Poor visibility into the network hampers progress in a number of important research areas, from network troubleshooting to Internet topology and performance mapping. This persistent, well-known problem has served as motivation for numerous proposals to build or extend existing Internet measurement platforms by recruiting larger, more diverse vantage points. Capturing the edge of the network, however, remains an elusive goal. We argue that at its root the problem is one of incentives. Today's measurement platforms build on the assumption that the goals of experimenters and those hosting the platform are the same. As much of the Internet growth occurs in residential broadband networks, this assumption no longer holds. We present a measurement experimentation platform that reaches the network edge by explicitly aligning the objectives of the experimenters with those of the users hosting the platform. Dasu-our current prototype-is designed to support both network measurement experimentation and broadband characterization. Dasu has been publicly available since July 2010 and has been installed by over 100 000 users with a heterogeneous set of connections spreading across 2431 autonomous systems (ASs) and 166 countries. We discuss some of the challenges we faced building and using a platform for the Internet's edge, describe its design and implementation, and illustrate the unique perspective its current deployment brings to Internet measurement.

Though the impact of file-sharing of copyrighted content has been discussed for over a decade, only in the past few years have countries begun to adopt legislation to criminalize this behavior. These laws impose penalties ranging from warnings and monetary fines to disconnecting Internet service. While their supporters are quick to point out trends showing the efficacy of these laws at reducing use of file-sharing sites, their analyses rely on brief snapshots of activity that cannot reveal long- and short-term trends.

People use P2P systems such as BitTorrent to share an unprecedented variety and amount of content with others around the world. The random connection pattern used by BitTorrent has been shown to result in reduced performance for users and costly cross-ISP traffic. Although several client-side systems have been proposed to improve the locality of BitTorrent traffic, their effectiveness is limited by the availability of local peers. We show that sufficient locality is present in swarms - if one looks at the right time. We find that 50% of ISPs have at least five local peers online during the ISP's peak hour, typically in the evening, compared to only 20% of ISPs during the median hour. To better discover these local peers, we show how to increase the overall peer discovery rate by over two orders of magnitude using client-side techniques: leveraging additional trackers, requesting more peers per sample, and sampling more frequently. We propose an approach to predict future availability of local peers based on observed diurnal patterns. This approach enables peers to selectively apply these techniques to minimize undue load on trackers.

Content Delivery Networks (CDNs) rely on the Domain Name System (DNS) for replica server selection. DNS-based server selection builds on the assumption that, in the absence of information about the client's actual network location, the location of a client's DNS resolver provides a good approximation. The recent growth of remote DNS services breaks this assumption and can negatively impact client's web performance.

The Domain Name System (DNS) is a fundamental component of today's Internet. Recent years have seen radical changes to DNS with increases in usage of remote DNS and public DNS services such as OpenDNS. Given the close relationship between DNS and Content Delivery Networks (CDNs) and the pervasive use of CDNs by many popular applications including web browsing and real-time entertainment services, it is important to understand the impact of remote and public DNS services on users' overall experience on the Web. This work presents a tool, namehelp, which comparatively evaluates DNS services in terms of the web performance they provide, and implements an end-host solution to address the performance impact of remote DNS on CDNs. The demonstration will show the functionality of namehelp with online results for its performance improvements.

A thorough understanding of the network impact of emerging large-scale distributed systems -- where traffic flows and what it costs -- must encompass users' behavior, the traffic they generate and the topology over which that traffic flows. In the case of BitTorrent, however, previous studies have been limited by narrow perspectives that restrict such analysis.

Evaluating and characterizing access ISPs is critical to consumers shopping for alternative services and governments surveying the availability of broadband services to their citizens. We present Dasu, a service for crowdsourcing ISP characterization to the edge of the network. Dasu is implemented as an extension to a popular BitTorrent client and has been available since July 2010. While the prototype uses BitTorrent as its host application, its design is agnostic to the particular host application. The demo showcases our current implementation using both a prerecorded execution trace and a live run.

Vehicular networks are emerging as a new distributed system environment with myriad possible applications. Most studies on vehicular networks are carried out via simulation, given the logistical and economical problems with large-scale deployments. This paper investigates the impact of realistic radio propagation settings on the evaluation of VANET-based systems. Using a set of instrumented cars, we collected IEEE 802.11b signal propagation measurements between vehicles in a variety of urban and suburban environments. We found that signal propagation between vehicles varies in different settings, especially between line-of-sight ("down the block") and non line-of-sight ("around the corner") communication in the same setting. Using a probabilistic shadowing model, we evaluate the impact of different parameter settings on the performance of an epidemic data dissemination protocol and discuss the implications of our findings. We also suggest a variation of a basic signal propagation model that incorporates additional realism without sacrificing scalability by taking advantage of environmental information, including node locations and street information.

We consider the problem of data dissemination in vehicular networks. Our main goal is to compare the application-level performance of fully distributed and centralized data dissemination approaches in the context of traffic advisory systems. Vehicular networks are emerging as a new distributed system environment with myriad promising applications. Wirelessly-connected, GPS-equipped vehicles can be used, for instance, as probes for traffic advisory or pavement condition information services with significant improvements in cost, coverage and accuracy. There is an ongoing discussion on the pros and cons of alternative approaches to data distribution for these applications. Proposed centralized, or infrastructure-based, models rely on road-side equipment to upload information to a central location for later use. Distributed approaches take advantage of the direct exchanges between participating vehicles to achieve higher scalability at the potential cost of data consistency. While distributed solutions can significantly reduce infrastructures' deployment and maintenance costs, it is unclear what the impact of "imprecise" information is to an application or what level of adoption is needed for this model to be effective. This paper investigates the inherent trade-offs in the adoption of distributed or centralized approaches to a traffic advisory service, a commonly proposed application. We based our analysis on a measurements study of signal propagation in urban settings and an extensive simulation-based experimentation in the Chicago road network.