell with distances 0.5-0.6 μm between mutually sensitive nodes.
Keywords
content-addressable memory, DICE, logical element, memory cell, single nuclear particle, topology.
Full Text
HyperLink Comparison Elements on STG DICE cell for Content-Addressable Memory and Simulation of Single-Event Transients
Page(s)
61-66
Doi
10.5937/telfor1701061S

Run-Time Reconfigurable Middleware in Device Network Architecture

M. Milovanović, I. T. Popović, and A. Ž. Rakić
Topic:
Applied Electronics
Abstract
In this paper we introduce a framework for implementation of automatically reconfigurable middleware layer for the system represented as a network of devices. The middleware enables communication between passive heterogeneous components of such a distributed system by introducing dedicated service agents. Service agents establish run-time configurable data paths and they enable quality of service based reconfiguration and redundancy in data transport. The case study is presented to demonstrate utilization of quality of service measurements for automatic middleware data path reconfiguration.
Keywords
automatically reconfigurable middleware, device network protocol, quality of service.
Full Text
HyperLink Run-Time Reconfigurable Middleware in Device Network Architecture
Page(s)
67-72
Doi
10.5937/telfor1701067M