Kolja Eger

Efficient simulation of large-scale p2p networks: packet-level vs. flow-level simulations

Peer-to-peer (P2P) networks can reduce the distribution cost of large media files for the original provider of the data significantly. Thereby, the BitTorrent protocol is widely used in the Internet today. Most research work studies the protocol analytically, by simulations at the flow-level or real world experiments. Thereby, for flow-level simulations the influence of neglecting packet-level characteristics is not yet quantified. Therefore, this paper compares packet-level simulation results with flow-level values and analytically derived bounds. Our findings show that BitTorrent is near to optimal at flow-level for different scenarios. Naturally, packet-level results deviate more from the optimal values but differences are at most around 30% in our simulations. Furthermore, we show that the propagation delay can significantly influence the download performance of BitTorrent.

Bandwidth Trading in Unstructured P2P Content Distribution Networks

Bandwidth trading schemes give peers an incentive to provide upload bandwidth to other peers in a P2P network for fast file distribution. A popular example is the tit-for-tat strategy used in the BitTorrent protocol. Although this game theoretical scheme provides an incentive to peers to contribute resources to the network it does not prevent unfairness and the performances of peers vary considerably. Therefore, we propose two new trading schemes, which are based on pricing. One uses explicit price information whereas the other scheme uses the download rates from other peers as the price. For both distributed algorithms the stable point provides a fair resource allocation as well as a Nash equilibrium, i.e. fairness is preserved although peers behave selfishly and try to maximise their own download rates only. We compare both pricing schemes with BitTorrent in simulations of static and dynamic networks. The pricing algorithms outperform BitTorrent with respect to fairness. With explicit prices the download rates converge faster to the fair equilibrium than with implicit ones

Bandwidth trading in BitTorrent-like P2P networks for content distribution

Bandwidth trading schemes give peers an incentive to provide upload bandwidth to other peers in a P2P network for fast file distribution. A popular example is the tit-for-tat strategy used in the BitTorrent protocol. Although this game theoretical scheme provides an incentive to peers to contribute resources to the network it does not prevent unfairness and the performances of peers vary considerably. Therefore, we propose two new trading schemes, which are based on pricing. One uses explicit price information whereas the other scheme uses the download rates from other peers as the price. For both distributed algorithms the stable point provides a fair resource allocation as well as a Nash equilibrium. Thus, fairness is preserved although peers behave selfishly and try to maximize their own download rates only. We compare both pricing schemes with BitTorrent in simulations of static and dynamic networks. In BitTorrent peers receive different download rates even if they provide the same upload bandwidth. Furthermore, peers with small upload capacities compared to others receive considerably more than what they contribute. The pricing algorithms outperform BitTorrent with respect to fairness. With both algorithms a peer receives a download performance proportional to its upload capacity. With explicit prices the download rates converge faster to the fair equilibrium than with implicit ones.

Fair resource allocation in peer-to-peer networks (extended version)

The first peer-to-peer (P2P) networks were based mainly on the altruistic behavior of the peers. Although newer implementations incorporate some kind of incentive mechanism to award sharing peers, no P2P network assures a guaranteed service rate. This article is meant as a first step towards the development of P2P networks with guaranteed service rate. We propose a distributed resource allocation algorithm where peers control the service rates to their neighbors. This algorithm is based on the congestion pricing principle known from IP networks and ensures some form of fairness. Hence a peer gets a fair share of the resources available in the P2P network weighted by its contribution to the network. We study the convergence properties of the distributed algorithm and validate them by simulation. Further simulation results present the functionality of the algorithm in large and varying networks. The results indicate a fair allocation of the resources even when the service rates of some peers deteriorate due to errors.

