Understanding the Impact of Network Infrastructure Changes using - - PowerPoint PPT Presentation

Understanding the Impact of Network Infrastructure Changes using Large-Scale Measurement Platforms

Computer Networks and Distributed Systems Jacobs University Bremen Bremen, Germany

June 2013 AIMS 2013, Barcelona Vaibhav Bajpai and Jürgen Schönwälder

{v.bajpai, j.schoenwaelder}@jacobs-university.de

Supported by: Leone Project: http://leone-project.eu

Introduction

[2/17]

• Large-Scale Broadband Measurement Use Case
• Internet Service Provider (ISP)
• Identify, isolate and fix problems in the access network.
• Evaluate the Quality of Experience (QoE) of the user.
• Benchmark and look into competitor insights.
• Consumers
• Does the ISP service adhere to the service level agreements (SLA)s?
• Diagnose impaired components in the private network.
• Regulators
• Need datasets to compare multiple broadband providers.
• Frame better policies to help regulate the broadband industry:

http://www.fcc.gov/measuring-broadband-america http://maps.ofcom.org.uk/broadband

[draft-linsner-lmap-use-cases-02].

State of the Art

[3/17]

• Early Studies
• Inject packet trains to infer broadband link characteristics [Dischinger-IMC-2007].
• Large-Scale Measurement Platforms:
• SamKnows and BISmark
• RIPE Atlas
• Google’s Measurement Lab (M-Lab)
• CAIDA’s Archipelago (Ark)

http://www.samknows.com http://atlas.ripe.net http://www.measurementlab.net http://www.caida.org/projects/ark

• Software-based Solutions:
• Speedtest.net a flash-tool to measure broadband throughput: http://www.speedtest.net.
• DIMES, a software agent that performs ping and traceroute measurements [Shavitt-CCR-2005].
• Glasnost, a Java-based applet that detects ISP-enforced traffic shaping [Dischinger-NSDI-2010].
• Netalyzr, a Java-based applet that performs DNS, NAT, HTTP

, IPv6-based tests. [Kreibich-IMC-2010].

• Fathom, a Firefox-extension to Netalyzr [Dhawan-IMC-2012].

State of the Art

• LMAP and IPPM Standardization
• Regulatory Implications
• Standards body collaboration: IETF + BBF + IEEE
• Registry for commonly-used metrics

[draft-bagnulo-ippm-new-registry-00] [draft-bagnulo-ippm-new-registry-independent-00]

MA

Measurement Test Traffic IPPM Framework [RFC 2330]

• IP Performance Metrics (IPPM) charter revision.
• Control and Report Protocol candidates

[draft-schoenw-lmap-netconf-00] [draft-bagnulo-lmap-ipfix-01] [draft-seedorf-lmap-lmap-alto-00]

• Data Model candidates

[draft-schoenw-lmap-yang-00]

Controller Collector

Control Protocol Report Protocol LMAP Framework [draft-eardley-lmap-terminology-01] [draft-eardley-lmap-framework-01]

• Large Scale Measurement of Access Network Performance (LMAP) Birds of a Feather (BOF) at IETF 86.

MA

Research Statement

[5/17]

• Understanding the Impact of Network Infrastructure Changes
• Measuring broadband performance from residential gateway.
• Helping regulators sketch better policy decisions.

using Large-Scale Measurement Platforms

Goals of Earlier Studies Extending the Goal

• Understanding the Impact of Network Infrastructure Changes

using Large-Scale Measurement Platforms

• Can we identify a Carrier-Grade NAT (CGNAT) from a residential gateway?
• Can we identify multiple layers of NATs from a residential gateway?
• Study IPv6 transition [Bajpai-AIMS-2012].
• Measure IPv6 adoption? [Dhamdhere-IMC-2012] [Allman-SIGMETRICS-2013] [Colitti-PAM-2010]

http://www.google.com/ipv6/statistics.html http://bgp.he.net/ipv6-progress-report.cgi

• Measure today’s IPv6 network.
• To what extend do web services centralize on Content Delivery Networks (CDNs)?
• To what extend does web experience depend on Regionalization?
• Study the blend of network centralization and decentralization
• How does the performance of IPv6 compare to that of IPv4?

