direkt zum Inhalt springen

direkt zum Hauptnavigationsmenü

Sie sind hier

TU Berlin

Inhalt des Dokuments

Es gibt keine deutsche Übersetzung dieser Webseite.

Past talks, student talks, and other events of 2012

On the Impact of Backpressure Algorithms on Packet Delay
Speaker:
Jinliang Xue (TU Berlin master studentl)
Type:
Master's thesis introductory talk
Time:
10 August 2012
13:15–13:45
Place:
Sputnik (TEL 18)
Abstract:
Backpressure algorithms represent an efficient class of maximum throughput routing algorithms which also ensure network stability. The main drawback of backpressure algorithms, however, is that they can induce very large packet delays. This can be especially detrimental for routing delay sensitive data streams.

In this thesis we propose to study the impact of backpressure algorithms on packet delay for several network topologies. We identify that looping and large queueing delays in single buffers are the key factors for the large packet delays. We use these insights to extend backpressure algorithms in order to reduce the delay and also to improve fairness in multi-stream scenarios.
Asynchrony from Synchrony
Speaker:
Eli Gafni (UCLA)
Type:
Talk
Time:
26 July 2012
16:00–17:00
Place:
Sputnik (TEL 17)
Abstract:
In the past adversaries that reduce the computational power of a system as they become more powerful, were investigated in the context of asynchronous shared-memory. The adversary's power was characterized by the set of processors it can crash. Here we investigate more refined adversaries: Adversaries that purge messages in a synchronous message-passing system. In particular, we consider a synchronous message passing system over a complete network where in each round all processors send to all. The computation power is modulated by an adversary that in each round, independently of previous rounds, can delete a subset of the messages. Deleting a message on a link in one direction does not necessarily imply the removal of the message in the other direction. The weakest adversary has no power, it may delete no message, while an adversary that can remove any set of messages is the strongest. Clearly the model power to solve tasks is inverse proportional to the adversary power. In the former model all tasks are solvable, while in the latter no none-trivial task is solvable.

We characterize the set of adversaries that are weak enough allowing the model to solve all read-write wait-free solvable tasks, and at the same time strong enough to preclude solution of any task which is not read-write wait-free solvable. A remarkable side-benefit of this characterization is a simple, as simple as can be, derivation of the Herlihy Shavit condition that equates wait-free solvability with a subdivided-simplex. We show how each step in the computation inductively takes a subdivided-simplex and further subdivides it in the simplest way possible, making the characterization of read-write wait-free widely accessible.
Real-Time Video Streaming in Multi-hop Wireless Static Ad Hoc Networks
Speaker:
Guy Even (School of Electrical Engineering, Tel-Aviv University)
Type:
Talk
Time:
17 July 2012
17:00–17:30
Place:
TEL 1118/19
Abstract:
We deal with the problem of streaming multiple video streams between pairs of nodes in a multi-hop wireless ad hoc network. The nodes are positioned on a plane, know their locations, and are synchronized (via GPS). We consider a 802.11g WiFi network in which NICs can hop between non-overlapping frequency channels in a synchronous fashion. We design a centralized algorithm that computes a frequency assignment and a schedule whose goal is to maximize the minimum throughput over all the video streams. In addition, we developed a localized flow-control mechanism to stabilize the queue lengths.

We simulated traffic scheduled by the algorithm using OMNET++/MixiM (i.e., physical SINR interference model with 802.11g) to test whether the computed throughput is achieved.

Joint work with Yaniv Fais, Moti Medina, Moni Shahar, and Alexander Zadorojniy.
Bio:
Guy Even joined the Faculty of Engineering in Tel-Aviv University in 1997. He works on algorithms in general, including: decoding error correcting codes, online algorithms, approximation algorithms, scheduling, frequency assignment in wireless networks, routing, and algorithms for finding optimal policies in MDPs. He also worked on theory of VLSI and computer arithmetic.
Hitting Sets Online and Unique-Max Coloring
Speaker:
Guy Even (School of Electrical Engineering, Tel-Aviv University)
Type:
Talk
Time:
16 July 2012
15:00–15:30
Place:
TEL 1118/!9
Abstract:
We consider the problem of hitting sets online. The range-space (consisting of points and ranges) is known in advance, and the ranges to be stabbed are input one-by-one in an online fashion. The online algorithm must stab each range upon arrival. An online algorithm may add points to the hitting set but may not remove already chosen points. The goal is to use the smallest number of points. Alon et al. presented in 2009 an online algorithms for general range spaces, the competitive ratio of which is O(log n ⋅ log m), where n and m denote the number of points and the number of ranges, respectively.

