ISMA 2013 AIMS-5 - Workshop on Active Internet Measurements : Talk Abstracts

This page contains names, talk abstracts (if presenting), and topics the the participants are interested in discussing, as well as any related URLs. Participants are encouraged to read these ahead of time to anticipate workshop discussion.

Participant Abstracts

NameAbstract
Emile Aben (RIPE NCC) Talk Title: Hurricane Sandy, as seen by RIPE Atlas

Talk Abstract: This talk will outline the research we did into Hurricane Sandy, specifically what we saw from RIPE Atlas.

Interested in Discussing: measurement infrastructure, ipv6, outages, Internet-wide events

Džiugas Baltrūnas (Simula Research Laboratory) Talk Title: Mobile Broadband measurements: representing results in a database-oriented fashion

Talk Abstract: Mobile Broadband networks are measured using many different metrics such as round-trip time, one-way delay, throughput. Both latency and throughput can be measured by generating TCP, UDP or ICMP packets of varying size and packets may be sent at different intervals. Additionally, each measurement is unique in terms of a physical location of the terminal, measurement type, a list of assessed mobile networks, periodicity, duration and other characteristics. Not only each measurement iteration creates one or two rows of data such as delay in seconds or bitrate in bits per second, but it is also associated with the metadata that consists of context parameters (e.g. signal strength, cell id, RRC state, network mode and submode, etc.) and is valid only at one particular time.

Having a few hundred of measurement nodes each equipped with multiple modems connected to different mobile networks, maintaining measurements data becomes a challenging task. We will share our experience creating the infrastructure around Mobile Broadband measurements and present our solution how we use a traditional relational database to efficiently store and organize the measurements data. We will provide the structure of our tables, discuss about possible options reusing the same tables for different type of measurements and show how standard database features can be used for data partitioning and extracting the metadata-enriched results using simple SQL queries for manual analysis or visualization. We will also explain how metadata can be extracted from 3G USB modems and stored in a database table for easy correlation with the measurement results. We believe our experience can be in particular useful for those who are running or planning to run continuos Mobile Broadband measurements of a large scale.

Interested in Discussing: Large scale Mobile Broadband measurements, assessing the robustness and resilience in Mobile Broadband networks, organizing the data in relational databases.

Robert Beverly (Naval Postgraduate School) Talk Title: Inferring Internet Server IPv4 and IPv6 Address Relationships

Talk Abstract: While IPv6 is finally experiencing non-trivial deployment, IPv4 and IPv6 are expected to co-exist for the foreseeable future, implying dual-stacked devices, and protocol inter-dependence. We develop and deploy a system for characterizing the association between IPv4 and IPv6 addresses (''siblings'') within network server infrastructure, with specific focus on Internet DNS and web servers. We develop novel active and passive techniques for finding DNS resolver sibling groups, and find only 14% are one-to-one, primarily due to complex deployment and inter-relationship between IPv4 and IPv6. We then describe a targeted method to actively interrogate candidate (IPv4, IPv6) pairs to determine if they are assigned to the same device. We find that the IPv4 and IPv6 addresses of Internet servers frequently belong to different interfaces, machines, and even autonomous systems. Our results have important implications on network resilience, security, and performance measurement, as well as the evolution of IPv6.

Interested in Discussing: Topology, IPv6, spoofing

William Brinkmeyer (Naval Postgraduate School) Talk Title: IPv6 Alias Resolution via Induced Fragmentation

Talk Abstract: Discovering router-level IPv6 topologies is important amid the growth of IPv6 and desires to understand its evolution and relation to IPv4. This work presents a fingerprinting-based IPv6 alias resolution technique that relies on inducing fragmented responses from IPv6 router interfaces. We leverage the way in which IPv6 implements fragmentation to provide highly reliable inferences. We demonstrate perfect inference accuracy of our alias resolution tool in a controlled environment and on a small subset of the production IPv6 Internet for which we have ground- truth. We then perform Internet-wide testing where we find over 70% of IPv6 interfaces we probe respond to the test. Our promising initial results suggest a valuable new tool to aid IPv6 topology discovery.

