Volume 49, Issue 5

Peer-to-Peer is a fundamental design principle and marks a paradigm shift for communication in computer networks. In this paper first a definition for peer-to-peer systems is given based on their characteristics. Using four quality attributes the DFG Research Group “QuaP2P” subdivided its research fields. Subsequently, the state-of-the-art in peer-to-peer research and its challenges are summarized, following this classification. This will give a structured overview to the reader in order to provide him with the relevant related work when looking into this highly active research area.

Structured peer-to-peer (p2p) networks are highly distributed systems with a potential to support business applications. There are numerous different suggestions on how to implement such systems. However, before legal p2p systems can become mainstream they need to offer improved efficiency, robustness, and stability. While Chord is the most researched and best understood mechanism, the Kademlia algorithm is widely-used in deployed applications. There are still many open questions concerning the performance of the latter. In this paper we identify the main problems of Kademlia by large scale simulations and present modifications which help to avoid those problems. This way, we are able to significantly improve the performance and robustness of Kademlia-based applications, especially in times of churn and in unstable states. In particular, we show how to increase the stability of the overlay, make searches more efficient, and adapt the maintenance traffic to the current churn rate in a self-organizing way. Zusammenfassung Strukturierte Peer-to-Peer (P2P) Netze sind eine vielversprechende Alternative zum klassischen Client-Server-Ansatz. Die Effizienz, die Robustheit und die Stabilität solcher Systeme ist allerdings noch nicht ausreichend untersucht, um diese neue Technik im Produktivumfeld einsetzen zu können. In dieser Arbeit greifen wir daher Kademlia als einen typischen Vertreter strukturierter P2P-Netze auf und analysieren die genannten Eigenschaften mittels Simulation großer Teilnehmerzahlen. Auf Basis der gewonnenen Einsichten schlagen wir Modifikationen vor, die die Stabilität erhöhen, Suchen effizienter gestalten und die in der Lage sind, den nötigen Verkehr zur Aufrechterhaltung der Struktur selbständig an die äußeren Umstände anzupassen.

The performance of peer-to-peer and overlay networks depends to a large extent on their awareness of the underlying network's properties. Several schemes for estimating end-to-end network distances have been proposed to simplify this task. The mOverlay framework identifies groups of nodes that are near to each other in the network topology. Instead of distances between nodes mOverlay measures distances between groups. However, mOverlay's locating procedure has a number of drawbacks. We propose an alternate method for identifying groups using Meridian's closest node search. Simulation results based on PlanetLab measurements indicate that the Meridian-based approach is able to outperform mOverlay in terms of joining delay, the size of the identified groups, and their suitability for a distance estimation service. This alternate method for identifying groups is our main contribution. Zusammenfassung Die Leistungsfähigkeit von Peer-to-Peer und Overlay-Netzen hängt weitgehend von der Kenntnis der Eigenschaften des unterliegenden Netzes ab. Verschiedene Mechanismen zur Distanzschätzung wurden hierzu in der Vergangenheit vorgeschlagen. Das Verfahren mOverlay identifiziert Gruppen von Knoten, welche in einer Netztopologie nahe beieinander liegen. Statt Distanzen zwischen einzelnen Knoten zu bestimmen, werden Distanzen zwischen Gruppen ermittelt. Das mOverlay-Verfahren hat jedoch gewisse Nachteile. Daher wird in diesem Artikel ein alternatives, auf Meridian basierendes Verfahren zum Ermitteln von Gruppen vorgeschlagen. Meridian unterstützt die Suche von nächsten Knoten zu einem gegebenen Knoten. Der Hauptbeitrag dieses Artikels besteht in diesem alternativen Verfahren. Simulationen, die auf Messungen in PlanetLab beruhen, zeigen, dass das alternative Verfahren hinsichtlich Gruppenbeitrittsverzögerungen, der Größe identifizierter Gruppen sowie der Qualität der Distanzschätzungen Verbesserungen erzielt.