We consider hypergraphs in which the union of two intersecting ranges is also a range. Our main result for such hypergraphs is as follows. The competitive ratio of the online hitting set problem is at least the unique-max coloring number and at most this number plus one. The proof of this result is via reductions from a unique-max coloring to an online algorithm, and vice-versa.

We also present improved competitive ratios for certain range spaces induced by connected induced subgraphs, half-planes, and unit discs.

Joint work with Shakhar Smorodinsky.
Bio:
Guy Even joined the Faculty of Engineering in Tel-Aviv University in 1997. He works on algorithms in general, including: decoding error correcting codes, online algorithms, approximation algorithms, scheduling, frequency assignment in wireless networks, routing, and algorithms for finding optimal policies in MDPs. He also worked on theory of VLSI and computer arithmetic.
Sharing a Sequential Program: Correctness and Relative Efficiency
Speaker:
Srivatsan Ravi
Type:
Project Group Meeting (PGT)
Time:
13 July 2012
11:00–12:00
Place:
Auditorium 2
Abstract:
It seems to be generally accepted that designing correct and efficient concurrent software is a sophisticated task that can only be held by experts. A crucial challenge then is to convert sequential code produced by a "mainstream" programmer into concurrent one. Various synchronization techniques may be used for this, e.g., locks or transactional memory, but what does it mean for the resulting concurrent implementation to be correct? And which synchronization primitives provide more efficiency at the end?
In this paper, we introduce a correctness criterion for a transformation that enables the use of a sequential data structure in a concurrent system. Informally, we require that the resulting concurrent implementation is locally sequential: concurrent threads simply run the given sequential code and let the implementation worry about the potential conflicts. To make sense globally, the implementation should also be linearizable with respect to the object type of the data structure.
We then evaluate the performance of different concurrent (locally sequential) implementations in terms of the sets of schedules (interleavings of steps of the sequential code) they accept. Intuitively, this captures the amount of concurrency that a given implementation can stand. This allowed us to analyze relative power of seemingly incomparable synchronization techniques, such as various forms of locking and transactional memory.
Impact of Network Effects on Applications Quality
Speaker:
Amir Mehmood
Type:
PhD Defense
Time:
9 July 2012
17:00–19:00
Place:
Auditorium 3
Measurement and Characterization of Internet Video Streaming Architectures
Speaker:
Thomas Krenc (TU Berlin master student)
Type:
Master's thesis final talk
Time:
3 July 2012
17:00–18:00
Place:
Auditorium 1
Abstract:
Internet video streaming traffic is booming as the demand of users for content and the broadband access speed continuously increase. Apart from the mainline commercial services such as IPTV operated by large ISPs, alternative services are also available. Among them are subscription-based services, e.g., Netflix, and free-of-charge One-Click-Hoster services hosted in datacenters, e.g., Rapidshare.