Interested in Discussing: IPv6 large-scale Internet measurement, security concerns, applicability to Department of Defense, cyber security.

Marc-Olivier Buob (UPMC-LIP6) Talk Title: Improving AS relationship inference by measuring occurrences of multi-exit routing

Talk Abstract: Traditional AS relationship inference techniques built on the valley-free model use BGP-level measurements without considering the BGP decision process underneath, and they often converge to an arbitrary acceptable solution among many. We show that considering the BGP decision process can give additional insights into the ASes' routing policies. We are interested in cases when an AS routes traffic towards a given destination through several next hop ASes. Such evidences of so-called multi-exit routings can help us to formulate new constraints that could be injected into existing approaches to reduce the space of acceptable solutions. Both BGP and traceroute measurements prove to be complementary sources to discover multi-exit occurrences, and we propose a unifying framework to jointly study the both of them. Our preliminary study shows that the phenomenon is widespread in the Internet and suggests that our approach may be a promising way to extend AS relationship inference techniques.

Interested in Discussing:

  • Routing policy inference,
  • BGP routing,
  • annotations across data sources, hybrid BGP/traceroute instrumentation
  • annotating router-level, or AS-level topologies with AS type classification, business relationships, geolocation, customer cone size

Ang Chen (University of Pennsylvania) Talk Title: Efficient Analytics over Route Data

Talk Abstract: Contrasting multi-level routes (e.g., IP, subnet, AS levels) is an analytical primitive underpinning many applications, such as route asymmetry and/or diversity measurement, route change characterization, efficient route-tracing design, and others. We are the first to identify that current approaches incur redundant node comparisons, because they treat each level independently. We propose a new approach called rtd that eliminates the redundancy, therefore improves the analysis efficiency, by integrating all levels recursively. Our extensive evaluations on simulated traces and real data from Ark, FastMapping, and iPlane datasets show that rtd eliminates 85% comparisons on average and doubles the analysis speed. Finally, we design a route clustering application using rtd, and demonstrate how it aids the monitoring of an ISP transition.

Interested in Discussing: Big data analysis in Internet measurement research

David Choffnes (University of Washington) Talk Title: Toward an Open Platform for Participatory Mobile Measurement and Using the Middle to Meddle with Mobile

Interested in Discussing: mobile performance data

Greg Cole (GLORIAD/Center for International Networking) Talk Title: GLORIAD's New Measurement & Monitoring System

Talk Abstract: The Global Ring Network for Advanced Applications Development (GLORIAD) is a network-enabled community providing advanced collaborative services to scientists, educators and students around the world. Its evolution since the late 1990s will be briefly described - with a focus on core principles, architecture, services, public/private partners and future plans. Special attention will be placed on its decentralized model of operations and governance - and the case made for how this serves the purpose of fostering the continuous, flexible, community-driven innovation required for modern science infrastructure.

During the past year, GLORIAD has been working on a new system for measuring and monitoring global network infrastructure focused less on performance needs of individual customers across end-to-end flows. To accomplish its goal of actively improving global infrastructure for individual customers, the new system is designed to to (1) understand the network needs and requirements of a global customer base by actively studying utilization; (2) identify poor performance of individual applications by constantly (and in near-real-time) analyzing information on such per-flow metrics as load, packet loss, jitter and routing asymmetries; (3) mitigate poor performing applications by identifying weaknesses in the global fabric and actively sharing that information with global partners; (4) build visual analysis applications such as GLORIAD-Earth and the new GloTOP to help make sense of the enormous volume of data.

To realize this new model of measurement and monitoring (focused less on links and more on individual customers), GLORIAD has recently moved from its netflow-based system (used since 1998 and storing approximately 1 million records per day) to a new, much more detailed system - collecting, storing and analyzing over 400 million network utilization records per day - based on deployment of open-source Argus software (www.qosient.com/argus). The talk will focus on the benefits and the technical challenges of this new and actively evolving work.

Interested in Discussing: We are particularly interested in bi-directional flow on WANs.