Resource pricing in peer-to-peer networks

In this letter resource pricing is applied to peer-to-peer (P2P) networks to ensure a fair allocation of resources. The differences to congestion pricing in IP networks are described and a distributed algorithm based on locally available information is proposed. Asymptotic stability for the global optimum of the proposed system is proven based on a Lyapunov function

Efficient Simulation of Large-Scale P2P Networks: Modeling Network Transmission Times

The ongoing process of globalization leads to a huge demand for highly scalable applications that are able to deal with millions of participants distributed all over the world. Peer-to-peer (P2P) technology enables an arbitrary large number of users to participate in distributed services like content distribution or collaboration tools. In order to verify a new protocols performance and scalability simulation is a commonly used tool. First, predicting the network and peer behavior in the real world is only feasible if the simulation, i.e. all applied models as well as the peer state, is as realistic as possible. Second, many properties of the system only become observable when the number of participants is sufficiently large. Therefore, verifying the scalability of a system requires simulating huge worldwide networks. Due to limited processing power, central memory and availabe time, both requirements can only be fullfilled if the applied models are very efficient. In this paper we take a closer look at the network layer. We compare the most commonly-used network models and present a very efficient model for applying real-world network transmission times in large scale simulations

Efficient Simulation of Large-Scale P2P Networks: Compact Data Structures

One of the most important design goals of current peer-to-peer (P2P) technology is to be able to offer its service to an arbitrary large number of users. Discrete event simulation is often applied to quantitatively and qualitatively evaluate the performance and scalability of such systems before they are deployed. However, the number of users, processes and events which can be simulated is limited by both the central memory and the time available. In this paper we present compact data structures and event design algorithms, which are intended to be a further step towards efficient simulation of large scale P2P systems. In particular, we give guidelines on how to increase the number of peers which can be simulated and show how to find a good tradeoff between computational time and memory consumption in large scale P2P simulation

TCPeer: rate control in P2P over IP networks

The prevalent mechanism to avoid congestion in IP networks is the control of the sending rate with TCP. Dynamic routing strategies at the IP layer are not deployed because of problems like route oscillations and out-of-order packet deliveries. With the adoption of P2P technology, routing is done also in these overlay networks. With multi-source download protocols peers upload and download to/from other peers in parallel. Based on congestion pricing for IP networks this paper proposes a rate control algorithm for P2P over IP networks. A peer adopts the functionality of TCP and extends the congestion window mechanism with information from the overlay network. Thus, a sending peer is able to shift traffic from a congested route to an uncongested one. This change in the rate allocation will be balanced by other peers in the overlay. Hence, the receiving peers experience no degradation of their total download rate.

Packet and flow level simulations of BitTorrent-like P2P networks

Peer-to-Peer (P2P) networks can reduce the distribution cost of large media files for the original provider of the data significantly. Thereby, the BitTorrent protocol is widely used in the Internet today. Most research work studies the protocol analytically, by simulations at the flow-level or real world experiments. For flow-level simulations the influence of neglecting packet-level characteristics (e.g. TCP) is not yet quantified. Therefore, this paper compares packet-level simulation results with flow-level values and analytically derived bounds. Our findings show that BitTorrent is near to optimal at flow-level for different scenarios. Naturally, packet-level results deviate more from the optimal values but differences are at most around 30% in our simulations. Furthermore, we show that the propagation delay can significantly influence the peer selection in BitTorrent and the download performance of the peers. Hence, the unchoking/choking algorithm in BitTorrent exploits implicitly network proximity.

Bandwidth Trading as Incentive

In P2P networks with multi-source download the file of interest is fragmented into pieces and peers exchange pieces with each other although they did not finish the download of the complete file. Peers can adopt different strategies to trade upload for download bandwidth. These trading schemes should give peers an incentive to contribute bandwidth to the P2P network. This chapter studies different trading schemes analytically and by simulations. A mathematical framework for bandwidth trading is introduced and two distributed algorithms, which are denoted as Resource Pricing and Reciprocal Rate Control, are derived. The algorithms are compared to the tit-for-tat principle in BitTorrent. Nash Equilibria and results from simulations of static and dynamic networks are presented. Additionally, we discuss how trading schemes can be combined with a piece selection algorithm to increase the availability of a full copy of the file. The chapter closes with an extension of the mathematical model which takes also the underlying IP network into account. This results in a TCP variant optimised for P2P content distribution.