In this project we first review the alternative Internet video streaming architectures that are currently utilized and comment on their operation. We then crawl the content that is hosted and delivered by some of them, namely Netflix and kinox.to. Our results show that the availability of movies highly depends on the video streaming architecture. Moreover, during our three-month measurement study, we observe significant changes in the availability of movies in each of the streaming architectures we monitor. We attribute this to contract expirations with companies owning movies rights, and to the fact that many underground portals are closed down due to copyright violation. We also highlight that peer-to-peer architectures still provide the highest availability of movies today.
Tracing the Birth of an OSN: Social Graph and Profile Analysis in Google+
Speaker:
Doris Schiöberg
Type:
Project Group Meeting (PGT)
Time:
3 July 2012
15:00–16:00
Place:
Auditorium 1
Abstract:
This paper leverages the unique opportunity of Google launching the Google+ OSN. Through multiple crawls of the Google+ OSN, before and after the official public release of the network, our results provide insights into the social graph dynamics of the birth of an OSN. Our findings underline the impact of peculiar aspects of Google+ such as (a) Google’s large initial user base taken over from other Google products and (b) Google+’s provision for asymmetric friendships, on its graph structure, especially in light of previously studied OSN graphs. In addition, we study the geographic distribution of the users and links of Google+, and correlate the social graph with additional information available from the public profiles.
CloudNet virtualization technology
Speaker:
Arno Töll (TU Berlin diplom student)
Type:
Diplom thesis introductory talk
Time:
26 June 2012
17:30–18:00
Place:
Auditorium 1
CloudNet embeddings
Speaker:
Milan Mehner (TU Berlin diplom student)
Type:
Diplom thesis introductory talk
Time:
26 June 2012
17:00–17:30
Place:
Auditorium 1
On Inter-Domain Traffic Classification
Speaker:
Philipp Richter (TU Berlin master student)
Type:
Master's thesis introductory talk
Time:
19 June 2012
15:00–16:00
Place:
Auditorium 1
Abstract:
raffic classification has received much attention over the past years and thereby numerous methods for classifying traffic into applications have been proposed. These methods range from looking at port numbers and analyzing flow properties to performing deep packet inspection and using sophisticated data mining techniques to capture the communication patterns between hosts. What most of the existing methods have in common is that they have been mainly designed to address the needs of ISPs with respect to performing application-dependent service differentiation and developing content-sensitive pricing models. Thus, they are intended for the classification of intra-domain traffic.

Internet Exchange Points (IXPs) play an important role in today's AS-level Internet and therefore obtaining a better insight into the inter-domain traffic that is exchanged over their switching fabric requires more attention. To this end, classifying that traffic into applications is a first important step towards better understanding the inter-AS Internet.

Traces captured at large IXPs are usually exported in the form of sFlow records, which yield certain characteristics, such as heavy random sampling and stripped frame contents. These limitations of sFlow records along with the absence of methods that are explicitly intended for inter-domain traffic and of data that can be used to establish ground truth in this particular context make application classification a challenging task.

In light of the above, this thesis will focus on 1. assessing to what extent the available traffic classification methods (e.g., port-based application classification) are applicable to sFlow records collected at a large IXP 2. enhancing and/or extending the existing methods to yield accurate results when applied to inter-domain traffic 3. devising and evaluating new methods to deal with traffic that the existing methods fail to classify or misclassify 4. developing a framework to classify inter-domain traffic with a high level of confidence, given the absence of ground truth 5. exploring the implications of classifying the traffic that is exchanged at large IXPs for the better understanding of today's complex inter-domain environment with respect to the inter-AS traffic matrix.
Evaluation of Plug Computers as a Flexible Home-Network Measurement Tool
Speaker:
Julian Vetter (TU Berlin master studentl)
Type:
Master's thesis final talk
Time:
19 June 2012
15:00–16:00
Place:
Auditorium 1
Abstract:
The number of home networks is increasing. Nowadays, almost every connections runs a router to connect more than one device to the Internet. The home networks have to fulfill specific requirements because of several reasons. The bandwidth is increasing, home users are used to QoS, and new devices crowd into those networks (e.g. smart phones, tablet computers, set-top boxes etc.); All from different vendors with different implementations. These points create new challenges for home networks. Hence, making measurements in home networks attractive and necessary for ISPs, regulatory authorities and companies. But legal and administrative reasons complicate the look into these edge networks. The interested entities are not allowed to examine the home networks without the explicit permission of all affected home users.

One approach to solve this problem is to deploy a measurement device into households to perform measurements. We propose to use a plug computer for this purpose. Plug computers are small energy saving devices, yet more powerful than an average home router.

In this work we want to verify if plug computers fulfill the requirements to work as flexible home-network measurement devices. We suggest the use of virtualization to share resources of the device with the home users and/or third parties in a save manner. With virtualization we want to increase the flexibility. It allows easier resource sharing, management and policy enforcement. Moreover with the advantages of virtualization we want to achieve a greater acceptance of the device among the home users.