Approach

[6/17]

• Requirements?
• Access to a large-scale measurement platform.
• Address allocations from Regional Internet Registries (RIR).
• Publicly available BGP data from route collectors.
• SamKnows and Jacobs University are partners of the Leone Consortium http://www.leone-project.eu.
• Work Flow
• Define metrics targeted to our research questions.
• Implement measurement tests that adhere to the metric definition.
• Deploy measurement tests on a large-scale measurement platform.
• Conglomerate measurement results from multiple Measurement Agents (MA)s.
• Correlate measurement results with data from RIRs and route collectors.
• Prepare data analysis tools that can mine this multidimensional data.
• Uncover the insights to answer the research questions.

Preliminary Work | Measuring Happy Eyeballs

[7/17]

1) native IPv6 routes ... 2) native IPv4 routes ... 3) IPv4-IPv6 Transitioning routes getaddrinfo(...) preference: TCP connection request

• getaddrinfo(...) behavior:
• The order is dictated by [RFC 6724] and /etc/gai.conf
• If the IPv6 connectivity is broken, an application remains unresponsive in the order of seconds.
• Returns a list of endpoints in an order that prioritizes IPv6-upgrade path.

Preliminary Work | Measuring Happy Eyeballs

[8/17]

• Happy Eyeballs Algorithm [RFC 6555]:
• Switch over with a TCP connect(...) to a different address family otherwise.
• The competition runs fair after 300ms.
• Initiate a TCP connect(...) with the first endpoint, give it 300ms.

t0 t0 + 300ms

time IPv6 IPv4 Happy Eyeballs [RFC 6555]

Preliminary Work | Measuring Happy Eyeballs

[9/17]

• Metrics and Implementation
• Uses getaddrinfo(...) to resolve service names.
• Uses non-blocking TCP connect(...) calls.

>> ./happy -q 1 -m www.google.com www.facebook.com HAPPY.0;1360681039;OK;www.google.com;80;173.194.69.105;8626 HAPPY.0;1360681039;OK;www.google.com;80;2a00:1450:4008:c01::69;8884 HAPPY.0;1360681039;OK;www.facebook.com;80;2a03:2880:10:6f01:face:b00c::8;170855 HAPPY.0;1360681039;OK;www.facebook.com;80;31.13.72.39;26665

• Applies a delay between connect(...) to avoid SYN floods.
• Service name resolution time is not accounted.
• Capability to produce both human-readable and CSV output.
• File locking capability.
• Capability to read multiple service names as arguments.
• Capability to read service names list from a file.
• Cross-compiled for OpenWrt platform. Currently running from SamKnows probes.

http://happy.vaibhavbajpai.com

happy 1) endpoint 2) endpoint 3) endpoint ... n) endpoint connection establishment times (µs) 1) service name 2) port

Preliminary Work | Measuring Happy Eyeballs

[10/17]

• How to compile a dual-stacked service names list?
• Hurricane Electric (HE) maintains a top 100 dual-stacked service names list.

http://bgp.he.net/ipv6-progress-report.cgi

• HE uses top 1M service names list from Alexa Top Sites (ATS).
• HE does not follow CNAMES.
• Prepared a custom top 100 dual-stacked service names list.
• Amazon has made the ATS top 1M service names list public.

http://s3.amazonaws.com/alexa-static/top-1m.csv.zip

• Explicitly follow CNAMES.
• Prepend a www to each service name and cross-check any AAAA response.

Preliminary Work | Measuring Happy Eyeballs

[11/17]

• From where to run the measurement test?

