Registration

Please visit the seminar information in the CAMPUS system.

Important Dates

Kick-off Meeting

May 4th, 2006: 12:00h - 13:00h

Informatikzentrum, room 4105, Chair for Computer Science IV

Slides of the kick-off meeting

Seminar Talks:

August 17th-18th, 2006: 09:00h - 17:00h

B-IT Center Bonn (transportation from RWTH Aachen and back will be provided)

Submission Date for Slides:

July 24th, 2006

Submission Date for Papers:

September 1st, 2006

Topics

All topics are supervised by Stefan Götz, Simon Rieche, or Olaf Lansiedel.

The Distributed Systems That Surround Us

• The IP Routing Infrastructure (H. Zhu, Stefan Götz)

  ISP? AS? SLA? BGP? Wtf? Who runs the Internet? How do providers cooperate? What does the routing backbone look like? - An Overview, data on and visualization on Internet traffic

• The Domain Name System (Stefan Götz)

  How are DNS servers organized? How important is caching? How vulnerable is the DNS? Who controls the DNS? - An Overview, Basic DNS Information, Secure DNS, DNS Threat Analysis, Philosophical DNS issues

• PlanetLab (Stefan Götz)

  More than 700 nodes around the globe at your finger tips, and what to do with them - The Home Page, Design Overview, Research practices on PlanetLab Next Generation DNS, Distributed Internet Monitoring

Peer-to-Peer Systems - A New Breed?

• Unstructured P2P-Systems (Simon Rieche)

  From Napster to Gnutella and Bittorrent, from centralized to pure P2P. The first and second generation (of file-sharing applications) of Peer-to-Peer systems.

  • Jörg Eberspächer, Rüdiger Schollmeier "First and Second Generation of Peer-to-Peer-Systems" in Ralf Steinmetz, Klaus Wehrle (Eds.), Peer-to-Peer Systems and Applications, LNCS 3485, p. 79-93, ISBN 3-540-29192-X, Springer, 2005
  • Bittorrent Overview

With Distributed Hash Tables or structured Peer-to-Peer systems, like CAN (Content Addressable Network), Chord or P-Grid, a far more efficient method for looking up data in a Peer-to-Peer system emerged since 2001. With their efficient, scalable, and self-organizing algorithms for data retrieval and management, DHTs offer crucial advantages compared to unstructured approaches The following five topics are concerned with such systems or applications.

• Consistent Hashing and Chord (D. Schulte, Simon Rieche)
  • Consistent Hashing and Random Trees: Distributed Caching Protocols for Relieving Hot Spots on the World Wide Web,
  • Chord: A Scalable Peer-to-peer Lookup Service for Internet Applications
• Content Adressable Networks: (T. Kurpick, Simon Rieche)
  • A Scalable Content Addressable Network
• Pastry and Tapestry: (Simon Rieche)
  • Pastry: Scalable, distributed object location and routing for large-scale peer-to-peer systems,
  • Tapestry: A Resilient Global-scale Overlay for Service Deployment
• p-Grid (Z. Hu, Simon Rieche)
  • Publications
• OpenDHT and Place Lab: (Simon Rieche)
  • Home Page
  • Publications Related to OpenDHT
• Extended Key Spaces (K. Meyer, Stefan Götz)

  Approaches like Skipnets and Skip Graphs go beyond the classic numeric key space, naturally supporting, e.g., strings for interesting locality and organizational properties, with Brushwood even generalizing the concept to arbitrary keys.

• When nodes go south (A. Ben Moussa, Stefan Götz)

  Frequent node arrivals, departures, and failures (so-called churn) are a common but difficult problem to handle in distributed systems. To understand churn, statistical analysis and modelling can be very helpful, as done by, e.g., Loguinovi and Stutzbach et. al. Existing approaches range from advance redundancy schemes to optimizing the routing information to neighbor selection based on their reliability. Related work in PlanetLab shows how severe churn and performance issues are even in well-administered networks.

• Trust in P2P (F. Schmidt, Stefan Götz)

  Establishing trust between anonymous peers seems impossible at first, however, current and past actions of a peer are a good indicator for its future behaviour. SLIC, for example, exploits this fact to create bandwidth incentives for other peers to cooperate, similar to BitTorrent. EigenTrust and an approach from EPFL rely on elaborate schemes for reputation management for creating global views of trust relationships. Sometimes, more simple probabilistic approaches seem to be good enough, though, to tell good from bad. Olaf says: cool stuff from Berkeley

• Mobility in P2P (S. Jansen, Olaf Landsiedel)

  For the use in the Internet domain, distributed hash tables (DHTs) have proven to be an efficient and scalable approach to distributed content storage and access. The goals of this Semiar topic are to explore how DHTs and mobile ad hoc networks (MANETs) fit together. One can easily notice, that both share key characteristics in terms of self organization, decentralization, redundancy requirements, and limited infrastructure. However, node mobility and so the continually changing physical topology pose a special challenge to scalability and the design of a DHT for mobile ad-hoc networks.

