[PPT] - Exhaustion of IPv4 Address in 2010 Exhaustion of IPv4 Address in PowerPoint Presentation

SLIDE 1

1 1

TAC2000/2000.7

802.16 IP Telephony Lab

Exhaustion of IPv4 Address in 2010 Exhaustion of IPv4 Address in 2010

Associate Professor

Dr. Quincy Wu

solomon@ipv6.club.tw Graduate Institute of Communication Engineering

National Chi Nan University

SLIDE 2

2 2

TAC2000/2000.7

802.16 IP Telephony Lab

Outline

IP(v4) Address

IP(v4) Address

IP Address Allocation

IP Address Allocation

Introduction to IPv6

Introduction to IPv6

國外現況

國外現況

標準組織

標準組織

Benefit of IPv6

Benefit of IPv6

其他國家

其他國家

IPv6 Address Allocation

IPv6 Address Allocation

SLIDE 3

3 3

TAC2000/2000.7

802.16 IP Telephony Lab

Current IP(v4) Address

bbs.ptt.cc

bbs.ptt.cc = 140.112.172.11 = 140.112.172.11

Every node needs a unique IP address to get connected to

Every node needs a unique IP address to get connected to Internet. Internet.

2

232

32 = 4 billion, but there are 6.5 billion people on earth.

= 4 billion, but there are 6.5 billion people on earth.

30 years ago, when hundreds of students and teachers were sharin

30 years ago, when hundreds of students and teachers were sharing a g a mainframe computer in university campus, this seems to be suffic mainframe computer in university campus, this seems to be sufficient. ient.

8

8-

bit Network ID (totally 256 networks) seems sufficient for the f

bit Network ID (totally 256 networks) seems sufficient for the foreseeable

reseeable
future. ~
future. ~ Vint

Vint Cerf Cerf

640K ought to be enough for anybody. ~ Bill Gates

640K ought to be enough for anybody. ~ Bill Gates

When computers become

When computers become “ “personal personal” ”, it implies that each person may , it implies that each person may have a computer. 4 billions is obviously insufficient. have a computer. 4 billions is obviously insufficient.

In the future, any single person may have multiple devices conne

In the future, any single person may have multiple devices connected to cted to Internet (PC, PDA, mobile phone, TV, XBOX, etc.) Internet (PC, PDA, mobile phone, TV, XBOX, etc.)

Fortunately (unfortunately?), only 1 billion of them have access

Fortunately (unfortunately?), only 1 billion of them have access to to Internet now. Internet now.

SLIDE 4

4 4

TAC2000/2000.7

802.16 IP Telephony Lab

IPv4 Address Allocation

Internet Assigned Numbers Authority (IANA)

Internet Assigned Numbers Authority (IANA)

manage the unallocated IPv4

manage the unallocated IPv4 unicast unicast address pool address pool

Regional Internet Registries (

Regional Internet Registries (RIRs RIRs) ) – – always receive /8 blocks always receive /8 blocks

ARIN (North America)

ARIN (North America)

RIPE NCC (Europe)

RIPE NCC (Europe)

APNIC (Asia/Pacific)

APNIC (Asia/Pacific)

LACNIC (Latin America)

LACNIC (Latin America) -

7 November 2002

7 November 2002

AfriNIC

AfriNIC (Africa) (Africa) -

8 April 2005

8 April 2005

National Internet Registries (

National Internet Registries (NIRs NIRs) )

JPNIC (Japan)

JPNIC (Japan)

KRNIC (Korea)

KRNIC (Korea)

TWNIC (Taiwan)

TWNIC (Taiwan)

SLIDE 5

5 5

TAC2000/2000.7

802.16 IP Telephony Lab

IANA /8 Allocation in 2005

SLIDE 6

6 6

TAC2000/2000.7

802.16 IP Telephony Lab

Status of /8s IPv4 Addresses in 2007

APNIC, 24 ARIN, 27 LACNIC, 4 RIPE NCC, 24 Multicast, 16 IANA Reserved, 48 Central Registry, 94 AfriNIC, 1 Experimental, 16 Public Use, 1 Private Use, 1

IP addressing in China and the myth of address shortage IP addressing in China and the myth of address shortage IP addressing in China and the myth of address shortage IP addressing in China and the myth of address shortage

http://www.apnic.net/news/hot-topics/index.html#ip-addressing

16 /8 blocks are allocated in two years!

SLIDE 7

7 7

TAC2000/2000.7

802.16 IP Telephony Lab

Deceleration & Acceleration in Growth

SLIDE 8

8 8

TAC2000/2000.7

802.16 IP Telephony Lab

Allocation of /16 Addresses in APNIC

Updated on 2008.03.27

ftp://ftp.apnic.net/pub/stats/apnic/

SLIDE 9

9 9

TAC2000/2000.7

802.16 IP Telephony Lab

Some Universities Have a Class B

inetnum

inetnum: 140.112.0.0 : 140.112.0.0 -

140.112.255.255

140.112.255.255

netname

netname: T : T-

NTU.EDU.TW

NTU.EDU.TW-

NET

NET

descr

descr: National Taiwan University : National Taiwan University

descr

descr: Taipei Taiwan : Taipei Taiwan

country: TW

country: TW

admin

admin-

c:

c: AN187 AN187-

TW

TW

tech

tech-

c:

c: AN187 AN187-

TW

TW

mnt

mnt-

by:

by: MAINT MAINT-

TW

TW-

TWNIC

TWNIC

remarks: This information has been partially mirrored by APNIC f

remarks: This information has been partially mirrored by APNIC from rom remarks: TWNIC. To obtain more specific information, please use remarks: TWNIC. To obtain more specific information, please use the the remarks: TWNIC remarks: TWNIC whois whois server at server at whois.twnic.net whois.twnic.net. .

changed:

changed: chuang@mail.moe.gov.tw chuang@mail.moe.gov.tw 19900324 19900324

status: ASSIGNED NON

status: ASSIGNED NON-

PORTABLE

PORTABLE

source: TWNIC

source: TWNIC

SLIDE 10

10 10

TAC2000/2000.7

802.16 IP Telephony Lab

Some Organizations Have a Class A

009/8 IBM

009/8 IBM 1992 1992-

08

08

013/8 Xerox Corporation

013/8 Xerox Corporation 1991 1991-

09

09

015/8 Hewlett

015/8 Hewlett-

Packard Company

Packard Company 1994 1994-

07

07

017/8 Apple Computer Inc.