Part of the work has been spent on investigating low-level details of the Linux network driver used for the plug computer's network card. Because the only available virtualization solution does not provide direct support for the network card. Another aspect of this work is the automated execution of measurements with a measurement framework specially developed for this purpose.
From this work we draw some general conclusions related to networking in virtualized Linux systems. In particular, we show how several network parameters vary depending on the virtualization and the number of oncurrently running virtualized systems.
Wireless Noise Analysis and Modeling for BOWL Network
Speaker:
Thomas Juhre (TU Berlin master student)
Type:
Master's thesis final talk
Time:
14 June 2012
13:00–14:00
Place:
Fino
(TEL 20)
Abstract:
In the field of wireless communication, environmental noise is defined as any unwanted signal from external sources interfering and distorting the monitored signal, and varies with time and location. Investigations revealed that environmental noise extensively affects the packet delivery under certain circumstances, in particular if the ratio between signal power and noise power level is very low. In this diploma thesis, real measurement data of environmental noise in the 2.4 GHz frequency band of the Berlin Open Wireless Lab (BOWL) testbed are analyzed. The BOWL project at Technische Universit ̈t Berlin a (TUB) maintains an 13-node Asus router wireless outdoor mesh network deployed on the rooftops of the TU Berlin campus and also an 9-node Wistron CM9 router indoor mesh network. The testbed is used for Internet access and wireless networking experimentation including real network traffic. One application of the noise analysis is in wireless network simulations, which depends on knowledge of the statistical distribution of packet reception for each node. One of the variables which is considered in the distribution function is noise. In order to have a better understanding on modeling environmental noise in BOWL outdoor and indoor network, two different measurement data sets, each generated by another wireless driver, for each node of BOWL are analyzed. Characteristics of noise, such as the night and day effect, time intervals with constant periods or repeating patterns sampled on each IEEE 802.11 Atheros radio in BOWL network are of main interest. In the first part, the data sets are examined with regard to consistency, correctness and plausibility, since they are generated by a new measurement tool called wprobe. The second part analyzes traces of a duration of 24 hours in order to get insights about short-term characteristics of noise and on the other hand, traces of a duration of 5 days in order to study long-term characteristics for the outdoor and the indoor network of BOWL. The last part introduces 2 models that aim to represent the noise floor of BOWL outdoor and indoor nodes.
Programming the Enterprise WLAN: An SDN Approach
Speaker:
Lalith Suresh (INESC-ID/Instituto Superior Tecnico)
Type:
Project Group Meeting (PGT)
Time:
12 June 2012
15:00–16:00
Place:
Auditorium 1
Abstract:
We present Odin, an SDN framework to introduce programmability in enterprise wireless local area networks (WLANs). Enterprise WLANs need to support a wide range of services and functionalities. This includes authentication, authorization and accounting, policy, mobility and interference management, and load balancing. WLANs also exhibit unique challenges. In particular, access point (AP) association decisions are not made by the infrastructure, but by clients. In addition, the association state machine combined with the broadcast nature of the wireless medium requires keeping track of a large amount of state changes. To this end, Odin builds on a light virtual AP abstraction that greatly simplifies client management. Odin does not require any client side modifications and its design supports WPA2 Enterprise. With Odin, a network operator can implement enterprise WLAN services as network applications. This talk will cover the motivation, design, implementation and evaluation of Odin.
Bio:
Lalith Suresh is a student of the European Masters in Distributed Computing (EMDC) programme across Instituto Superior Tecnico (IST), Lisbon and Royal Institute of Technology (KTH), Stockholm. He is presently conducting work leading to his master’s thesis from the INET group at T-Labs.
Fibasim: A Simulation-based Evaluation of FIB Aggregation Algorithms
Speaker:
Robert Wuttke (TU Berlin Master Student)
Type:
Master's thesis introductory talk
Time:
12 June 2012
14:00–15:00
Place:
Auditorium 1
Abstract:
The Internet is growing rapidly and so are the Forwarding Information Bases (FIBs) in Internet routers. Various FIB aggregation algorithms have been proposed to reduce FIB memory requirements which could increase the lifetime of a router. In this work we evaluate FIB aggregation schemes using fibasim, a modular simulator which reads FIB snapshots and a stream of consecutive FIB updates, processes them using an aggregation algorithm, and computes summaries about the original and the aggregated FIB while also tracking the running time of the aggregation algorithm itself. A lot of effort is put into the verification of the correctness of the simulator to ensure the validity of the generated results.
Overview of our recent research in white space networking
Speaker:
Bozidar Radunovic, Microsoft Research, Cambridge
Type:
Talk
Time:
31 May 2012
11:00–12:00
Place:
Consilium (TEL 1315)
Abstract:
In the talk I will discuss some unique challenges facing wireless networking in unused analogue TV frequencies (white spaces), and I will give an overview of our research that addresses some of the challenges. In particular, I will focus on the power asymmetry problem and the spectrum fragmentation problem. Due to various regulatory constraints white space networks give rise to power heterogeneity. This further pronounces the hidden terminal problem and yields frequent starvation of low-power nodes. Also, non-contiguous spectrum availability and diverse bandwidth capability of various devices makes it difficult to efficiently use fragmented spectrum. I will present two of our recent projects with novel PHY/MAC designs that address these two challenges.
Bio:
Bozidar Radunovic is a Researcher in the Networks, Economics and Algorithms group at Microsoft Research, Cambridge. His research interests are in architecture and performance evaluation of computer systems with particular interest in wireless communication, cross-layer design and application of advanced communication techniques in network system design.
Bozidar received his PhD in technical sciences from EPFL, Switzerland, in 2005, and his BSc at the School of Electrical Engineering, University of Belgrade, Serbia, in 1999. He was a PhD student at LCA, EPFL from 2000–2005. Then he did a one year post-doc at TREC, at ENS Paris, in 2006. He also did a 6 month internship in IBM Zurich Research Labs in winter 2004/05. In 2008 he has been awarded IEEE William R. Bennett Prize Paper Award in the Field of Communications Systems.
Fast car counting from planet-scale web cameras and profiling of global urban traffic dynamics
Speaker:
Bai Du (TU Berlin master student)
Type:
Master's thesis introductory talk
Time:
29 May 2012
15:00–16:00
Place:
Auditorium 1
Abstract:
The increase in number of vehicles has created a problem of traffic congestion in many cities across the globe. Longitudinal and large-scale vehicular datasets are very important in the study of traffic. Although there are many ways to collecting the urban traffic information, but most of them, just like GPS coordinates, are usually expensive. My master thesis is to find an effective way to use the data from traffic web cameras, which is free to get and used worldwide, and extract the useful information from them, then use this information to analysis the urban traffic dynamics. Processing data from a global network of web camera is a big challenge, due to the low quality of the cameras, the low rate of the video (“snapshots”), and the large diversity of the mounting angles of the cameras; I need to make a fast and unified algorithm to process the huge amount of data, which is about 7.5 terabytes. And finally I will discuss the traffic dynamic in two cities with the current data.
Bio:
Bai Du is a master student at TU Berlin, he studies electronic technology. His bachelor degree was from Beijing Institute of Technology, and his specialty is Information Engineering.
Real-world Polymorphic Attack Detection
Speaker:
Evangelos Markatos, FORTH-ICS and Univ. of Crete
Type:
Talk
Time:
23 April 2012
10:00–11:00
Place:
TEL 1118/19
Abstract:
As state-of-the-art attack detection technology becomes more prevalent, attackers have started to employ evasion techniques such as code obfuscation and polymorphism to defeat existing defenses. We have recently proposed network-level emulation, a heuristic detection method that scans network traffic to detect polymorphic attacks. Our approach uses a CPU emulator to dynamically analyze every potential instruction sequence in the inspected traffic, aiming to identify the execution behavior of certain malicious code classes, such as self-decrypting polymorphic shellcode. In this work, we present results and experiences from deployments of network-level emulation in production networks. After more than a year of continuous operation, our prototype implementation has captured more than a million attacks against real systems, while so far has not resulted to any false positives. The observed attacks employ a highly diverse set of exploits, often against less widely used vulnerable services, and in some cases, sophisticated obfuscation schemes.
Bio:
Prof. Evangelos Markatos received his diploma in Computer Engineering from the University of Patras in 1988, and the M.S and Ph.D. degrees in Computer Science from the University of Rochester, NY in 1990 and 1993 respectively. Since 1992, he collaborates with the Institute of Computer Science of the Foundation for Research and Technology — Hellas (ICS-FORTH) where he is currently the founder and head of the Distributed Computing Systems Laboratory. He conducts research in several areas including distributed and parallel systems, the World-Wide Web, Internet Systems and Technologies, as well as Computer and Communication Systems Security. He has been the project manager of the LOBSTER and NoAH projects, both funded in part by the European Union and focusing on developing novel approaches to network monitoring and network security. He is currently the project manager of the i-code and SysSec projects. Since 1992, he has also been affiliated with the Computer Science Department of the University of Crete, where he is currently a full Professor.
Interactions of Legacy Internet Traffic with Multihop Wireless Networks
Speaker:
Harald Schiöberg
Type:
PhD Defense
Time:
21 March 2012
15:00–16:00
Place:
Auditorium 3
Measurement and Characterization of Content Distribution in BitTorrent
Speaker:
Thomas Krenc (TU Berlin master student)
Type:
Master's thesis final talk
Time:
27 March 2012
15:00–16:00
Place:
Auditorium 1
Abstract:
BitTorrent is the most popular P2P system. It is primarily used for file sharing. Despite the fact that recent studies show a decline of BitTorrent traffic, it is still very popular in many regions of the world and there is an increasing trend to use BitTorrent-like systems in commercial products. Key issues in BitTorrent are the availability of content and the efficient content distribution. These problems have been studied from a theoretical point of view, but a large-scale measurement and characterization study of content availability and distribution in the wild has to still be done.