Ann Cox (DHS) Talk Title: Internet Measurement and Attack Modeling Project

Talk Abstract: This talk will describe the ten projects funded under BAA 11-02 TTA 5,6, and 7 that are in the IMAM portfolio. I will also discuss the open BAA that is available to fund other projects of interest.

Casey Deccio (Sandia National Laboratories) Talk Title: Profiling Internet Spammers Using Passive Dual-stack Observation

Talk Abstract: One of the challenges facing deployment of IPv6 to production mail servers is the lack of IPv6 reputation data, whereby the servers can reject up-front incoming mail transfers from non-reputable hosts, as is done with DNS-based blacklists in IPv4. In this talk I will discuss work to measure the presence of IPv6 capabilities among hosts whose IPv4 addresses are blacklisted and use our findings to characterize these hosts. I will also discuss the potential to expand this experiment to allow a larger, collaborative testbed to obtain larger results.

Interested in Discussing: IPv6 - and I'd like to have feedback not only on the analysis we've performed thus far but also willing participants and thoughts on future direction and expansion.

Related URLs:

Ramakrishnan Durairajan (University of Wisconsin, Madison) Talk Title: Internet Atlas: A Geographic Database of the Physical Internet

Talk Abstract: This talk describes Internet Atlas, a new database and visualization portal of the physical interconnection structure of the Internet. Our goal is to create a comprehensive and geographically accurate catalogue and analysis environment for (i) the locations of points of presence (PoPs) that house switching and routing equipment, (ii) the conduits (links) that connect these locations, and (iii) relevant meta data (e.g., source provenance). Our work is based on using search to identify primary source data such as maps and other repositories of service provider network information. This data is then carefully entered into the database using a combination of manual and automated processes including consistency checks and methods for geocoding both PoP and link data. The repository currently contains over 8K PoP locations and nearly 13K links for over 185 networks (including all tier 1 providers) around the world. The openly available web portal is based on the widely-used ArcGIS geographic information system, which enables visualization and diverse spatial analyses of the data.

Interested in Discussing: Topology mapping, active measurements

Related URLs:

  • Internet Atlas, A Comprehensive Repository of the Physical Internet at Wisconsin Advanced Internet Laboratory (WAIL)
Ahmed Elmokashfi (Simula Research Laboratory ) Talk Title: Update on MBB measurements in Norway

Talk Abstract: MBB networks are increasingly becoming the medium of choice to access the Internet, thanks to their increasing speed, pervasiveness, and the availability of scores of powerful smartphones and tablets. Despite their rapid adoption, various aspects of MBB performance and robustness have not been fully understood yet. A major hurdle in this respect is the lack of large scale measurement infrastructures. To scale this hurdle, we started back in 2010 with building an infrastructure of 90 measurement hosts in 10 municipalities across Norway to actively measure the performance of three MBB operators. Armed with findings and learned lessons, we are now moving to the second phase of the project. We are deploying more nodes, covering larger geographic spread, measuring the performance of five MBB networks, and using carefully designed measurement boxes. To contextualize and ease the task of interpreting our measurements, we now have more focus on annotating them with meta-data. We use various annotations including operator, geo-location, cellular cell-id, network modes and submodes, and RRC states. In this talk, I will give an update on the current status of our MBB measurement deployment. I will also present, how we are making our measurements more accessible through augmented metadata.

Interested in Discussing: IPv6 deployment, MBB, big data retention

Related URLs:

Nick Feamster (Georgia Tech) Talk Title: Exposing Inconsistent Web Search Results with Bobble

Talk Abstract: Personalized Web search can potentially provide users with search results that are tailored to their geography, the device from which they are searching, and a variety of other preferences and predis- positions. Although most major search engines employ some type of personalization, the algorithms used to implement this person- alization remain a "black box" to users, who are not aware of the effects of these personalization algorithms on the results that they ultimately see. Indeed, many users may be unaware that such per- sonalization is taking place at all. This papers take a first look at the nature of inconsistent search results that result from location- based personalization and search history. We present the design and implementation of Bobble, a tool that executes a single user query from a variety of different vantage points and under a range of different conditions and comparsed the consistency of the results that are returned from each query. Using more than 75,000 search queries from about 175 users over a nine-month period, we explore the nature of inconsistencies that arise in different search categories and regions and find that 98% of all Google search queries from Bobble users resulted in some inconsistency, and that geography is more important than search history in influencing the nature of the inconsistency.