017/8 Apple Computer Inc. 1992 1992-

07

07

018/8 MIT

018/8 MIT 1994 1994-

01

01

019/8 Ford Motor Company

019/8 Ford Motor Company 1995 1995-

05

05

041/8

041/8 AfriNIC AfriNIC 2005 2005-

04

04

196/8

196/8 AfriNIC AfriNIC 1993 1993-

05

05

http://www.iana.org./assignments/ipv4-address-space

SLIDE 11

11 11

TAC2000/2000.7

802.16 IP Telephony Lab

It is harder to get a complete /16 now!

www.csie.mcu.edu.tw

www.csie.mcu.edu.tw = 210.59.106.65 = 210.59.106.65

inetnum

inetnum: 210.59.104.0 : 210.59.104.0 -

210.59.107.255

210.59.107.255

netname

netname: MCU : MCU-

TY

TY-

NET

NET

descr

descr: Ming : Ming-

Chuan

Chuan University(Taoyuan University(Taoyuan Campus) Campus)

descr

descr: : Taoyuan Taoyuan Taiwan Taiwan

country: TW

country: TW

admin

admin-

c:

c: AM7 AM7-

TW

TW

admin

admin-

c:

c: AM6 AM6-

TW

TW

tech

tech-

c:

c: AM7 AM7-

TW

TW

mnt

mnt-

by:

by: MAINT MAINT-

TW

TW-

TWNIC

TWNIC

remarks: This information has been partially mirrored by APNIC f

remarks: This information has been partially mirrored by APNIC from rom remarks: TWNIC. To obtain more specific information, please use remarks: TWNIC. To obtain more specific information, please use the the remarks: TWNIC remarks: TWNIC whois whois server at server at whois.twnic.net whois.twnic.net. .

changed:

changed: chuang@mail.moe.gov.tw chuang@mail.moe.gov.tw 20021230 20021230

status: ASSIGNED NON

status: ASSIGNED NON-

PORTABLE

PORTABLE

source: TWNIC

source: TWNIC

http://wq.apnic.net/apnic-bin/whois.pl

SLIDE 12

12 12

TAC2000/2000.7

802.16 IP Telephony Lab

IPv4 Lifetime Projections

SLIDE 13

13 13

TAC2000/2000.7

802.16 IP Telephony Lab

The Clock is Ticking!

Projected IANA Unallocated

Projected IANA Unallocated Address Pool Exhaustion: 05 Address Pool Exhaustion: 05-

Mar

Mar-

2011

2011

Projected RIR Unallocated

Projected RIR Unallocated Address Pool Exhaustion: 22 Address Pool Exhaustion: 22-

May

May-

2012

2012

On-line IPv4 Address Report: http://www.potaroo.net/tools/ipv4/index.html

SLIDE 14

14 14

TAC2000/2000.7

802.16 IP Telephony Lab

Summary

Network Address Translation

Network Address Translation (NAT) and (NAT) and Classless Inter Classless Inter-

Domain Routing

Domain Routing (CIDR) did their jobs and bought the 10 years (CIDR) did their jobs and bought the 10 years needed to get IPv6 standards and products developed. needed to get IPv6 standards and products developed.

Now is the time to recognize the end to sustainable growth of

Now is the time to recognize the end to sustainable growth of the IPv4 the IPv4-

based Internet has arrived

based Internet has arrived

It is time to move to IPv6.

It is time to move to IPv6.

These steps will take time

These steps will take time— —in many cases multiple years. in many cases multiple years.

Staff will need to be trained,

Staff will need to be trained,

Management tools will need to be enhanced,

Management tools will need to be enhanced,

Routers and operating systems will need to be updated,

Routers and operating systems will need to be updated,

IPv6

IPv6-

enabled versions of applications will need to be deployed.

enabled versions of applications will need to be deployed.

SLIDE 15

15 15

TAC2000/2000.7

802.16 IP Telephony Lab

IPv4 Address Depletion

Hiroshi Esaki, Ph.D. Professor, The University of Tokyo

SLIDE 16

16 16

TAC2000/2000.7

802.16 IP Telephony Lab

Japanese MIC formed study group

Mission:

Mission: “ “How to achieve smooth IPv6 introduction How to achieve smooth IPv6 introduction” ”

Some interesting discussion and analysis

Some interesting discussion and analysis

1. 1. There is no free lunch ! All must pay money on IPv6. There is no free lunch ! All must pay money on IPv6. 2. 2. Squeezing out the global IPv4 address from existing network look Squeezing out the global IPv4 address from existing network looks so hard s so hard… …. . JPNIC had got less than 2% of address space JPNIC had got less than 2% of address space… ….. .. 3. 3. IPv4 address exchanging market may generate the company accounti IPv4 address exchanging market may generate the company accounting issue, since ng issue, since IPv4 address may become as an asset. IPv4 address may become as an asset. 4. 4. Broadband Internet consumes a lot of global IP(v4) addresses Broadband Internet consumes a lot of global IP(v4) addresses 5. 5. RIPE is large IPv4 address consumption as well as RIPE is large IPv4 address consumption as well as BRICs BRICs area. area. 6. 6. The largest sacrifice is business/service deployment for new com The largest sacrifice is business/service deployment for new companies and for panies and for legacy companies legacy companies 7. 7. Contents provider and system integrator should join Contents provider and system integrator should join 8. 8. Translator between large clouds will not work Translator between large clouds will not work… ….. Put it at the edge, e.g., SOHO .. Put it at the edge, e.g., SOHO router and router and iDC iDC

SLIDE 17

17 17

TAC2000/2000.7

802.16 IP Telephony Lab

Japanese MIC formed study group

Mission:

Mission: “ “How to achieve smooth IPv6 introduction How to achieve smooth IPv6 introduction” ”

Some interesting discussion and analysis

Some interesting discussion and analysis

1. 1. There is no free lunch ! All must pay money on IPv6. There is no free lunch ! All must pay money on IPv6. 2. 2. Squeezing out the global IPv4 address from existing network look Squeezing out the global IPv4 address from existing network looks so hard s so hard… …. . JPNIC had got less than 2% of address space JPNIC had got less than 2% of address space… ….. .. 3. 3. IPv4 address exchanging market may generate the company accounti IPv4 address exchanging market may generate the company accounting issue, since ng issue, since IPv4 address may become as an asset. IPv4 address may become as an asset. 4. 4. Broadband Internet consumes a lot of global IP(v4) addresses Broadband Internet consumes a lot of global IP(v4) addresses 5. 5. RIPE is large IPv4 address consumption as well as RIPE is large IPv4 address consumption as well as BRICs BRICs area. area. 6. 6. The largest sacrifice is business/service deployment for new com The largest sacrifice is business/service deployment for new companies and for panies and for legacy companies legacy companies 7. 7. Contents provider and system integrator should join Contents provider and system integrator should join 8. 8. Translator between large clouds will not work Translator between large clouds will not work… ….. Put it at the edge, e.g., SOHO .. Put it at the edge, e.g., SOHO router and router and iDC iDC

6. The largest sacrifice is

business/service deployment for “new” companies and for “legacy” companies

7. Contents provider and system

integrator should/must join, see the Google’s IPv6 site (http://google.com/ipv6)

6. The largest sacrifice is

business/service deployment for “new” companies and for “legacy” companies

7. Contents provider and system

integrator should/must join, see the Google’s IPv6 site (http://google.com/ipv6)

SLIDE 18

18 18

TAC2000/2000.7

802.16 IP Telephony Lab

Japanese MIC formed study group

Mission:

Mission: “ “How to achieve smooth IPv6 introduction How to achieve smooth IPv6 introduction” ”

Some interesting discussion and analysis

Some interesting discussion and analysis

1. 1. There is no free lunch ! All must pay money on IPv6. There is no free lunch ! All must pay money on IPv6. 2. 2. Squeezing out the global IPv4 address from existing network look Squeezing out the global IPv4 address from existing network looks so hard s so hard… …. . JPNIC had got less than 2% of address space JPNIC had got less than 2% of address space… ….. .. 3. 3. IPv4 address exchanging market may generate the company accounti IPv4 address exchanging market may generate the company accounting issue, since ng issue, since IPv4 address may become as an asset. IPv4 address may become as an asset. 4. 4. Broadband Internet consumes a lot of global IP(v4) addresses Broadband Internet consumes a lot of global IP(v4) addresses 5. 5. RIPE is large IPv4 address consumption as well as RIPE is large IPv4 address consumption as well as BRICs BRICs area. area. 6. 6. The largest sacrifice is business/service deployment for new com The largest sacrifice is business/service deployment for new companies and for panies and for legacy companies legacy companies 7. 7. Contents provider and system integrator should join Contents provider and system integrator should join 8. 8. Translator between large clouds will not work Translator between large clouds will not work… ….. Put it at the edge, e.g., SOHO .. Put it at the edge, e.g., SOHO router and router and iDC iDC

8. We may need carrier-class NAT

boxes in the network. But, serious technical issue will

ccur, regarding the number of

TCP sessions………

8. We may need carrier-class NAT

boxes in the network. But, serious technical issue will

ccur, regarding the number of

TCP sessions………

SLIDE 19

19 19

TAC2000/2000.7

802.16 IP Telephony Lab

From IPv4 only To v4/v6 Dual Stack

NTT NTT Communications Communications Shin Shin Miyakawa Miyakawa, , Ph.D Ph.D

NAT can not help you

SLIDE 20

20 20

TAC2000/2000.7

802.16 IP Telephony Lab

Introduction of “Carrier-Grade NAT”

Access Concentrator Internet Global v4 address CPE With NAT Global v4 address End Host Private v4 address FTTH ADSL Access Concentrator With NAT Internet Global v4 address CPE With NAT (newly defined) Private v4 address End Host Private v4 address

SLIDE 21

21 21

TAC2000/2000.7

802.16 IP Telephony Lab

However……

Limitation

Limitation on the number of session states for NAT

n the number of session states for NAT
peration
peration
Each user could use certain number of sessions

Each user could use certain number of sessions

How many sessions ?

How many sessions ?

Even as the best case,

Even as the best case, 65,536

65,536 is the maximum

is the maximum number of sessions, number of sessions, shared by customers shared by customers accommodated into a single IPv4 address accommodated into a single IPv4 address

When the number of users is

When the number of users is 2,000

2,000, it will be

, it will be only

nly

30 sessions 30 sessions

This means

This means…… …….. ..

SLIDE 22

22 22

Application Application # of TCP sessions # of TCP sessions No operation No operation 5 5～～10 10 Yahoo Yahoo top page top page 10 10～～20 20 Google Google image search image search 30 30～～60 60 ニコニコ動画ニコニコ動画 50 50～～80 80 OCN OCN photo friend photo friend 170 170～～200+ 200+ iTunes iTunes 230 230～～270 270 iGoogle iGoogle 80 80～～100 100 楽天楽天( (Rakuten Rakuten) ) 50 50～～60 60 Amazon Amazon 90 90 HMV HMV 100 100 YouTube YouTube 90 90

SLIDE 29

29 29

TAC2000/2000.7

802.16 IP Telephony Lab

This may be why Google has turned on IPv6

SLIDE 30

30 30

TAC2000/2000.7

802.16 IP Telephony Lab

IPv6 Design Goals

Addressing and Routing

Addressing and Routing

Address Space

Address Space

Header Format

Header Format

Minimizing Administrative Workload

Minimizing Administrative Workload

Auto

Auto-

configuration

configuration

Multi

Multi-

media Support

media Support

Security

Security

AH/ESP

AH/ESP

Mobility

Mobility

Transition Mechanism

Transition Mechanism

SLIDE 31

31 31

TAC2000/2000.7

802.16 IP Telephony Lab

IPv6 Address

128

128-

bit long

bit long

2

2128

128 = 3.4

= 3.4× ×10 1038

38 →

→ 665

665× ×10 1021

21 addresses per m

addresses per m2

2 of

f

earth surface. earth surface.

