Sven Ubik

Evaluating Application-Layer Classification Using a Machine Learning Technique over Different High Speed Networks

Application–layer classification is needed in many monitoring applications. Classification based on machine learning offers an alternative method to methods based on port or payload based techniques. It is based on statistical features computed from network flows. Several works investigated the efficiency of machine learning techniques and found algorithms suitable for network classification. A classifier based on machine learning is built by learning from a training data set that consists of data from known application traces. In this paper, we evaluate the efficiency of application-layer classification based on C4.5~machine learning algorithm used for classification network flows from different high speed networks, such as 100~Mbit, 1~Gbit and 10~Gbit networks. We find a significant decrease in the classification efficiency when classifier built for one network is used to classify other network. We recommend to build classifier from data collected from all available networks for best results. However, if different networks are not available, good results can be obtained from data traces to the commodity Internet.

Real-time long-distance transfer of uncompressed 4K video for remote collaboration

Better-than-high-definition-resolution video content (such as 4K) is already being used in some areas, such as scientific visualization and film post-production. Effective collaboration in these areas requires real-time transfers of such video content. Two of the main technical issues are high-data volume and time synchronization when transferring over an asynchronous network such as the current Internet. In this article, we discuss design options for a real-time long-distance uncompressed 4K video transfer system. We present our practical experience with such transfers and show how they can be used to increase productivity in film post-production, as an application example.

Realistic Passive Packet Loss Measurement for High-Speed Networks

Realistic and accurate packet loss measurement of production traffic has been challenging, since the frequently-used active monitoring approaches using probe packets cannot capture the packet loss experienced by the traffic of individual user applications. In this paper, we present a new approach for the accurate measurement of the packet loss rate faced by actual production traffic based on passive network monitoring. In contrast to previous work, our method is able to pinpoint the packet loss rate experienced by the individual traffic flows of concurrently running applications. Experimental results suggest that our approach measures packet loss with 100% accuracy for network speeds as high as 12 Gbit/s, while traditional ICMP-based approaches were usually much less accurate. We also report experiences from a real-world deployment of our method in several 10 Gbit/s links of European research networks, where it has been successfully operational for several months.

ABW--Short-Timescale Passive Bandwidth Monitoring

Bandwidth usage monitoring is important for network troubleshooting and planning. Traditionally, used bandwidth is computed from router interface byte counters read by SNMP. This method only allows to check long-term averages of the total used bandwidth without information about short-term dynamics and without knowledge of what applications are consuming most bandwidth. We describe the architecture of a novel passive bandwidth usage monitoring application. This application uses packet capture and advanced processing to continuously provide real-time information about bandwidth usage. The produced characteristics include information about short- term peaks and about the percentage of bandwidth used by different protocols in different layers of the OSI model hierarchy, including detection of application protocols that use dynamic ports.

Real-time stereoscopic streaming of robotic surgeries

Modern robotic surgery using stereoscopic video views of surgical elements can also enable novel E-health applications, such as remote medical students training or presentations of surgical procedures on symposia. The requirement is that such video signals are transmitted remotely in high quality and low latency. In this paper we describe our experience with real-time long distance stereoscopic transmissions of robotic surgeries using a system for low latency streaming over packet networks that we have developed.

Real-time anonymization in passive network monitoring

Passive network monitoring that observes user traffic has many advantages over active monitoring that uses test packets. It can provide characteristics of real user traffic, that cannot be detected actively. However, when processing user traffic, we must guarantee user privacy. This is a task of packet header anonymization that removes sensitive information, while keeping as much as possible of the original traffic properties. In this paper we present design and implementation of an FPGA-based packet header anonymization that unlike previous approaches operates in real time and prevents sensitive information from getting to the monitoring PC and beyond.

MTPP - Modular Traffic Processing Platform

High-speed (10 Gb/s and above) network monitoring and traffic processing requires hardware acceleration. Different applications require different functions to be placed in hardware. Current packet capture cards include fixed firmware, which is difficult to extend. In this paper we propose an architecture for Modular Traffic Processing Platform (MTPP), which enables end users to easily modify hardware processing without any FPGA development. On the other hand, developers can create new processing modules with much reduced effort thanks to simple module interfaces and isolation of module time constraints.

Cyber performances, technical and artistic collaboration across continents

Our objective was to verify whether modern computer network and audiovisual technologies can enable collaborative work between performing artists when they are distributed across large distances and what the requirements and limitations are. Such distributed collaborative environments will bring new opportunities both for the artists and the audience.We describe our experiences gained from laboratory experiments and during cyber performances at APAN meetings that took place in Korea and Taiwan. We conducted several experiments of playing music together over distance.The maximum acceptable delay is between 15 and 30 ms.Order of importance for performers is good sound, feeling other musicians and good vision.Good sound means similar acoustic space characteristics of the local and remote sound.With higher delay, distributed music and dance can be successfully contributed to a common performance.

LZ4 compression algorithm on FPGA

This paper describes analysis and implementation of a LZ4 compression algorithm. LZ4 is derived from a standard LZ77 compression algorithm and is focused on the compression and decompression speed. The LZ4 lossless compression algorithm was analyzed regarding its suitability for hardware implementation. The first step of this research is based on software implementation of LZ4 with regard to the future hardware implementation. As a second step, a simple hardware implementation of LZ4 is evaluated for bottlenecks in the original LZ4 code. Xilinx Virtex-6 and 7-Series FPGAs are used to obtain experimental results. These results are compared to the industry competitor.

Remote Access to 3D Models for Research, Engineering, and Art

Stereoscopic (3D) models and visualizations can add value to current videoconferencing systems by offering a higher level of immersion when accessing remote 3D models. The low-latency remote access to 3D models in high resolution over high-speed networks could enable more effective collaboration between physically distributed teams in research, engineering, and humanities. Many research fields, from engineering and science to the arts and humanities, could benefit from remote access to 3D models. This article surveys the state of the art in this field as well as the various system architectures available. The authors also describe several experiments conducted over long distances (more than 10,000 km) using prototype systems developed at the Czech Technical University. Their results show that when the low processing delay on the sender and receiver side is added to the inevitable network propagation delay the resulting response time is still acceptable and can provide an interactive feeling.