  • Mobile Hash Tables (MHT)
  • Ekta: An Efficient DHT Substrate for Distributed Applications in Mobile Ad Hoc Networks
  • Multi Level Peer Index (MPI)
  • A Distributed Search Service for Peer-to-Peer File Sharing in Mobile Applications
  • GHT: A Geographic Hash Table for DataCentric Storage
• Anonymous Communication in P2P (C. Lücking, Olaf Landsiedel)
  • Untraceable electronic mail, return addresses, and digital pseudonyms
  • Tor: The Second-Generation Onion Router
  • Tarzan: A Peer-to-Peer Anonymizing Network Layer
  • Low-Cost Traffic Analysis of Tor
  • Core: Connectionless Onion Routing

  Sensor Networks - Small Is Beautiful

  • Applications & Deployments (K. Herings, Olaf Landsiedel)
    • Wireless sensor networks, Introduction
    • Habitat monitoring with sensor networks
    • Sensor Network-Based Countersniper System
    • A Wireless Sensor Network for Structural Monitoring
    • An Analysis of a Large Scale Habitat Monitoring Application
  • Hardware (TBA, Olaf Landsiedel)
    • Wireless sensor networks, Introduction
    • The platforms enabling wireless sensor networks
    • Telos: Enabling Ultra-Low Power Wireless Research
    • Design of a Wireless Sensor Network Platform for Detecting Rare, Random, and Ephemeral Events
    • The Intel Mote platform: A Bluetooth based sensor network for industrial monitoring applications
  • Operating Systems / TinyOS (T. Lohmann, Olaf Landsiedel)
    • Wireless sensor networks, Introduction
    • System architecture directions for network sensors
    • The Emergence of Networking Abstractions and Techniques in TinyOS
    • Towards a Sensor Network Architecture: Lowering the Waistline
  • Medium Access Control (MAC) (D. König, Olaf Landsiedel)
    • Wireless sensor networks, Introduction
    • Versatile Low Power Media Access for Wireless Sensor Networks
    • A Unifying Link Abstraction for Wireless Sensor Networks
    • An energy-efficient MAC protocol for wireless sensor networks
    • An Adaptive Energy-Efficient MAC Protocol for Wireless Sensor Networks
  • Routing (T. Wellens, Olaf Landsiedel)
    • Wireless sensor networks, Introduction
    • Taming the Underlying Challenges of Reliable Multihop Routing in Sensor Networks
    • Beacon Vector Routing: Scalable Point-to-Point Routing in Wireless Sensornets
    • Directed Diffusion: A Scalable and Robust Communication Paradigm for Sensor Networks
    • Networking support for query processing in sensor networks

Seminar Details

The seminar provides a broad overview on existing distributed system and current research topics in the areas of peer-to-peer and sensor networks. It is our paramount interest to see that you gain a thorough understanding of your seminar topic and that you are able to convey this understanding to the other participants. For you to benefit from the presented knowledge, we encourage active participation and interaction between students. You will both synchronize with each other while preparing your talk as some topics overlap and discuss the contents after the talks themselves.

Your contribution to the seminar is twofold: you present your topic concisely in a 30-minute talk to the other seminarists and you supply a paper providing more detail on the topic than the talk. Furthermore, you are expected to engage in discussions about each talk. Plagiarism of any form is unacceptable and will lead to your immediate suspension from the seminar. The recommendations at http://www-i4.informatik.rwth-aachen.de/ -> Teaching -> Seminar are a good starting point for designing and writing your slides and your paper. Please adhere to them to avoid disappointment when discussing your work with your supervisor.

Talks

The goal of the talk is to give the audience a good understanding of the whole topic and to dive into a few interesting details of the subject matter. The ratio of one to the other depends on the topic and needs to be determined with your supervisor. Each talk is scheduled to be 40 minutes, with 35 and 45 minutes being hard lower and upper bounds, respectively. Ten minutes of discussion follow each talk. Please use the DS template for your slides

Papers

While the talk focusses on the overview and details of interest, the paper allows you to discuss your topic in its full breadth and depth. It covers all aspects of the talk and provides additional insights to related work and specifics. For example, source code snippets or interface descriptions would not go into your talk but may well show up in your paper where applicable.

The papers are based on the official ACM conference style and are expected to be between ten and twelve pages in length (hard limits). ACM provides official templates for a number of formats but we encourage you to use LaTeX for writing your paper.