Considering some inefficient usage (e.g. 911

Considering some inefficient usage (e.g. 911 prefix in telephony), it is estimated to support prefix in telephony), it is estimated to support 8 8× ×10 1017

17 to 2

to 2× ×10 1033

33 addresses.

addresses.

8

8× ×10 1017

17 →

→ 1,564 address per

1,564 address per m m2

2.

.

SLIDE 32

32 32

TAC2000/2000.7

802.16 IP Telephony Lab

Text Representation of Addresses

Colon

Colon-

Hex

Hex 2001:e10:6840:20:20f:eaff:fe56:ea22

2001:e10:6840:20:20f:eaff:fe56:ea22

Compressed Format:

Compressed Format:

2001:0b00:0c18:0001:0000:0000:0000:0010 2001:0b00:0c18:0001:0000:0000:0000:0010 becomes becomes 2001:b00:c18:1::10 2001:b00:c18:1::10

IPv4

IPv4-

compatible:

compatible:

0:0:0:0:0:0:163.22.2.1

0:0:0:0:0:0:163.22.2.1

r ::163.22.2.1
r ::163.22.2.1
6to4 Address

6to4 Address

2002:8C6E:3C2E::8C6E:3C2E

2002:8C6E:3C2E::8C6E:3C2E

140.110.60.46 = 8C6E:3C2E

140.110.60.46 = 8C6E:3C2E

SLIDE 33

33 33

TAC2000/2000.7

802.16 IP Telephony Lab

Multicast Support

Multicast is optional for IPv4. Not all routers support

Multicast is optional for IPv4. Not all routers support multicast. multicast.

All IPv6 hosts and routers are required to support

All IPv6 hosts and routers are required to support multicast. multicast.

There are no broadcast addresses in IPv6, their

There are no broadcast addresses in IPv6, their function being superseded by multicast addresses. function being superseded by multicast addresses.

Link

Link-

local

local

Site

Site-

local (

local (obsoleted

bsoleted by Unique

by Unique-

Local)

Local)

Global scope

Global scope

SLIDE 34

34 34

TAC2000/2000.7

802.16 IP Telephony Lab

Multicast IPv6

Special multicast IPv6 address

Special multicast IPv6 address

FF01::1

FF01::1

Node

Node-

local scope all

local scope all-

nodes multicast address

nodes multicast address

FF02::1

FF02::1

Link

Link-

local scope all

local scope all-

nodes

nodes multicast address multicast address

FF01::2

FF01::2

Node

Node-

local scope all

local scope all-

routers multicast address

routers multicast address

FF02::2

FF02::2

Link

Link-

local scope all

local scope all-

Routers multicast address

Routers multicast address

FF05::5

FF05::5

site

site-

local scope

local scope all all-

routers

routers multicast address multicast address

Use low

Use low-

order 32 bits, each group ID maps to a unique Ethernet MAC addre
rder 32 bits, each group ID maps to a unique Ethernet MAC address(RFC

ss(RFC 2373) 2373)

4 32 bits 8 group ID scope flags 11111111 4 80 000……000

SLIDE 35

35 35

TAC2000/2000.7

802.16 IP Telephony Lab

Example on FreeBSD

$ $ ping6 ping6 -

c 5 FF02::2%em0

c 5 FF02::2%em0 PING6(56=40+8+8 bytes) fe80::20f:eaff:fe4e:6a8c%em0 PING6(56=40+8+8 bytes) fe80::20f:eaff:fe4e:6a8c%em0 --

-> ff02::2%em0

> ff02::2%em0 16 bytes from fe80::20d:28ff:fe49:bea0%em0, 16 bytes from fe80::20d:28ff:fe49:bea0%em0, icmp_seq icmp_seq=0 =0 hlim hlim=64 time=0.715 ms =64 time=0.715 ms 16 bytes from fe80::20d:65ff:fee9:6c00%em0, 16 bytes from fe80::20d:65ff:fee9:6c00%em0, icmp_seq icmp_seq=0 =0 hlim hlim=64 time=0.862 ms =64 time=0.862 ms(DUP!) (DUP!) 16 bytes from fe80::20d:28ff:fe49:bea0%em0, 16 bytes from fe80::20d:28ff:fe49:bea0%em0, icmp_seq icmp_seq=1 =1 hlim hlim=64 time=0.613 ms =64 time=0.613 ms 16 bytes from fe80::20d:65ff:fee9:6c00%em0, 16 bytes from fe80::20d:65ff:fee9:6c00%em0, icmp_seq icmp_seq=1 =1 hlim hlim=64 time=0.860 ms =64 time=0.860 ms(DUP!) (DUP!) 16 bytes from fe80::20d:28ff:fe49:bea0%em0, 16 bytes from fe80::20d:28ff:fe49:bea0%em0, icmp_seq icmp_seq=2 =2 hlim hlim=64 time=0.610 ms =64 time=0.610 ms 16 bytes from fe80::20d:65ff:fee9:6c00%em0, 16 bytes from fe80::20d:65ff:fee9:6c00%em0, icmp_seq icmp_seq=2 =2 hlim hlim=64 time=0.745 ms =64 time=0.745 ms(DUP!) (DUP!) 16 bytes from fe80::20d:28ff:fe49:bea0%em0, 16 bytes from fe80::20d:28ff:fe49:bea0%em0, icmp_seq icmp_seq=3 =3 hlim hlim=64 time=0.730 ms =64 time=0.730 ms 16 bytes from fe80::20d:65ff:fee9:6c00%em0, 16 bytes from fe80::20d:65ff:fee9:6c00%em0, icmp_seq icmp_seq=3 =3 hlim hlim=64 time=0.864 ms =64 time=0.864 ms(DUP!) (DUP!) 16 bytes from fe80::20d:28ff:fe49:bea0%em0, 16 bytes from fe80::20d:28ff:fe49:bea0%em0, icmp_seq icmp_seq=4 =4 hlim hlim=64 time=0.721 ms =64 time=0.721 ms

