Failures, Latency and Happy Eyeballs Contributions Takeaway - - PowerPoint PPT Presentation

▶

Dec 08, 2023 474 likes •649 views

Motivation Vaibhav Bajpai March 20, 2019 https://doi.org/10.1109/TNET.2019.2895165 March 2019 IEEE/ACM Transactions on Networking Garching bei Mnchen, Germany NetSys 2019 (Invited Talk) Technische Universitt Mnchen Failures, Latency

SLIDE 1

Motivation Contributions CDN Penetration Complete Failures Partial Failures IPv6 Preference Latency Content Caches Takeaway

A Longitudinal View of Dual-stacked Websites − Failures, Latency and Happy Eyeballs

Vaibhav Bajpai Technische Universität München NetSys 2019 (Invited Talk) Garching bei München, Germany

IEEE/ACM Transactions on Networking March 2019 https://doi.org/10.1109/TNET.2019.2895165

March 20, 2019

1 / 15

SLIDE 2

Motivation Contributions CDN Penetration Complete Failures Partial Failures IPv6 Preference Latency Content Caches Takeaway

Motivation

▶ Literature focus largely on IPv6 adoption. ▶ Very little work on measuring IPv6 performance. Do users experience benefjt (or sufger) from web content delivery over IPv6? What factors con- tribute to the difgerence? How has content delivery

ver IPv6 evolved over years?

∼100 dual-stacked SamKnows probes (∼66 origin ASes)

2009 2010 2011 2012 2013 2014 2015 2016 2017 2018 2019 0% 6% 12% 18% 24% 30% Google IPv6 Adoption

shaded region represents the duration of the longitudinal study. 01 04 07 10 13 16 19 22 25 28 Nov 2018 20% 25% Google IPv6 Adoption ▶

1⁄4 of connections to Google are made over IPv6.

▶ IPv6 penetration more in home deployments.

2 / 15

SLIDE 3

Motivation Contributions CDN Penetration Complete Failures Partial Failures IPv6 Preference Latency Content Caches Takeaway

Research Contributions

▶ Penetration of Content Delivery Networks (CDNs) ▶ Latency, Failures and Happy Eyeballs ▶ IP Path Lengths and Content Cache Deployments Tiis is the fjrst study to provide a longitudinal view (6 years)

f web content delivery over IPv6.

▶ Relevance: ▶ Network operators in early stages of IPv6 deployment. ▶ Content providers to see how their service delivery over IPv6 compares to IPv4. ▶ Drive related standards work in the IETF.

3 / 15

SLIDE 4

Motivation Contributions CDN Penetration Complete Failures Partial Failures IPv6 Preference Latency Content Caches Takeaway

Content Delivery Networks (CDNs)

0.0 0.2 0.4 0.6 0.8 1.0 CDF ALEXA 10K with AAAA entries 0% 20% 40% 60% 80% 100% CDN Penetration 0.0 0.2 0.4 0.6 0.8 1.0 CDF Cloudflare Google Akamai Total IPv6 IPv4

▶ CDNs play a leading role in technology adoption. ▶ Leading players:

1. Cloudfmare (∼35%)
2. Google (∼16%) and
3. Akamai (∼9%)

2009 2010 2011 2012 2013 2014 2015 2016 2017 2018 2019 0.0% 10.0% 20.0% W6D W6LD ALEXA 1M with AAAA entries

▶

1⁄5 of 1M websites announce AAAA entries (2019).

▶ Cloudfmare added AAAA entries for all websites [1]. ▶ Cloudfmare’s impact (shaded) > W6D (or W6LD).

CDNs serve (2018) more than 1⁄2 of 10K websites.

4 / 15

SLIDE 5

Motivation Contributions CDN Penetration Complete Failures Partial Failures IPv6 Preference Latency Content Caches Takeaway

Complete Failures

2009 2010 2011 2012 2013 2014 2015 2016 2017 2018 2019 0.0% 20.0% 40.0% W6D W6LD ALEXA 1M with AAAA entries HTTP Failure

▶ Failures reduced from 40% (2009) to 2% (2019).