Provider (IPv4, IPv6) Location Platform

(dfn, AS680), (-)

Jacobs University Bremen SamKnows

(Kabel Deutschland, AS31334), (HE, AS6939)

Bremen SamKnows

(Gaertner Datensystems GmbH, AS24956), (-)

Braunchsweig SamKnows

(Deutsche Telekom AG, AS3320), (-)

Bremen SamKnows

(British Sky Broadcasting Limited, AS5607), (-)

London SamKnows

(Telekom Italia, AS3269), (-)

Torino SamKnows

(BT Spain, AS8903), (-)

Madrid SamKnows

(ROEDUNET, AS2614), (-)

Timisoara SamKnows

(LambdaNet Communications, AS13237), (Teredo)

Berlin GNU/Linux

(dfn, AS680), (-)

Jacobs University Bremen Mac OS X

(-) means the IPv6 provider and AS are same as that for IPv4.

Preliminary Work | Measuring Happy Eyeballs

[12/17]

• How does the performance (mean) of IPv6 compare to that of IPv4?
• Native IPv4 and IPv6 connectivity via DTAG - Deutsche Telekom AG [AS 3320]

Preliminary Work | Measuring Happy Eyeballs

• How does the performance (variation) of IPv6 compare to that of IPv4?
• Native IPv4 and IPv6 connectivity via DTAG - Deutsche Telekom AG [AS 3320]

[13/17]

Preliminary Work | Measuring Happy Eyeballs

• Do major portion of the web services centralize on CDNs?

IPv4 Aggregation Cloud

www.google.com.br Google Ireland Limited 2a00:1450::/29 www.flipkart.com Flipkart India Pvt Ltd 2001:df0:23e::/48 www.google.com.sg www.google.com.sa Internet Assigned Numbers Authority /0 www.goo.gl Akamai Technologies 2a02:26f0::/32 www.google.by www.google.co.za www.google.be www.google.sk www.google.bg www.google.com.bd www.google.ae Mozilla Corporation 2620:101:8000::/40 www.google.se Facebook Ireland Ltd 2a03:2880::/32 www.youtube.com www.aol.com America Online 2001:4b0::/32 www.google.fi SoftLayer Technologies Inc. 2607:f0d0::/32 www.att.com www.orkut.com www.google.co.id Magyar Telekom plc. 2001:4c48::/29 www.google.co.in www.google.co.il www.balagana.net www.google.co.ma www.google.dz www.blogspot.jp www.google.fr www.google.nl www.google.no www.google.co.ve www.google.com.ly www.google.com.mx VNET s. r. o. 2a01:390::/32 BUL.NET 2a01:9e40:195::/48 www.google.ch www.blogspot.co.uk www.google.cn www.youm7.com 665 Third Street 2400:cb00::/32 www.google.com.vn www.mapquest.com www.google.ca www.blogspot.com.es www.blogger.com www.rtl.de RTL Interactive Frankfurt 2a03:d680::/48 www.google.co.nz www.bitsnoop.com 2a02:29b8:1925::/64 www.blogspot.in www.softlayer.com www.google.ro www.yify-torrents.com COOLHOUSING s.r.o. 2a01:5f0::/32 www.google.co.jp www.google.ru www.comcast.net Akamai Technologies 2a02:26f0:5::/48 www.facebook.com www.google.cl www.google.kz www.google.gr www.blogspot.com www.google.cz www.google.com.hk www.google.com.ua www.google.de www.google.dk www.google.com.ec www.android.com www.google.com.eg www.google.co.th EdgeCast Networks, Inc. 2606:2800::/32 www.google.co.kr www.google.lk www.google.tn www.google.hr www.bing.com www.google.co.uk www.google.com.au www.netflix.com Amazon Data Services Ireland LTD 2a01:578::/32 www.google.lt www.blogspot.kr www.google.com.tr www.google.es 2607:f0d0:3001:ae::/64 www.google.az www.gravatar.com www.google.at www.sprint.com Sprint 2600::/29 www.google.com.ph www.blogspot.fr www.google.com.pk www.networkedblogs.com www.google.com.pe www.google.com.co www.mozilla.org www.irs.gov www.google.ie www.google.pl www.autoblog.com www.google.com www.anitube.jp www.google.com.do www.google.it www.google.co.hu www.googleusercontent.com www.google.iq www.brainyquote.com www.google.pt

IPv6 Aggregation Cloud

[14/17]

Preliminary Work | Measuring Happy Eyeballs

• To what extend is IPv6 preferred when connecting to a dual-stacked service?