-- FF02::2%em0 ping6 statistics

FF02::2%em0 ping6 statistics ---

5 packets transmitted, 5 packets received, +4 duplicates, 0.0% p

5 packets transmitted, 5 packets received, +4 duplicates, 0.0% packet loss acket loss round round-

trip min/

trip min/avg avg/max/std /max/std-

dev = 0.610/0.747/0.864/0.093 ms

dev = 0.610/0.747/0.864/0.093 ms

SLIDE 36

36 36

TAC2000/2000.7

802.16 IP Telephony Lab

Example on Linux

$ $ ping6 ping6 -

I eth0

I eth0 -

c 2 ff02::1

c 2 ff02::1 PING ff02::1(ff02::1) from fe80::21a:64ff:fec2:6fb8 eth0: 56 dat PING ff02::1(ff02::1) from fe80::21a:64ff:fec2:6fb8 eth0: 56 data bytes a bytes 64 bytes from fe80::21a:64ff:fec2:6fb8: 64 bytes from fe80::21a:64ff:fec2:6fb8: icmp_seq icmp_seq=0 =0 ttl ttl=64 time=0.032 ms =64 time=0.032 ms 64 bytes from fe80::202:55ff:fed6:d904: 64 bytes from fe80::202:55ff:fed6:d904: icmp_seq icmp_seq=0 =0 ttl ttl=64 time=0.149 ms =64 time=0.149 ms (DUP!) (DUP!) 64 bytes from fe80::214:5eff:fe22:52c9: 64 bytes from fe80::214:5eff:fe22:52c9: icmp_seq icmp_seq=0 =0 ttl ttl=64 time=0.155 ms =64 time=0.155 ms (DUP!) (DUP!) 64 bytes from fe80::20d:60ff:fe6d:3d10: 64 bytes from fe80::20d:60ff:fe6d:3d10: icmp_seq icmp_seq=0 =0 ttl ttl=64 time=0.157 ms =64 time=0.157 ms (DUP!) (DUP!) 64 bytes from fe80::216:17ff:fec0:d718: 64 bytes from fe80::216:17ff:fec0:d718: icmp_seq icmp_seq=0 =0 ttl ttl=64 time=0.249 ms =64 time=0.249 ms (DUP!) (DUP!) 64 bytes from fe80::221:5eff:fe57:9ea0: 64 bytes from fe80::221:5eff:fe57:9ea0: icmp_seq icmp_seq=0 =0 ttl ttl=64 time=0.253 ms =64 time=0.253 ms (DUP!) (DUP!) 64 bytes from fe80::216:17ff:fec0:d718: 64 bytes from fe80::216:17ff:fec0:d718: icmp_seq icmp_seq=0 =0 ttl ttl=64 time=0.255 ms =64 time=0.255 ms (DUP!) (DUP!) 64 bytes from fe80::ce01:80ff:fe02:3a76: 64 bytes from fe80::ce01:80ff:fe02:3a76: icmp_seq icmp_seq=0 =0 ttl ttl=64 time=0.257 ms =64 time=0.257 ms (DUP!) (DUP!)

SLIDE 37

37 37

TAC2000/2000.7

802.16 IP Telephony Lab

IPv4 Header

20 Octets+Options : 13 fields, include 3 flag bits

0 bits 31 Ver IHL Total Length Identifier Flags Fragment Offset 32 bit Source Address 32 bit Destination Address 4 8 24 16 Service Type Options and Padding Time to Live Header Checksum Protocol

Removed Changed

SLIDE 38

38 38

TAC2000/2000.7

802.16 IP Telephony Lab

IPv6 Header

40 Octets, 8 fields

31 Version Priority Flow Label Payload Length Next Header Hop Limit 128 bit Source Address 128 bit Destination Address 4 12 24 16

SLIDE 39

39 39

TAC2000/2000.7

802.16 IP Telephony Lab

IPv6 enables efficient routing

Total IPv6 header size is only twice as large.

Total IPv6 header size is only twice as large.

Most IPv6 extension headers are not examined or

Most IPv6 extension headers are not examined or processed by intermediate nodes (in contrast with IPv4, processed by intermediate nodes (in contrast with IPv4, where IP options typically cause a major performance where IP options typically cause a major performance loss for the packet at every intermediate router). loss for the packet at every intermediate router).

Priority

Priority and and flow label flow label can be used to identify flows can be used to identify flows even when the payload is encrypted. even when the payload is encrypted.

Traditionally, routers must inspect layer

Traditionally, routers must inspect layer-

4 headers to get

4 headers to get QoS QoS information. information.

SLIDE 40

40 40

TAC2000/2000.7

802.16 IP Telephony Lab

Application-level Security Solutions

SSH for TELNET

SSH for TELNET

SSL for WWW

SSL for WWW

PGP for email

PGP for email The contents may be encrypted, but the traffic flow can The contents may be encrypted, but the traffic flow can still be observed. still be observed. IPv6 provides Network IPv6 provides Network-

level security.

level security. IPsec IPsec is a mandatory part of IPv6, and is optional for use is a mandatory part of IPv6, and is optional for use with IPv4. with IPv4.

SLIDE 41

41 41

TAC2000/2000.7

802.16 IP Telephony Lab

At boot time, an IPv6 host build a Link-Local address, then its global IPv6 address(es) from RA

RA indicates SUBNET PREFIX

IPv6 Auto-Configuration

Stateless (RFC2462)

Stateless (RFC2462)

Host autonomously configures its own

Host autonomously configures its own Link Link-

Local

Local address address

Router

Router solicitation solicitation are sent by are sent by booting booting nodes nodes to to request request RAs RAs for for configuring configuring the the interfaces. interfaces.

