... to flood Internet access and backbone ISPs with massive amounts of new traffic. We recently measured 200,000 IPTV users for a single program, receiving at an aggregate simultaneous rate of 100 gigabits/second. Although many architectures are possible for IPTV video distribution, several chunkdriven P2P architectures have been successfully deployed in the Internet. In order to gain insight into chunk-driven P2P IPTV systems and the traffic loads they place on ISPs, we have undertaken an in-depth measurement study of one of the most popular IPTV systems, namely, PPLive. We have developed a dedicated PPLive crawler, which enables us to study the global characteristics of the chunk-driven PPLive system. We have also collected extensive packet traces for various different measurement scenarios, including both campus access network and residential access networks. The measurement results obtained through these platforms bring important insights into IPTV user behavior, P2P IPTV traffic overhead and redundancy, peer partnership characteristics, P2P IPTV viewing quality, and P2P IPTV design principles.

A pervasive requirement of distributed systems is to deal with churn -- change in the set of participating nodes due to joins, graceful leaves, and failures. A high churn rate can increase costs or decrease service quality. This paper studies how to reduce churn by selecting which subset of a set of available nodes to use. First,

Distributed storage systems provide reliable access to data through redundancy spread over individually unreliable nodes. Application scenarios include data centers, peer-to-peer storage systems, and storage in wireless networks. Storing data using an erasure code, in fragments spread across nodes, requires less redundancy than simple replication for the same level of reliability. However, since fragments must be periodically replaced as nodes fail, a key question is how to generate encoded fragments in a distributed way while transferring as little data as possible across the network. For an erasure coded system, a common practice to repair from a node failure is for a new node to download subsets of data stored at a number of surviving nodes, reconstruct a lost coded block using the downloaded data, and store it at the new node. We show that this procedure is sub-optimal. We introduce the notion of regenerating codes, which allow a new node to download functions of the stored data from the surviving nodes. We show that regenerating codes can significantly reduce the repair bandwidth. Further, we show that there is a fundamental tradeoff between storage and repair bandwidth which we theoretically characterize using flow arguments on an appropriately constructed graph. By invoking constructive results in network coding, we introduce regenerating codes that can achieve any point in this optimal tradeoff. I.

With the increasing functionality and complexity of distributed systems, resource failures are inevitable. While numerous models and algorithms for dealing with failures exist, the lack of public trace data sets and tools have prevented meaningful comparisons. To facilitate the design, validation, and comparison of fault-tolerant models and algorithms, we have created the Failure Trace Archive (FTA) as an online public repository of availability traces taken from diverse parallel and distributed systems. Our main contributions in this study are the following. First, we describe the design of the archive, in particular the rationale of the standard FTA format, and the design of a toolbox that facilitates automated analysis of trace data sets. Second, applying the toolbox, we present a uniform comparative analysis with statistics and models of failures in nine distributed systems. Third, we show how different interpretations of these data sets can result in different conclusions. This emphasizes the critical need for the public availability of trace data and methods for their analysis. I.

Live peer-to-peer (P2P) streaming applications have been successfully deployed in the Internet. With relatively simple peer selection protocol design, modern live P2P streaming applications are able to provide millions of concurrent users adequately satisfying viewing experiences. That said, few existing research has provided sufficient insights on the time-varying internal characteristics of P2P topologies in live streaming. With 120 GB worth of traces in late 2006 from a commercial P2P live streaming system of UUSee Inc. in Beijing, this paper represents the first attempt in the research community to explore topological properties in practical P2P streaming, and how they behave over time. Starting from classical graph metrics, such as degree, clustering coefficient, and reciprocity, we explore and extend them in specific perspectives of streaming applications. We also compare our findings with existing insights from topological studies of P2P file sharing applications, which shed new and unique insights specific to streaming. Our characterization reveals the scalability of the commercial P2P streaming application even in case of large flash crowds, the clustering phenomenon of peers in each ISP, as well as the reciprocal behavior among peers, all of which play important roles in achieving its current success. 1

