Emile Aben
RIPE NCC Measurements - World IPv6 Day• IPv6 Eyechart and 6to4 (not in this talk)• Active measurements
– Sources: 49 vantage points (RIPE TTM, CAIDA Ark, ...)– Destinations: 53 participant or already dual-stacked sites
– From 2011-06-01 to 2011-06-11 we measured– DNS: A and/or AAAA records– ping(6)/traceroute(6)– HTTP over IPv4 and IPv6
2
Emile Aben
Measurement Vantage Points
3
Lesson: Control
Emile Aben
When Did World IPv6 Day Start?• Less then 2 days before World IPv6 Day:
5
0 10 20 30 40 50 60 70 80 90
1m 2m 5m 10m 20m 1h 2h 5h 10h 1d 2d
Num
ber o
f par
ticip
atin
g w
ebsi
tes
minimum of: minimum TTL in SOA, TTL of SOA
DNS negative caching for websites participating in World IPv6 Day
Emile Aben
Percentage Of Vantage Points Seeing AAAA
6
v6day
Emile Aben
Percentage Of Vantage Points Seeing AAAA
6
v6day
Negative caching
Emile Aben
Percentage Of Vantage Points Seeing AAAA
6
v6day
Negative cachingTTL caching
Emile Aben
Control - Lessons Learned• Know your on/off switch
– Control your DNS• Set low TTL in case of
roll-back• Set low negative TTL
– By means of minTTL in SOA / TTL of SOA
7
Lesson:Test And Monitor
Emile Aben
You Don’t Want This To Happen
9
On IPv6:
On IPv4:
Emile Aben
Or This ...
10
0 %
20 %
40 %
60 %
80 %
100 %
00:0004:00
08:0012:00
16:0020:00
00:0004:00
08:0012:00
Comparing DNS,ping and HTTP IPv6 measurementsto www.commerce.gov from 2011-06-09 0:00 UTC to 2011-06-10 12:00 UTC
% AAAA queries returning an IPv6 address% HTTP responses == 200
% ICMPv6 ping success
Emile Aben
Or This ...
10
0 %
20 %
40 %
60 %
80 %
100 %
00:0004:00
08:0012:00
16:0020:00
00:0004:00
08:0012:00
Comparing DNS,ping and HTTP IPv6 measurementsto www.commerce.gov from 2011-06-09 0:00 UTC to 2011-06-10 12:00 UTC
% AAAA queries returning an IPv6 address% HTTP responses == 200
% ICMPv6 ping success
v6 on www turned off at end of v6day
Emile Aben
Or This ...
10
0 %
20 %
40 %
60 %
80 %
100 %
00:0004:00
08:0012:00
16:0020:00
00:0004:00
08:0012:00
Comparing DNS,ping and HTTP IPv6 measurementsto www.commerce.gov from 2011-06-09 0:00 UTC to 2011-06-10 12:00 UTC
% AAAA queries returning an IPv6 address% HTTP responses == 200
% ICMPv6 ping success
but v6 service stillannounced
v6 on www turned off at end of v6day
Emile Aben
Test And Monitor - Lessons Learned
11
Emile Aben
Test And Monitor - Lessons Learned• Test when deploying something
– The more real-life, the less likely you !@#$-ed up• Monitor your infrastructure• People2people reachability
– Avoidable situations like Level3 and Dept. Commerce– Contact info up to date in RIR databases (whois)– Monitor the web (NANOG, *NOG, Twitter, ...)
11
Global View
Emile Aben
IPv4/IPv6 Performance On World IPv6 Day
• Bell-shaped– centered around 0– with fatter IPv4-side (62%)
• Dual-stack = two chances for best performance!– Real-time apps can use this (voice/game)
13
Distribution of IPv4/IPv6 relative performance
best performing protocol, as percentage−wise increase over the least performing protocol
Freq
uenc
y
+ 40
0 %
+ 35
0 %
+ 30
0 %
+ 25
0 %
+ 20
0 %
+ 15
0 %
+ 10
0 %
+ 50
%+
50 %
+ 40
%+
30 %
+ 20
%+
10 %
+ 0
%+
0 %
+ 10
%+
20 %
+ 30
%+
40 %
+ 50
%+
50 %
+ 10
0 %
+ 15
0 %
+ 20
0 %
+ 25
0 %
+ 30
0 %
+ 35
0 %
+ 40
0 %
IPv4 performs better IPv6 performs better
Emile Aben
Performance - Tunnels?• 9.2% of paths had 1 (or more!) hops with ‘tunnel’
in the hostname• TTM mesh tunnel-discovery: < 2% of paths
14
0 %
10 %
20 %
30 %
40 %
50 %
60 %
012004
012005
012006
012007
012008
012009
012010
012011
012012
Percentage of paths in TTM IPv6 mesh where tunnels are detected
Emile Aben
Performance - Same Path: No Difference
15
Traceroutes from Univ. in Greece to Univ. in Los Angeles
IPv6
IPv4
Emile Aben
Partial Reachability• Internet is a collection of interconnecting
networks, and can be different on IPv4/IPv6• Some vantage points couldn’t reach all dests.