In this thesis, we design, develop and run a system that monitors the activity of BitTorrent users. The system periodically connects to peers and reports the availability and propagation of content in millions of users and thousands of swarms. We utilize our system to collect and analyze measurements that span a period of more than three months. Our study sheds light on the operation of BitTorrent in the wild. In particular, our results show that the popularity of BitTorrent as well as the propagation of content in BitTorrent highly depends on the type of content, the ISP, as well as the swarm demographics. Furthermore, we uncover diurnal and flash-crowd cycles in content availability. We also show that a large fraction of BitTorrent content is available in diverse network locations.
Flow-based Intrusion Detection
Speaker:
Anna Sperotto (University of Twente)
Type:
Project Group Meeting (PGT)
Time:
6 March 2012
15:00–16:00
Place:
Auditorium 1
Abstract:
The spread of 1–10 Gbps technology has in recent years paved the way to a flourishing landscape of new, high-bandwidth Internet services. At the same time, we have also observed increasingly frequent and widely diversified attacks. To this threat, the research community has answered with a growing interest in intrusion detection, aiming to timely detect intruders and prevent damage. We believe that the detection problem is a key component in the field of intrusion detection. Our studies, however, made us realize that additional research is needed, in particular focusing on validation and automatic tuning of Intrusion Detection Systems (IDSs).