Abstract — The fundamental advantage of peer-to-peer (P2P) multimedia streaming applications is to leverage peer upload capacities to minimize bandwidth costs on dedicated streaming servers. The available bandwidth among peers is of pivotal importance to P2P streaming applications, especially as the number of peers in the streaming session reaches a very large scale. In this paper, we utilize more than 230 GB of traces collected from a commercial P2P streaming system, UUSee, over a four-month period of time. With such traces, we seek to thoroughly understand and characterize the achievable bandwidth of streaming flows among peers in large-scale real-world P2P live streaming sessions, in order to derive useful insights towards the improvement of current-generation P2P streaming protocols, such as peer selection. Using continuous traces over a long period of time, we explore evolutionary properties of inter-peer bandwidth. Focusing on representative snapshots of the entire topology at specific times, we investigate distributions of interpeer bandwidth in various peer ISP/area/type categories, and statistically test and model the deciding factors that cause the variance of such inter-peer bandwidth. Our original discoveries in this study include: (1) The ISPs that peers belong to are more correlated to inter-peer bandwidth than their geographic locations; (2) There exist excellent linear correlations between peer last-mile bandwidth availability and inter-peer bandwidth within the same ISP, and between a subset of ISPs as well; and (3) The evolution of inter-peer bandwidth between two ISPs exhibits daily variation patterns. Based on these insights, we design a throughput expectation index that facilitates highbandwidth peer selection without performing any measurements.

Contributory applications allow users to donate unused resources on their personal computers to a shared pool. Applications such as

Despite interest in structured peer-to-peer overlays and their scalability to millions of nodes, few, if any, overlays operate at that scale. This paper considers the distributed hash table extensions supported by modern BitTorrent clients, which implement a Kademlia-style structured overlay network among millions of BitTorrent users. As there are two disjoint Kademlia-based DHTs in use, we collected two weeks of traces from each DHT. We examine churn, reachability, latency, and liveness of nodes in these overlays, and identify a variety of problems, such as median lookup times of over a minute. We show that Kademlia’s choice of iterative routing and its lack of a preferential refresh of its local neighborhood cause correctness problems and poor performance. We also identify implementation bugs, design issues, and security concerns that limit the effectiveness of these DHTs and we offer possible solutions for their improvement. 1

Resource discovery is an important process for finding suitable nodes that satisfy application requirements in large loosely-coupled distributed systems. Besides inter-node heterogeneity, many of these systems also show a high degree of intra-node dynamism, so that selecting nodes based only on their recently observed resource capacities for scalability reasons can lead to poor deployment decisions resulting in application failures or migration overheads. In this paper, we propose the notion of a resource bundle— a representative resource usage distribution for a group of nodes with similar resource usage patterns—that employs two complementary techniques to overcome the limitations of existing techniques: resource usage histograms to provide statistical guarantees for resource capacities, and clustering-based resource aggregation to achieve scalability. Using trace-driven simulations and data analysis of a month-long PlanetLab trace, we show that resource bundles are able to provide high accuracy for statistical resource discovery (up to 56 % better precision than using only recent values), while achieving high scalability (up to 55% fewer messages than a non-aggregation algorithm). We also show that resource bundles are ideally suited for identifying group-level characteristics such as finding load hot spots and estimating total group capacity (within 8 % of actual values). 1.

The TCP/IP stack has been extremely successful for reliable delivery of best-effort, time insensitive elastic type data traffic. Nowadays, the Internet is rapidly evolving to become an equally efficient platform for multimedia content delivery. Key examples of this evolution are, to name few, YouTube, Skype Audio/Video, IPTV, P2P video distribution such as Coolstreaming or Joost. While YouTube streams videos using the Transmission Control Protocol (TCP), applications that are time-sensitive such as Skype VoIP or Video Conferencing employ the UDP because they can tolerate small loss percentages but not delays due to TCP recovery of losses via retransmissions. Since the UDP does not implement congestion control, these applications must implement those functionalities at the application layer in order to avoid congestion and preserve network stability. In this paper we investigate Skype Video in order to discover to what extent this application is able to throttle its sending rate to match the unpredictable Internet bandwidth while preserving resource for co-existing best-effort TCP traffic.