0.1K 1K 10K 100K 1000K ALEXA Rank 0.0 0.2 0.4 0.6 0.8 1.0 CCDF Failing AAAA Websites (3.7K) [Jan '19]

▶ 88% failing websites rank > 100K. ▶ 1% rank < 10K, fjve websites rank < 300.

100 www.bing.com 200 400 www.detik.com 100 www.engadget.com 200 400 www.nifty.com 2013 2014 2015 2016 2017 2018 200 250 300 350 www.sakura.ne.jp IPv6 IPv4 TCP Connect Times (ms)

Failures in content delivery over IPv6 have reduced to 2%

5 / 15

SLIDE 6

Motivation Contributions CDN Penetration Complete Failures Partial Failures IPv6 Preference Latency Content Caches Takeaway

Partial Failures

ALEXA top 100 websites with AAAA entries. ▶ 27% show some rate of failure over IPv6. ▶ 9% exhibit more than 50% failures over IPv6.

20 40 60 80 100 Success Rate (%) 0.0 0.2 0.4 0.6 0.8 1.0 CDF IPv6 (100) IPv4 (100)

▶ Limiting to root webpage can lead to

verestimation of IPv6 adoption numbers

▶ Unclear whether websites with partial failures can be deemed IPv6-ready ▶ ISOC now supporting [2] development of tools that identify such partial failures

# Webpage Success Rate (%) W6LD IPv6(↓) IPv4 01 www.bing.com 100 ✓ 02 www.detik.com 100 ✓ 03 www.engadget.com 100 ✓ 04 www.nifty.com 100 05 www.qq.com 100 06 www.sakura.ne.jp 100 07 www.flipkart.com 09 99 ✓ 08 www.folha.uol.com.br 13 100 09 www.aol.com 48 100 ✓ 10 www.comcast.net 52 100 ✓ 11 www.yahoo.com 72 100 ✓ 12 www.mozilla.org 84 100 ✓ 13 www.orange.fr 86 100 ✓ 14 www.seznam.cz 89 100 ✓ 15 www.mobile.de 90 100 ✓ 16 www.wikimedia.org 90 100 17 www.t-online.de 93 100 ✓ 18 www.free.fr 95 100 19 www.usps.com 95 100 20 www.vk.com 95 100 ✓ 21 www.wikipedia.org 95 100 ✓ 22 www.wiktionary.org 95 100 23 www.elmundo.es 96 100 ✓ 24 www.uol.com.br 96 100 ✓ 25 www.marca.com 97 100 ✓ 26 www.terra.com.br 98 100 ✓ 27 www.youm7.com 99 100 6 / 15

SLIDE 7

Motivation Contributions CDN Penetration Complete Failures Partial Failures IPv6 Preference Latency Content Caches Takeaway

Partial Failures | Root Cause Analysis

30 60 90 www.youm7.com (1%) www.terra.com.br (2%) www.marca.com (3%) www.uol.com.br (4%) www.elmundo.es (4%) www.wiktionary.org (5%) www.wikipedia.org (5%) www.vk.com (5%) www.usps.com (5%) www.free.fr (5%) www.t-online.de (7%) www.wikimedia.org (10%) www.mobile.de (10%) www.seznam.cz (11%) www.orange.fr (14%) www.mozilla.org (16%) www.yahoo.com (28%) www.comcast.net (48%) www.aol.com (52%) www.folha.uol.com.br (87%) www.flipkart.com (91%) www.sakura.ne.jp (100%) www.qq.com (100%) www.nifty.com (100%) www.engadget.com (100%) www.detik.com (100%) www.bing.com (100%) Network Level

CURLE_OK CURLE_COULDNT_RESOLVE_HOST CURLE_COULDNT_CONNECT CURLE_OPERATION_TIMEDOUT CURLE_GOT_NOTHING CURLE_RECV_ERROR

30 60 90 Contribution (%) Content Level

*/css */html */javascript, */json */octet-stream */plain */rdf */xml image/*

30 60 90 Service Level

SAME ORIGIN CROSS ORIGIN

Website failing over IPv6

▶ Failures due to DNS resolution error on

image/*, */javascript, */json and */css content.