Joint work with Xinyu Xing, Wei Meng, Dan Doozan, Wenke Lee, and Alex C. Snoeren.

Interested in Discussing: Censorship.

Related URLs:

Phillipa Gill (The Citizen Lab/Stony Brook University) Talk Title: Characterizing Global Web Censorship: Why is it so hard?

Talk Abstract: I've been working with a group that has been performing measurements of Internet censorship for the past 5 years. I will present challenges in collecting this type of data (gaining access to realistic vantage points in the country, informed consent from participants) as well as general challenges they've faced managing longitudinal data.

Interested in Discussing: interdomain routing measurements (especially issues around IXPs/settlement-free peering). The intersection of politics and Internet measurement.

James Grace (Florida International University) Talk Title: Technology Augmented Research

Talk Abstract: The presentation will focus on the goals of the AmLight project to operationalize the use of passive and active monitors, and performance measurement tools to measure performance and flows on the AmLight links.

The Coral Reef implementation will be discussed. We will show a diagram of the deployment. We will describe the project we have with CAIDA to enhance Coral Reef to represent vlans. We will mention AMPATH has deployed a DYNES controller that uses the OSCARS protocol, and its IDC is peering with Brazil's IDC and Internet2 ION. This is in the context of hybrid networking services.

Interested in Discussing: Technology Augmented Research

Sarthak Grover (Georgia Tech) Talk Title: End-to-end Routing Behavior in the Internet: A Re-Appraisal from Access Networks

Talk Abstract: Recent testing and reports on ISP and broadband performance by FCC show an overall improvement in delivery of advertised performance, but the effects of traffic engineering policies and practices on end users and end-to-end application performance remains largely unknown. A first step in such a study is understanding how these wide-area routes change, as a result of traffic engineering practices. As part of this study, we analyze more than 250,000 traceroutes from October-November 2012 from access networks to a variety of Internet destinations and re-examine these paths in terms of "Paxon's persistence" and "prevalence" metrics. We find there is often not a single prevalent path from access networks to many Internet destinations, and that these paths are often not persistent.

Interested in Discussing: Improved Internet measurement techniques and tools, Data analysis tools for handling and storing such real-time measurement data

Mehmet Gunes (University of Nevada, Reno) Talk Title: Cheleby Internet Topology Mapping System

Talk Abstract: In this talk, I will present Cheleby Internet mapping system that takes the snapshots of the underlying topology. Cheleby utilizes efficient algorithms to process large scale data-sets collected from various vantage points and provide topology graphs at link-layer. In particular, we focus on subnet-level analysis and annotation of the graph.

Interested in Discussing: Internet topology sampling/modeling

Yihua Guo (University of Michigan) Talk Title: Throughput prediction based on mobile device context in Cellular Network

Talk Abstract: With the prevalence of cellular networks, e.g. 3G and LTE, and mobile devices, it is significant to understand the performance and power characteristics through active measurement. One important aspect of performance is the throughput, which affects energy consumption and user experience. Since the context of mobile devices, such as signal strength and TCP connection state, can affect the TCP throughput, we perform measurements to explore the correlation between throughput and device context. We would like to use the light-weight information instead of active probing to predict throughput based on existed measurement data, which provides indication on scheduling network data transfer and reducing energy consumption on mobile devices.

Interested in Discussing: Throughput prediction based on mobile device context in Cellular Network

John Heidemann (USC/Information Sciences Institute) Talk Title: Long-term analysis of outages at the Internet edge

Talk Abstract: This talk will discuss our approach to network outage detection based on analysis of active probing of internet edge networks via ping. We will discuss collection and analysis of long-term, general purpose datasets (approximately quarterly since 2006), and the role of long-term datasets to developing adaptive probing methods specific to this problem.