This talk presents the main contribution of the speaker's PhD thesis, defended in October 2010 at the University of Twente, The Netherlands. The contribution of this thesis is that it develops a structured approach to intrusion detection that focuses on (i) system validation and (ii) automatic system tuning. We developed our approach by focusing on network flows, which offer an aggregated view of network traffic and help to cope with scalability issues.
Bio:
Anna Sperotto is a PostDoc at the Design and Analysis of Communication Systems (DACS) group, which belongs to the Faculty of Electrical Engineering, Mathematics, and Computer Science at the University of Twente. Currently she is participating in the European FP Project Univerself. Previously, she took part in the European Network of Excellence on Management Solutions for Next Generation Networks (EMANICS). On Octber 14, 2010, Anna defended her thesis "Flow-based Intrusion Detection", at the University of Twente. Previously, she graduated in Computer Science from the Ca' Foscari University, Venice, Italy, in 2006 (Master of Science) and in 2004 (Bachelor). Her main topic of interest is Intrusion Detection, Self-Learning, Network Modeling and Graph Theory.
Evaluation of Plug Computers as a flexible Home-Network Measurement Tool
Speaker:
Julian Vetter (TU Berlin master student)
Type:
Master's thesis introductory talk
Time:
21 February 2012
14:00–15:00
Place:
Auditorium 1
Abstract:
The increasing number of high speed broadband internet connections raises the need for performance verification of these connections. Customers want to verify if they chose the right contract. On the other hand ISPs need to measure problematic connections and the regulatory authorities want to protect customers by verifying that the ISPs fulfill their SLAs.