▶ Failures silently exist; clients do not notice them due to IPv4 fallback. ▶ Identifjcation of operational issues relevant for upcoming IPv6-only networks

7 / 15

SLIDE 8

Motivation Contributions CDN Penetration Complete Failures Partial Failures IPv6 Preference Latency Content Caches Takeaway

Partial Failures | Root Cause Analysis

30 60 90 Contribution (%) www.youm7.com (1%) www.terra.com.br (2%) www.marca.com (3%) www.uol.com.br (4%) www.elmundo.es (4%) www.wiktionary.org (5%) www.wikipedia.org (5%) www.vk.com (5%) www.usps.com (5%) www.free.fr (5%) www.t-online.de (7%) www.wikimedia.org (10%) www.mobile.de (10%) www.seznam.cz (11%) www.orange.fr (14%) www.mozilla.org (16%) www.yahoo.com (28%) www.comcast.net (48%) www.aol.com (52%) www.folha.uol.com.br (87%) www.flipkart.com (91%) www.sakura.ne.jp (100%) www.qq.com (100%) www.nifty.com (100%) www.engadget.com (100%) www.detik.com (100%) www.bing.com (100%) *.youm7.com *.terra.com.br *.marca.com *.uol.com.br *.elmundo.es *.wiktionary.org *.wikipedia.org *.vk.com *.usps.com *.free.fr *.t-online.de *.wikimedia.org *.mobile.de *.seznam.cz *.orange.fr *.mozilla.org *.yahoo.com *.comcast.net *.aol.com *.uol.com.br *.flipkart.com *.sakura.ne.jp *.qq.com *.nifty.com *.engadget.com *.detik.com *.bing.com SAME ORIGIN

▶ 12% of websites have more than 50% webpage elements that belong to the same origin source and fail over IPv6. ▶ CDN infrastructure does not have IPv6 turned on by default for all same-origin webpage elements.

# Webpage Same Origin (↓) 01 www.bing.com 100% 02 www.detik.com 100% 03 www.engadget.com 100% 04 www.nifty.com 100% 05 www.usps.com 100% 06 www.qq.com 100% 07 www.sakura.ne.jp 100% 08 www.comcast.net 85% 09 www.yahoo.com 83% 10 www.terra.com.br 74% 11 www.marca.com 70% 12 www.wikimedia.org 65% 13 www.elmundo.es 37% 14 www.vk.com 31% 15 www.t-online.de 30% 16 www.youm7.com 24% 17 www.wiktionary.org 22% 18 www.wikipedia.org 22% 19 www.free.fr 13% 20 www.folha.uol.com.br 12% 21 www.mozilla.org 7% 22 www.uol.com.br 7% 23 www.mobile.de 7% 24 www.aol.com 5% 25 www.orange.fr 5% 26 www.seznam.cz 4% 27 www.flipkart.com 1% 8 / 15

SLIDE 9

Motivation Contributions CDN Penetration Complete Failures Partial Failures IPv6 Preference Latency Content Caches Takeaway