Interested in Discussing: network outages, dataset curation and archival, shared datasets

Related URLs:

Julio Ibarra (Florida International University)

Interested in Discussing: Technology Augmented Research

Ethan Katz-Bassett (University of Southern California) Talk Title: Active BGP measurement to understand path changes

Talk Abstract: I can give a short talk on one or two of the things we've used the Mux for. It doesn't have a long-term measurement / dataset aspect, but I think it is interesting in that it is active BGP measurements (rather than the typical passive RouteViews stuff). I'll talk about our Sigcomm 2012 paper (LIFEGUARD - locating and avoiding Internet failures)) and maybe touch on the Sigcomm submission we are working on right now (locating the root cause of BGP path changes).

Interested in Discussing: mobile measurements

Related URLs:

  • LIFEGUARD examples: LIFEGUARD has 2 parts: (1) locate failures and (2) route around them. I'm not going to talk much about (1), and we don't discuss it much in the paper either. Dave talked about preliminary work towards (1) at AIMS a few years ago, I think, although I wasn't there. Here are some examples of (1) that will hopefully convince people that we can do it well enough to serve as the basis for (2)
Robert Kisteleki (RIPE NCC) Talk Title: (Provisional) Data storage at the RIPE NCC

Talk Abstract: RIPE NCC is using a Hadoop/HBase based infrastructure to store all data behind RIPE Atlas (about 50 million results a day) and RIPEstat (including all data from RIS, the Routing Information Service, and besides other sources).

Interested in Discussing: Measurement data storage and retreival, time series, batch and real-time processing.

Sándor Laki (Eötvös Loránd University) Talk Title: Analysing the spatial structure of the Internet topology

Talk Abstract: The geographic layout of the physical Internet inherently determines important network properties and traffic characteristics. In this talk, I focus on the spatial properties of the Internet topology and routing. To represent the network we conducted a geographically dispersed traceroute campaign, and embedded the extracted topology into the geographic space by applying a novel IP geolocalization service, called Spotter. Our investigations include the characterization of the length distribution of Internet links and the quantification of the circuitousness and asymmetry of end-to-end Internet routes. We also analyzed the direction-dependence of geographic deviations and a description of router density in terms of the geographic layout of end-to-end paths has been given. Our intention with the presented study is to shed light on previously hidden structural aspects of the Internet.

Interested in Discussing: geolocation, topology measurements and discovery, routing

Related URLs:

Dan Massey (Colorado State University) Talk Title: Combining Active Route Measurements with Passive BGP Data

Talk Abstract: There is considerable work on monitoring Internet routes and the results can have implications on topology, routing security, censorship, and network management. This talk has two key objectives; an update on BGP data collection services and a discussion of combining BGP data collection with active measurements.

First, we would like to update the community on progress in Internet route monitoring. This progress includes the creation of new route collector locations which will be in place by the workshop, new tools in CPAN for processing BGP data, and new services such as ROVER that provide additional BGP information. The goal is both to raise awareness of new services and too solicit community input on how the services should evolve. The second objective is to discuss preliminary work on combining passive BGP measurements with active measurement systems and discuss early results and directions from recently funded work that includes both data traffic capture and BGP monitoring. Results include well developed work on fast and cost effective packet forwarding, bonnet identification, and analysis of IP address allocations. Again, the goal is both report results and solicit feedback and/or collaboration on future directions.

Interested in Discussing: IP address allocations, routing, forwarding, security.

Related URLs:

  • BGPmon on CPAN - BGPmon-core is a collection of Perl functions that are useful for obtaining and parsing BGPmon XML messages.
Riad Mazloum (UPMC Sorbonne Universités) Talk Title: Improving AS relationship inference by measuring occurrences of multi-exit routings

Talk Abstract: Traditional AS relationship inference techniques built on the valley-free model use BGP-level measurements without considering the BGP decision process underneath, and they often converge to an arbitrary acceptable solution among many. We show that considering the BGP decision process can give additional insights into the ASes' routing policies. We are interested in cases when an AS routes traffic towards a given destination through several next hop ASes. Such evidences of so-called multi-exit routings can help us to formulate new constraints that could be injected into existing approaches to reduce the space of acceptable solutions. Both BGP and traceroute measurements prove to be complementary sources to discover multi-exit occurrences, and we propose a unifying framework to jointly study the both of them. Our preliminary study shows that the phenomenon is widespread in the Internet and suggests that our approach may be a promising way to extend AS relationship inference techniques.