Mobile ad-hoc networks and distributed hash tables (DHT) share key characteristics in terms of self organization, decentralization, redundancy requirements, and limited infrastructure. However, node mobility and the continually changing physical topology pose a special challenge to scalability and the design of a DHT for mobile ad-hoc networks. In this paper, we show that with some local knowledge we can build a scalable and mobile structured peer-to-peer network, called Mobile Hash Table (MHT). Furthermore, we discuss practical challenges such as Churn, load balancing and security of MHT. A special focus is put on the differences and new challenges that the use of a DHT in a mobile environment poses.

In order to distribute multimedia content in peer-to-peer based approaches, all participants choose other peers as neighbors, thus spanning an overlay network of logical nodes and connecting edges. In building such an overlay for live multimedia streaming, two important properties of the evolving topologies have to be considered: The efficiency regarding the transmission cost in the network and the stability towards random node failures and deliberate attacks. The majority of existing approaches focuses on the construction of efficient overlays and only a few aim to achieve resilience towards the failure of nodes. Furthermore, no approaches exist which propose a satisfying solution to the creation of both efficient and stable topologies. In order to close this gap, this article presents a new approach for an adjustable construction of both efficient and stable overlay streaming topologies. For this purpose, the properties of stability and efficiency are formalized as a set of cost metrics and a distributed algorithm for dynamic construction of such topologies is introduced, which locally minimizes the total cost. A simulation study shows that this approach allows for a good trade-off between the construction of efficient and stable topologies.

The benefits of the peer-to-peer paradigm have been proven through various applications besides file sharing. The requirements for the design of peer-to-peer overlay networks vary according to its purpose. In order to compare existing overlay networks and determine their suitability for specific purposes, requirements are defined with abstract quality attributes. Once the benchmarking set (quality attribute, metrics, and scenarios) is identified, experiments should be applied under the same circumstances on each overlay in order to obtain comparable results. This paper presents PeerfactSim.KOM, a simulator providing a benchmarking platform for peer-to-peer systems, especially for overlay networks. It supports defined benchmarking sets for all kinds of peer-to-peer overlays through an implemented catalogue of metrics and a simple but comprehensive scenario specification. Various peer distributions and churn rates are given which also supports geographical-location dependence. The platform is extensible due to its modular design and can scale up to around 10 6 ; peers for simple overlays such as Gnutella and 10 5 ; for more complex overlays like Kademlia. Zusammenfassung In vielen Anwendungen jenseits der Dateitauschbörsen zeigen sich die Vorteile des Peer-to-Peer Kommunikationsparadigmas. Abhängig vom Zweck variieren die Anforderungen an das Design eines geeigneten Peer-to-Peer-Overlay-Netzes. Um existierende Overlay-Netze vergleichen zu können und um zu entscheiden, ob sie für einen festgelegten Zweck geeignet sind, werden die Anforderungen mit abstrakten Qualitätsattributen beschrieben. Hat man einen Benchmarking-Satz (Qualitätsattribute, Metriken und Szenarios) identifiziert, sollte dieser, um vergleichbare Resultate zu erzielen, bei der Messung eines jeden Overlay-Netzes eingesetzt werden. Der Schwerpunkt dieser Arbeit ist Peerfact.KOM, ein Simulator, der eine Benchmarking-Plattform für Peer-to-Peer-Overlay-Netze bereitstellt. Für alle verschiedenen Arten von Peer-to-Peer-Overlay-Netzen bietet er festgelegte Benchmarking-Sätze, einen implementierten Katalog von Metriken und eine einfache, aber ausdrucksstarke Methode, um Szenarien zu spezifizieren. Verschiedene Churn-Raten und Peer-Verteilungen, die optional auch auf geographische Ortsangaben bezogen sind, werden geboten. Der Simulator ist durch sein modulares Design erweiterbar. Er skaliert bis zu 10 6 ; Peers in einfachen Overlay-Netzen wie Gnutella und 10 5 ; in komplexeren Overlay-Netzen wie Kademlia.

Main topic of the ongoing DFG-funded Research Unit 500 “Computer aided destruction of complex structures using controlled explosives” is the realistic computer simulation of demolition of structures using controlled explosives. Thereby, in particular real world uncertainties are considered to improve the planning process and to determine a verified optimal demolition strategy. With respect to the readership, this article focuses on concepts and implementation of the software system for planning of demolition projects, which is based on simulation models for the collapse of complex high-rise reinforced concrete buildings using multibody systems.