Partial Failures | Root Cause Analysis

*.adition.com *.ajax.googleapis.com *.aolcdn.com *.cimcontent.net *.creativecommons.org *.d5nxst8fruw4z.cloudfront.net *.demdex.net *.dmtry.com *.doubleclick.net *.el-mundo.net *.elmundo.es *.expansion.com *.f.i.uol.com.br *.flixcart.com *.globaliza.com *.images1.folha.com.br *.imedia.cz *.imguol.com *.imguol.com.br *.interactivemedia.net *.ioam.de *.jsuol.com.br *.leguide.com *.ligatus.com *.mail.ru *.mozilla.net *.navdmp.com *.netbiscuits.net *.omtrdc.net *.optimizely.com *.outbrain.com *.proxad.net *.quantserve.com *.sblog.cz *.scorecardresearch.com *.szn.cz *.tag.navdmp.com *.telva.com *.theadex.com *.toi.de *.trrsf.com *.unidadeditorial.es *.voila.fr *.woopic.com *.www1.folha.com.br *.xiti.com

▶ Tiird-party advertisements (.doubleclick.net) ▶ Analytics (.scorecardresearch.com, .quantserve.com) ▶ User-centric content (.facebook.com, .ajax.googleapis.com) ▶ Static content (.wikimedia.org, *.creativecommons.org) ▶ Enabling IPv6 on few cross cross-origin sources (creativecommons.org, doubleclick.net) will help reduce partial failure of multiple websites.

9 / 15

SLIDE 10

Motivation Contributions CDN Penetration Complete Failures Partial Failures IPv6 Preference Latency Content Caches Takeaway

IPv6 Preference

95% 96% 97% 98% 99% 100% 0.0 0.2 0.4 0.6 0.8 1.0 CCDF Web (458) Audio (458) Video (458)

['14 - '16]

IPv6 Preference

▶ RFC 6724 [3] makes apps prefer connections made over IPv6. ▶ RFC 6555 [4] allows apps to fallback to IPv4 when IPv6 connectivity is bad. ▶ TCP connections over IPv6 are preferred at least 97% of the time. Clients prefer web and video content over IPv6

10 / 15

SLIDE 11

Motivation Contributions CDN Penetration Complete Failures Partial Failures IPv6 Preference Latency Content Caches Takeaway

Latency

10−1 100 101 102 103 0.0 0.2 0.4 0.6 0.8 1.0

CDF

250 ms 10−1 100 101 102 103 TCP Connect Times (ms) 0.0 0.2 0.4 0.6 0.8 1.0

CDF

250 ms 2018 2017 2016 2015 2014 2013 IPv6 IPv4

Latency (2013 → 2018) reduced by: ▶ 29% (21 ms → 15 ms) over IPv4 ▶ 57% (28 ms → 16 ms) over IPv6 ▶ Latency over IPv6 has reduced by more than half in past six years (2013 - 2018).

t0 t0 + 250 ms time IPv6 IPv4 Happy Eyeballs [RFC 8305]

▶ Only 3% of samples above HE timer > 250 ms ▶ Clients prefer IPv6 when slower in 81% cases. ▶ HE timer can be reduced to 150 ms. ▶ RFC 6555 should have used 150 ms timer. ▶ Measurements should inform IETF work.

11 / 15

SLIDE 12

Motivation Contributions CDN Penetration Complete Failures Partial Failures IPv6 Preference Latency Content Caches Takeaway

Latency

▶ 50% samples used to 5 ms or more slower

ver IPv6 (2013); reduced to 8% (2018).

−1000−100 −10 −1 0 1 10 100 1000 TCP Connect Times [∆sa (ms)] 0.0 0.2 0.4 0.6 0.8 1.0 CDF IPv4 faster IPv6 faster 2018 2017 2016 2015 2014 2013 −100 −10 −1 1 10 100 TCP Connect Times [∆sa (ms)] 0.0 0.2 0.4 0.6 0.8 1.0 CDF (ALEXA 10K) IPv4 faster IPv6 faster [12/2018]

▶ 56% websites are faster over IPv6. ▶ 95% of the rest are at most 1 ms slower. ▶ 2% are at least 25 ms slower. ▶ More than half of ALEXA 10K websites are faster over IPv6 (2018)

12 / 15

SLIDE 13

Motivation Contributions CDN Penetration Complete Failures Partial Failures IPv6 Preference Latency Content Caches Takeaway

How benefjcial are the content caches?