Interested in Discussing:

  • Routing policy inference,
  • BGP routing,
  • annotations across data sources, hybrid BGP/traceroute instrumentation
  • annotating router-level, or AS-level topologies with AS type classification, business relationships, geolocation, customer cone size

Sachit Muckaden (Georgia Institute of Technology) Talk Title: MySpeedTest: Active and Passive Measurements of Cellular Network Performance

Talk Abstract: As the number and diversity of applications available to mobile users increases, there is an increasing need for developers, network service providers, and users to understand how users perceive the network performance of these applications. My SpeedTest is a measurement tool that actively probes the network to determine not only TCP throughput and round trip time, but also the proximity to popular content providers, IP packet delay variation, and loss. It also records other metadata that could affect user experience, such as signal strength, service provider, connection type, battery state, device type, manufacturer, time of day, and location. The tool also takes passive measurements of the applications installed on the device and the network usage of these applications. My SpeedTest is available on the Google Play Store and currently has 600+ active users.

We present the effect of metrics like latency, throughput, loss and IP packet delay variation on performance. We also attempt to infer the effect of these metrics on user behavior on different kinds of applications, thereby helping developers and network service providers identify the key quality metrics and areas for improvement.

Interested in Discussing: Challenges with validating measurements.

Eric Osterweil (Verisign Labs)

Interested in Discussing: DNS and/or routing stability

Aaron Schulman (University of Maryland)

Interested in Discussing: Why do residential Internet connections fail?

Srikanth Sundaresan (Georgia Tech) Talk Title: Localizing Performance Bottlenecks in Home Networks + State of Project BISmark

Talk Abstract: I will give an update on the state of Project BISmark, a platform to measure home networks. As part of a related project, we develop techniques to characterize and identify performance bottlenecks in homes using measurements from the access point. The goal is to localize the bottleneck to the access link, the wireless network, or the application. I will also present early results from a deployment of these techniques in 20+ homes.

Interested in Discussing: Mobile and broadband networks

Related URLs:

  • Project BISmark is a project led by Georgia Tech and the University of Napoli Federico II to develop an OpenWRT-based platform for performing measurements of ISP performance, as well as traffic inside the home.
  • Network Dashboard is a portal designed to help home users visualize properties of their home Internet connections.
John-Paul Verkamp (Indiana University) Talk Title: Internet measurement and cenorship

Talk Abstract: I will be speaking about ongoing research involving active and passive measurements of DNS, BGP, and other data and how they relate to censorship.

Interested in Discussing: Censorship. Regarding this workshop, how AS-level topology and passive DNS data relate.

Related URLs:

Nicholas Weaver (ICSI) Talk Title: ICSI Updates: Netalyzr

Talk Abstract: Netalyzr continues to be developed and improved. Three particular areas will be discussed in this talk

  1. Developing a mobile version of Netalyzr, targeting the Android platform. This version is a full version of Netalyzr, with complete functionality, not a stripped down version.
  2. Proxy detection and analysis, including server->client proxy traceroute functionality.
  3. DNSSEC evaluation

Interested in Discussing: Mobile

Related URLs:

  • A Netalyzr execution from the command line client on my home connection - This shows what we currently test.
  • The client-side transcript - This shows how the tests work.
  • The JSON representation - This shows the framework others can use to embed Netalyzr in their own projects.
Yuanyuan Zhou (University of Michigan) Talk Title: Mobiperf

Talk Abstract: MobiPerf is an open source application for measuring network performance on mobile platforms. You can measure your network's throughput and latency, as well as other useful network metrics. MobiPerf also performs measurements at regular intervals in the background. The data is collected either anonymously or from your selected account, which allows you to see your own data. The user credentials collected are not shared outside of this site, and any data used in research projects in universities are anonymized before use.

Interested in Discussing: Network measurement on mobile devices

Published
Last Modified