Stateful

Stateful

DHCPv6

DHCPv6 ( (RFC 3315 RFC 3315) )

Renumbering

Renumbering

Hosts Hosts renumbering renumbering is is done done by by modifying modifying the the RA to RA to announce announce the the old

ld prefix

prefix with with a short a short lifetime lifetime and and the the new new prefix prefix. . Router renumbering protocol Router renumbering protocol (RFC 2894) (RFC 2894), to allow , to allow domain domain-

interior routers to learn of prefix introduction

interior routers to learn of prefix introduction / withdrawal / withdrawal

SUBNET PREFIX + MAC ADDRESS SUBNET PREFIX + MAC ADDRESS SUBNET PREFIX + MAC ADDRESS SUBNET PREFIX + MAC ADDRESS SUBNET PREFIX + MAC ADDRESS SUBNET PREFIX + MAC ADDRESS SUBNET PREFIX + MAC ADDRESS SUBNET PREFIX + MAC ADDRESS

SLIDE 42

42 42

TAC2000/2000.7

802.16 IP Telephony Lab

Dual Stack Approach & DNS

In a dual stack case, an application that:

In a dual stack case, an application that:

Is IPv4 and IPv6

Is IPv4 and IPv6-

enabled

enabled

Asks the DNS for all types of addresses

Asks the DNS for all types of addresses

Chooses one address and, for example, connects to the

Chooses one address and, for example, connects to the IPv6 address IPv6 address

DNS Server IPv4 IPv6 www.a.com = * ? 2001:DB8 ::1 2001:DB8::1 10.1.1.1

SLIDE 43

43 43

TAC2000/2000.7

802.16 IP Telephony Lab

Cisco IOS Dual Stack Configuration

Cisco IOS is IPv6

Cisco IOS is IPv6-

enable

enable: :

If IPv4 and IPv6 are configured on

If IPv4 and IPv6 are configured on one

ne interface,

interface, the router the router is is dual dual-

stacked

stacked

Telnet, Ping,

Telnet, Ping, Traceroute Traceroute, SSH, DNS client, TFTP, , SSH, DNS client, TFTP,… …

IPv6 and IPv4 Network Dual-Stack Router

IPv4: 140.110.199.1 IPv6: 2001:C58:213:1::/64 eui-64

router# interface Ethernet0 ip address 140.110.199.1 255.255.255.0 ipv6 address 2001:C58:213:1::/64 eui-64

SLIDE 44

44 44

TAC2000/2000.7

802.16 IP Telephony Lab

WWW Server & Browser

SLIDE 45

45 45

TAC2000/2000.7

802.16 IP Telephony Lab

Summary

IPv6 integrates many built

IPv6 integrates many built-

in features which are optional and

in features which are optional and sometimes mutually sometimes mutually-

exclusive in IPv4.

exclusive in IPv4.

Many routers and operating systems support IPv6 now.

Many routers and operating systems support IPv6 now.

Windows 2000/XP/Vista/7

Windows 2000/XP/Vista/7

Linux

Linux

FreeBSD

FreeBSD

Mac OS

Mac OS

Android 2.1

Android 2.1

Transition mechanism is crucial to the success of IPv6.

Transition mechanism is crucial to the success of IPv6.

Dual

Dual-

Stack

Stack

Tunneling

Tunneling

Translation

Translation

SLIDE 46

46 46

TAC2000/2000.7

802.16 IP Telephony Lab

IPv6 in IETF

IPv6 WG (formerly known as

IPv6 WG (formerly known as IPng IPng WG) is developing WG) is developing IPv6. IPv6.

IPv6 was standardized as RFC 2460 (in 1998) and

IPv6 was standardized as RFC 2460 (in 1998) and many related many related RFCs RFCs. .

Joint meeting with 3GPP was held in May 2001 and

Joint meeting with 3GPP was held in May 2001 and published an internet draft published an internet draft “ “Recommendations for IPv6 Recommendations for IPv6 in 3GPP Standards. in 3GPP Standards.” ”

“

“Minimum IPv6 Functionality for Cellular Host Minimum IPv6 Functionality for Cellular Host” ” is is under discussion. under discussion.

SLIDE 47

47 47

TAC2000/2000.7

802.16 IP Telephony Lab

IPv6 in 3GPP

GPRS supports IPv6 transport service

GPRS supports IPv6 transport service

ptionally.
ptionally.
3GPP decided to support IPv6 exclusively for

3GPP decided to support IPv6 exclusively for IP Multimedia Subsystem (IMS) in June 2000. IP Multimedia Subsystem (IMS) in June 2000.

IMS is being specified as part of 3GPP Release

IMS is being specified as part of 3GPP Release 5 specification. 5 specification.

Collaboration agreement with IETF in June

Collaboration agreement with IETF in June 2001. 2001.

SLIDE 48

48 48

TAC2000/2000.7

802.16 IP Telephony Lab

IPv6 Implementation - Host

Apple

Apple

Compaq

Compaq

FreeBSD

FreeBSD

Hitachi, Ltd.

Hitachi, Ltd.

HP

HP

IBM

IBM

Linux
Microsoft
Mentat
SGI
Sun
etc.

SLIDE 49

49 49

TAC2000/2000.7

802.16 IP Telephony Lab

IPv6 Implementation - Router

3

3Com Com

6Wind

6Wind

Cisco Systems

Cisco Systems

Hitachi, Ltd.

Hitachi, Ltd.

NTHU

NTHU

Nokia

Nokia

Nortel Networks

Nortel Networks

Ericsson/

Ericsson/Telebit Telebit Communications Communications

Juniper Networks

Juniper Networks

http://playground.sun.com/pub/ipng/html/ipng-implementations.html

SLIDE 50

50 50

TAC2000/2000.7

802.16 IP Telephony Lab

Former Prime Minister of

Former Prime Minister of Japan Japan -

Yoshiro Mori

Yoshiro Mori

Governmental

Governmental

IPv6 Council

IPv6 Council

JGN (Japan Gigabit Network)

JGN (Japan Gigabit Network) IPv6 IPv6

Industrial

Industrial

IAJapan

IAJapan IPv6 Deployment IPv6 Deployment Committee Committee

JPNIC IPv6 project

JPNIC IPv6 project

100% IPv6 readiness by 2005

Japan

SLIDE 51

51 51

TAC2000/2000.7

802.16 IP Telephony Lab

NTT-JP

SLIDE 52

52 52

TAC2000/2000.7

802.16 IP Telephony Lab

Live E! is a consortium that promotes the deployment of new infrastructure

that can generate, collect, process, and share all the “Environmental Information”, associated with the Earth

As the first step, we picked up "Digital Weather Station“
Individuals, non-commercial and commercial organization install sensor

nodes, and let the information available from anyone on the Internet.