2 4 6 8 10 12 14 16 18 20 22 TTL 0.0 0.2 0.4 0.6 0.8 1.0 CDF

no cache (IPv4) cache (IPv4) no cache (IPv6) cache (IPv6)

▶ 90% caches are reachable within 6 IP hops.

10 20 30 40 50 60 RTT [ms] 0.0 0.2 0.4 0.6 0.8 1.0 CDF

no cache (IPv4) cache (IPv4) no cache (IPv6) cache (IPv6)

▶ Caches are reachable within 20 ms. ▶ Caches reduce latencies by: ▶ IPv4: 25 ms → 17 ms; 1⁄3 improvement ▶ IPv6: 29 ms → 16 ms; 1⁄2 improvement ▶ GGC caches reduce IP path lengths by 1⁄2 over both AF. ▶ GGC caches reduce latencies over IPv6 by 1⁄2 and over IPv4 by 1⁄3.

13 / 15

SLIDE 14

Motivation Contributions CDN Penetration Complete Failures Partial Failures IPv6 Preference Latency Content Caches Takeaway

Dual-stacked content caches

▶ Caches always reduce IP path lengths. ▶ Path lengths difger when caches not hit.

15 10 5 5 10 15 TTL delta 0.0 0.2 0.4 0.6 0.8 1.0 CDF

IPv6 slower IPv6 faster

IPv4 only IPv6 only Both versions Neither version

1000 100 10 1 0 1 10 100 1000 RTT delta [ms] 0.0 0.2 0.4 0.6 0.8 1.0 CDF IPv6 slower IPv6 faster

IPv4 only IPv6 only Both versions Neither version

▶ Dual-stacked: ∼80% within 1 ms. ▶ IPv4-only: ∼80% faster over IPv4. ▶ IPv6-only: 1⁄4 still slower over IPv6. IPv6 performance degrades when content caches are not dual-stacked.

14 / 15

SLIDE 15

Motivation Contributions CDN Penetration Complete Failures Partial Failures IPv6 Preference Latency Content Caches Takeaway

Takeaway

▶

1⁄4 of connections to Google are made (mostly from home networks) over IPv6.

▶ CDNs serve more than 1⁄2 of ALEXA 10K websites (2018). ▶ IPv6 failures reduced to 2%. IPv6 latency reduced by more than half in six years. ▶ More than 1⁄2 of ALEXA 10K websites are faster over IPv6 (2018). ▶ Caches can be reached in six IP hops; reduce IP path lengths by 1⁄2 over both AF. ▶ Caches reduce latencies over IPv6 by 1⁄2 and over IPv4 by 1⁄3. ▶ Latency over IPv6 degrades when caches are not dual-stacked.

▶ Reproducibility Considerations:

https://github.com/vbajpai/2018-ton-v6-longitudinal-websites

www.vaibhavbajpai.com bajpaiv@in.tum.de | @bajpaivaibhav

15 / 15

SLIDE 16

Motivation Contributions CDN Penetration Complete Failures Partial Failures IPv6 Preference Latency Content Caches Takeaway

References

[1] “98.01% of sites on Cloudfmare now use IPv6,” https://blog.cloudflare.com/98-percent-ipv6, [Online; accessed 15-Apr-2017]. [2] “NAT64 Check,” nat64check.ipv6-lab.net, [Accessed 15-Apr-2017]. [3]

D. Tialer, R. Draves, A. Matsumoto, and T. Chown, “Default Address

Selection for Internet Protocol Version 6 (IPv6),” RFC 6724, Internet Engineering Task Force, Sep. 2012, https://tools.ietf.org/html/rfc6724. [4]

D. Wing and A. Yourtchenko, “Happy Eyeballs: Success with

Dual-Stack Hosts,” RFC 6555, IETF, 2012, https://tools.ietf.org/html/rfc6555. 15 / 15