Native IPv4 and IPv6 connectivity via DTAG - Deutsche Telekom AG [AS 3320] IPv4 connectivity via LambdaNet Communications [AS 13237]. IPv6 connectivity via Teredo.

[15/17]

Preliminary Work | Measuring Happy Eyeballs

• Data Analysis Insights
• A 300ms advantage leaves a MA 1% chance to prefer IPv4.
• Measurement Agent (MA) will never use Teredo IPv6 unless IPv4 connectivity is broken.
• Higher connection times and variations over IPv6.
• A number of disparate services (bing, comcast, irs) show similar performances.
• whois data reveals they resolve to same RIR allocated blocks owned by a CDN.
• IPv4 and IPv6 whois aggregation clouds reveal many services centralize at Google and Akamai CDNs.

[16/17]

Understanding the Impact of Network Infrastructure Changes using Large-Scale Measurement Platforms

Research Questions

• Study IPv6 transition [Bajpai-AIMS-2012].
• Measure today’s IPv6 network.
• Study the blend of network centralization and decentralization
• Can we identify a Carrier-Grade NAT (CGNAT) from a residential gateway?
• Can we identify multiple layers of NATs from a residential gateway?
• How does the performance of IPv6 compare to that of IPv4?
• To what extend do web services centralize on Content Delivery Networks (CDNs)?
• To what extend does web experience depend on Regionalization?

Dissemination:

[17/17]

• Technical Article: Evaluating the Effectiveness of Happy Eyeballs, RIPE Labs, June 2013:

https://labs.ripe.net/Members/vaibhav_bajpai/evaluating-the-effectiveness-of-happy-eyeballs

• Publication: PhD Workshop Paper, AIMS, June 2013
• Tutorial: Large Scale Measurement Platforms, AIMS, June 2013
• Invited Talk: Measuring the Effectiveness of Happy Eyeballs, RIPE 66, May 2013: http://ripe66.ripe.net/archives/video/1208/

References

[1] M. Dischinger, et al., Characterizing Residential Broadband Networks, ACM Conference on Internet Measurement Conference (IMC), 2007. [2]

• Y. Shavitt, et al., DIMES: Let the Internet Measure Itself, ACM

Computer Communications Review (CCR), 2005. [3] M. Dischinger, et al., Glasnost: Enabling End Users to Detect Traffic Differentiation, USENIX Symposium on Networked Systems Design and Implementation (NSDI), 2010 [4] C. Kreibich, et al., Netalyzr: Illuminating the Edge Network, ACM Conference on Internet Measurement Conference (IMC), 2010 [5] M. Dhawan, et al., Fathom: A Browser-based Network Measurement Platform, ACM Conference on Internet Measurement Conference (IMC), 2012

References

[6]

• V. Bajpai, et al., Flow-based Identification of Failures caused by IPv6 Transition Mechanisms, 6th

Conference on Autonomous Infrastructure, Management and Security (AIMS), 2012 [7] A. Dhamdhere, et al., Measuring the Deployment of IPv6: Topology, Routing and Performance, ACM Conference on Internet Measurement Conference (IMC), 2012 [8] M. Allman, et al., Accessing IPv6 Adoption, ACM Special Interest Group (SIG) for Computer Systems Performance Evaluation (SIGMETRICS), 2013 [9] L. Colitti, et al., Evaluating IPv6 Adoption in the Internet, Proceedings of Passive and Active Measurements (PAM), 2010