– Are our vantage points representative?– Working on this: RIPE Atlas
– Network partitioning, examples we encountered:– Level3 - Hurricane Electric– Cogent - Hurricane Electric– See http://en.wikipedia.org/wiki/Comparison_of_IPv6_support_by_major_transit_providers
16
Emile Aben
Partial Reachability• Internet is a collection of interconnecting
networks, and can be different on IPv4/IPv6• Some vantage points couldn’t reach all dests.
– Are our vantage points representative?– Working on this: RIPE Atlas
– Network partitioning, examples we encountered:– Level3 - Hurricane Electric– Cogent - Hurricane Electric– See http://en.wikipedia.org/wiki/Comparison_of_IPv6_support_by_major_transit_providers
16
(fixed as of Aug 31)
Emile Aben
Long Term Effects - Content
17
Raw data: Dan Wing (http://banjo.employees.org/~dwing/aaaa-stats.html)
0 %0.5 %
1 %1.5 %
2 %2.5 %
3 %3.5 %
4 %
2011-01
2011-02
2011-03
2011-04
2011-05
2011-06
2011-07
2011-08
2011-09
2011-10
2011-11
Percentage of web sites in Alexa 1M that can be reached over IPv6
Alexa 1MAlexa top 25k
Emile Aben
Long Term Effects - Content
17
Raw data: Dan Wing (http://banjo.employees.org/~dwing/aaaa-stats.html)
8 June:3.8%
0 %0.5 %
1 %1.5 %
2 %2.5 %
3 %3.5 %
4 %
2011-01
2011-02
2011-03
2011-04
2011-05
2011-06
2011-07
2011-08
2011-09
2011-10
2011-11
Percentage of web sites in Alexa 1M that can be reached over IPv6
Alexa 1MAlexa top 25k
Emile Aben
Long Term Effects - Content
17
Raw data: Dan Wing (http://banjo.employees.org/~dwing/aaaa-stats.html)
Pre/post 8 June:0.4% -> 0.7%
8 June:3.8%
0 %0.5 %
1 %1.5 %
2 %2.5 %
3 %3.5 %
4 %
2011-01
2011-02
2011-03
2011-04
2011-05
2011-06
2011-07
2011-08
2011-09
2011-10
2011-11
Percentage of web sites in Alexa 1M that can be reached over IPv6
Alexa 1MAlexa top 25k
Emile Aben
Long Term Effects - Content
17
Raw data: Dan Wing (http://banjo.employees.org/~dwing/aaaa-stats.html)
Pre/post 8 June:0.4% -> 0.7%
Around 3 July:0.7% -> 1.0%(DE?!)
8 June:3.8%
0 %0.5 %
1 %1.5 %
2 %2.5 %
3 %3.5 %
4 %
2011-01
2011-02
2011-03
2011-04
2011-05
2011-06
2011-07
2011-08
2011-09
2011-10
2011-11
Percentage of web sites in Alexa 1M that can be reached over IPv6
Alexa 1MAlexa top 25k
Emile Aben
Long Term Effects - IPv6 ASes
• Higher growth pre-event: Deployments pushed earlier?• Lower growth post-event: Summer-vacation? Economy?• http://v6asns.ripe.net
18
8 June
-50-40-30-20-1001020304050
2011-04-01
2011-05-01
2011-06-01
2011-07-01
2011-08-01
2011-09-01
2011-10-01
2011-11-01
Dai
ly in
crea
seof
IPv6
ASe
s
9.5 %
10 %
10.5 %
11 %
11.5 %
12 %
12.5 %
% o
f ASe
san
noun
cing
IPv6
pfx
Emile Aben
Conclusions - What We Learned• IPv6/dual-stack works just fine, but make sure
that– It is properly tested and monitored (like IPv4)– Your network can reach all others (like IPv4)
• Dual-stack = Two chances for best performance• Days like this ‘work’
– Raise awareness– Give people a target to work towards– We’re ready for a next IPv6(day|week|month|year|∞)
19
Emile Aben
More Information• Web interface to the measurements
– http://v6day.ripe.net/• Analysis on RIPE Labs
– http://labs.ripe.net/ipv6day• Raw data availability
– http://labs.ripe.net/datarepository/data-sets/ripe-ncc-active-measurements-of-world-ipv6-day-dataset
20
Questions?