Larger number of participation leads richer information and for all, and

creates innovative applications and usage of information.

Single information can be used multiple purposes

Chair :

Hiroshi ESAKI (Univ. of Tokyo)

Co-Chair :

Reiji AIHARA （（（（Hiroshima Univ.））））

! "#$%$

&'! &()* +$, -.* + &,/$01+*2/+23$%$ ++ .45..* +6763+*2+ 3 45..+*2/+ %% "8$%,$ /"%, $.%$ "" + ,!81!%$ 9.!(.,.$, .!* + Cooperation Organization:

Digital Weather Station

①Education Materials

②Public Services ③Business applications

Live E! Project

Environmental Information System Environmental Information System

(http://www.live-e.org/)

SLIDE 53

53 53

TAC2000/2000.7

802.16 IP Telephony Lab

New Zealand will install Indonesia has installed Taiwan has installed Thailand has installed

SLIDE 54

54 54

TAC2000/2000.7

802.16 IP Telephony Lab

Four Sensor Nodes installed in Taiwan Four Sensor Nodes installed in Taiwan

SLIDE 55

55 55

TAC2000/2000.7

802.16 IP Telephony Lab

IPv6 Deployment Policy in Korea IPv6 Deployment Policy in Korea

Transition Roadmap by Government (23 Feb 2001)

Transition Roadmap by Government (23 Feb 2001)

Complete native IPv6 Commercial IPv6 Service (wire/wireless)

Phase I (~2001) Phase I (~2001) Phase II (2002~2005) Phase II (2002~2005) Phase III (2006~2010) Phase III (2006~2010) Phase IV (2011~) Phase IV (2011~) IPv4 Only IPv4 Ocean IPv6 Island IPv4 Island IPv6 Ocean IPv6 Only

Validation
Operation
Promotion
IMT2000 Service
Translation Service

IPv4/IPv6 Translation Required

Experimental IPv6 Network

SLIDE 56

56 56

TAC2000/2000.7

802.16 IP Telephony Lab

Korean IPv6 Network

Experimental IPv6 Network

– 6Bone-KR (since 1998) *www.6bone.ne.kr

Research & Education IPv6 Network

– KOREN IPv6 Network (Since 1999) – KREONET2 IPv6 Network (Since 1999) – TEIN IPv6 Network (Since 2001) 6GN (Gigabit IPv6 Infrastructure) (Since 2004)

IPv6 Internet Exchange

– 6NGIX (IPv6 Next Generation Internet Exchange) (Since 2001)

Commercial IPv6 Network & Trials

– Pre-Commercial IPv6 networks – KOREAv6 Pilot Project Network (Since 2004)

SLIDE 57

57 57

TAC2000/2000.7

802.16 IP Telephony Lab

Gigabit IPv6 Infrastructure: 6GN

11 members
2 backbones
1 IX

2G/6 CNU KISTI 1G/6 155M/6 1G/6 1G/6 1G/6 1G/6 1G/6 1G/6 1G/6 GIST HYU 1G/6 1G/6 KNU 1G/6 1G/6 1G/6 6NGIX KT TNL 1G/6 NCA 1G/6

6KANet

2G/6

KOREN KREONET2

ICU ETRI KAIST 1G/6 1G/6 KMA

Phase I Phase II

SLIDE 58

58 58

TAC2000/2000.7

802.16 IP Telephony Lab

Multihoming test on 6GN

ET-SHIM6 Testbed

ET-SHIM6 - Stream Sever ET-SHIM6 – Stream Client

. 6:8$ ;$ <-1 <-1

3 3 3 3 Multihomed Sites

SLIDE 59

59 59

TAC2000/2000.7

802.16 IP Telephony Lab

CERNETv6

SLIDE 60

60 60

TAC2000/2000.7

802.16 IP Telephony Lab 40 GPOS 10GPOS 2.5GPOS 比威 BE12016 华为 NE80 华为 NE5000 Juniper T640 CiscoCRS Hitachi GR4000 北京－清华西安武汉合肥上海－交大广州成都沈阳天津重庆厦门南京兰州大连哈尔滨长春济南杭州长沙郑州北京－北大北京－北邮北京－北航上海－复旦上海－同济 CNGI -6IX 上海 CNGI -6IX 北京

SLIDE 61

61 61

TAC2000/2000.7

802.16 IP Telephony Lab

SLIDE 62

62 62

TAC2000/2000.7

802.16 IP Telephony Lab

Australia

AARNet

AARNet (Australian Academic and Research Network) core (Australian Academic and Research Network) core and edge is dual stack since 2003 and edge is dual stack since 2003

Dual stack deployed across Juniper M320 core using OSPF3

Dual stack deployed across Juniper M320 core using OSPF3 and BGP as routing protocols and BGP as routing protocols

IPv6 is used within AARNet

IPv6 is used within AARNet

www.aarnet.edu.au

www.aarnet.edu.au IPv6 enabled IPv6 enabled

Infrastructure is IPv6 enabled

Infrastructure is IPv6 enabled

IPv6 Multicast is enabled

IPv6 Multicast is enabled

SSM supported

SSM supported

Currently use a static RP for ASM

Currently use a static RP for ASM

SLIDE 63

63 63

TAC2000/2000.7

802.16 IP Telephony Lab

SLIDE 64

64 64

TAC2000/2000.7

802.16 IP Telephony Lab

Still work to be done…

DNS about to be implemented

DNS about to be implemented

Mail issues

Mail issues – – 3 3rd

rd parties?

parties?