One solution would involve deploying an active measurement device into multiple households and remotely run network measurements, potentially uninterrupted for some time period without direct user interaction.

The scope of this master thesis is the evaluation of the platform type of plug computers as a flexible home-network measurement tool. Main goal is to run a plug computer (namely Sheevaplug) with a microkernel (L4), with multiple compartments of Linux, to not only do home network measurements but also share parts of the hardware resources with the customer or other 3rd parties (i.e., researchers). With the potentially added value of sharing resources with the customer we hope to achieve a greater acceptance and adoption of the device. Our experiments include
  • measuring the Sheevaplug itself, how much overhead is introduced by microkernel virtualization compared to native Linux.
  • How many resources can potentially be handed over to the customer, without interfering with our measurements
  • measuring the performance of a broadband Internet access with the device plugged in to the local area network and finally deployment of the device to multiple homes and running measurements remotely.
Another important aspect of the thesis is the remote management of the device with the goal of deploying new experiments dynamically on demand.
Analyse und Untersuchung des World-of-Warcraft-Netzwerkprotokolls
Speaker:
Matthias Siebke (TU Berlin Diplom student)
Type:
Diplom thesis final talk
Time:
2 February 2012
12:00–13:00
Place:
Sputnik
Abstract:
Mit der steigenden Zahl an Online-Computerspielen im letzten Jahrzehnt stieg auch der durch sie verursachte Datenverkehr im Internet. Daher besteht die Notwendigkeit den speziellen Netzwerkverkehr von Spielen zu untersuchen, um zu verstehen, wie diese Anwendungen das Internet als Kommunikationsmedium nutzen und beeinflussen.

Ziel der Diplomarbeit 'Analyse und Untersuchung des World-of-Warcraft-Netzwerkprotokolls' war es, die Funktionsweise der Kommunikationsprotokolle eines der größten Online-Spiele der letzten Jahre zu ermitteln und dessen verursachten Internetdatenverkehr zu untersuchen. Aufbauend auf der Analyse der Funktionsweise der Protokoll wurden zwei Protokoll-Analyzer für das Network Intrusion Detection System Bro implementiert, die in der Lage sind World-of-Warcraft-Netzwerkverkehr portunabhängig zu klassifizieren und bestimmte Daten daraus zu extrahieren. Zur Untersuchung des World-of-Warcraft-Datenverkehrs wurden diese Analyzer auf einen Internetverkehrsmitschnitt angewendet, um die spezifischen Eigenschaften exemplarisch anhand der im Mitschnitt vorhandenen World-of-Warcraft-TCP-Verbindungen untersuchen zu können. Da sich der Netzwerkverkehr von Computerspielen vom Verkehr klassischer Internetanwendungen wie WWW, File-Sharing oder E-Mail unterscheidet und zusätzlich durch das individuelle Verhalten der Spieler beeinflusst ist, wurden in der Arbeit neben den allgemeinen Paketeigenschaften der World-of-Warcraft-TCP-Verbindungen zusätzlich Bewegungsinformationen im Spiel selbst und das Nutzerverhalten analysiert.

Im Abschlussvortrag zur Diplomarbeit sollen nun die Ergebnisse der Untersuchung vorgestellt werden, um einen Überblick darüber zu geben, wie moderne Computerspiele kommunizieren, welchen Internetverkehr sie verursachen und durch welche Faktoren der Verkehr beeinflusst ist.

Zusatzinformationen / Extras

Direktzugang:

Schnellnavigation zur Seite über Nummerneingabe