Still need to deploy IPv6 measurement

Still need to deploy IPv6 measurement

IPv6 monitoring is still in its infancy within our

IPv6 monitoring is still in its infancy within our infrastructure infrastructure

IPv4

IPv4 Netflow Netflow is heavily deployed is heavily deployed – – IPv6 isn IPv6 isn’ ’t at the t at the moment moment

SLIDE 65

65 65

TAC2000/2000.7

802.16 IP Telephony Lab

USA

IPv6 continues to perform well on the Internet2

IPv6 continues to perform well on the Internet2 Network Network

Transition from Abilene layer3 network to new

Transition from Abilene layer3 network to new Internet2 network infrastructure almost complete [done Internet2 network infrastructure almost complete [done by 30 by 30-

Sep]

Sep]

Using same Juniper T640 routers; still running dual

Using same Juniper T640 routers; still running dual-

stack IPv4 and IPv6

stack IPv4 and IPv6

The backbone now supports 32 bit

The backbone now supports 32 bit ASNs ASNs

SLIDE 66

66 66

TAC2000/2000.7

802.16 IP Telephony Lab

Monitoring

Monitoring and network test points still being updated

Monitoring and network test points still being updated for the new network [Should be finished by end Sept.] for the new network [Should be finished by end Sept.]

We will have at every router node

We will have at every router node

Latency (

Latency (owamp

wamp)

)

Throughput (

Throughput (bwctl bwctl) [1G, 10G upon request] ) [1G, 10G upon request]

On

On-

demand testing (NDT)

demand testing (NDT)

All supporting IPv6

All supporting IPv6

We also have firewall filters installed in the Juniper

We also have firewall filters installed in the Juniper routers to capture IPv6 utilization (and other specific routers to capture IPv6 utilization (and other specific port counts) port counts)

SLIDE 67

67 67

TAC2000/2000.7

802.16 IP Telephony Lab

IPv6 Utilization

http://vixen.grnoc.iu.edu/jfirewall-viz/index-bits.html

SLIDE 68

68 68

TAC2000/2000.7

802.16 IP Telephony Lab

IPv6 Address Allocations

RFC 3177

RFC 3177

/32

/32 for an ISP for an ISP

/48

/48 for an organization in the general case for an organization in the general case

/64

/64 when it is known that one and only one subnet is needed when it is known that one and only one subnet is needed

/128

/128 when it is absolutely known that one and only one when it is absolutely known that one and only one device is connecting device is connecting

This document also describe the advantage of the fixed

This document also describe the advantage of the fixed boundary specifically at /48 boundary specifically at /48

SLIDE 69

69 69

TAC2000/2000.7

802.16 IP Telephony Lab

Initial IPv6 Prefix Allocation for RIRs Initial IPv6 Prefix Allocation for RIRs

IPv6 Prefix Range Assignment 2001:0000::/29-2001:01F8::/29 IANA 2001:0200::/29-2001:03F8::/29 APNIC 2001:0400::/29-2001:05F8::/29 ARIN IPv6 Prefix Range Assignment 2001:0600::/29-2001:07F8::/29 RIPE NCC 2001:1200::/29-2001:13F8::/29 LACNIC

International IPv6 Address Management International IPv6 Address Management

SLIDE 70

70 70

TAC2000/2000.7

802.16 IP Telephony Lab

IANA IPv6 Allocations to RIRs

issued in Oct 2006 2800:0000::/12 2800:0000::/12 LACNIC LACNIC 2A00:0000::/12 2A00:0000::/12 RIPE NCC RIPE NCC 2600:0000::/12 2600:0000::/12 ARIN ARIN 2400:0000::/12 2400:0000::/12 APNIC APNIC 2C00:0000::/12 2C00:0000::/12 AfriNIC AfriNIC

IPv6 Address IPv6 Address RIR RIR

Some /23s from the previous slide are incorporated in these /12s

SLIDE 71

71 71

TAC2000/2000.7

802.16 IP Telephony Lab

IANA IPv6 Allocations to RIRs

issued as /23s prior to Oct 2006

1 73 13 2 198 50 100 150 200 250 AfriNIC APNIC ARIN LACNIC RIPE NCC

http://www.apnic.net/info/reports/index.html

SLIDE 72

72 72

TAC2000/2000.7

802.16 IP Telephony Lab

IANA IPv6 Allocations to RIRs

Mar 2008

http://www.apnic.net/info/reports/index.html

=
'
6#

6 6 26

SLIDE 73

73 73

TAC2000/2000.7

802.16 IP Telephony Lab

2001:238::/32

2001:238::/32-

HiNet

HiNet

2001:288::/32

2001:288::/32-

TANet

TANet

2001:C08::/32

2001:C08::/32-

ASNet

ASNet

2001:C50::/32

2001:C50::/32-

TTN

TTN

2001:C58::/32

2001:C58::/32-

6REN

6REN

2001:CA0::/32

2001:CA0::/32-

CHT TL

CHT TL

2001:CD8::/32

2001:CD8::/32-

SeedNet

SeedNet

2001:D20::/32

2001:D20::/32-

TFN

TFN

2001:D40::/32

2001:D40::/32-

TW NTT

TW NTT

2001:E10::/32

2001:E10::/32-

TWAREN

TWAREN

2001:ED8::/32

2001:ED8::/32-

ITRI

ITRI

2001:F18::/32

2001:F18::/32-

NCTU

NCTU

SLIDE 77

77 77

TAC2000/2000.7

802.16 IP Telephony Lab

Conclusion

The recent consumption rates of IPv4 will deplete the

The recent consumption rates of IPv4 will deplete the central pool before the end of this decade. central pool before the end of this decade.

Organizations would be wise to start the process of

Organizations would be wise to start the process of planning for an IPv6 deployment now. planning for an IPv6 deployment now.

Those who delay may find that the IANA pool for

Those who delay may find that the IANA pool for IPv4 has been exhausted before they have completed IPv4 has been exhausted before they have completed their move to IPv6. their move to IPv6.

Students are advised to learn IPv6 programming,

Students are advised to learn IPv6 programming, because we foresee an urgent demand after two years. because we foresee an urgent demand after two years.