IPv6-workshop-tm-gdv-f

© 2012 Cisco and/or its affiliates. All rights reserved. 1

Cisco “Tech Session” Preparing for BYOD & the Internet of Everything, an IPv6 Workshop

Tim Martin

CCIE #2020

Solutions Architect

Spring 2014


•  BYOD and the Internet of Everything •  IPv6 Addressing Deep Dive •  IPv6 Design Considerations •  IPv6 Campus Design •  IPv6 Data Center Transitions •  IPv6 Translation Techniques •  IPv6 Internet Edge Design •  Summary


89%

10%

1% 23%

36%

26% 75%

22%

Desktops Smart Phones Tablets •  Boomers are retiring, GenX is “tech savvy”, GenY is “tech dependent”

•  2016 GenY (the millennia's (18-34)) become the largest workforce segment

•  43% of 18-24 year-olds say that texting is just as meaningful as a phone conversation -eMarketer

•  40% of GenY believe that blogging about workplace issues is acceptable –Iconoculture

•  24% of GenY say that technology use is what makes their generation unique -Pew Research

•  74% of GenY used a smartphone for work purposes in the last year, compared to 37 percent of Baby Boomers -CompTIA

© 2012 Cisco and/or its affiliates. All rights reserved. 4 4


National IPv6 Strategies STEM

IPv6

IPv4 Address Depletion

Infrastructure Evolution IPv6 OS, Content & Applications

2011

•  Eliminate Complexities Associated with RFC 1918, NAT, ALG’s, CIDR

Mandate

Pref. by App’s in W7, S2008, OSX 4G, DOCSIS 3.0, CGN


• Early Adopters, from ~2001-2005 (6bone) • Chasm, Refinement from 2005-2009 (Tunneling) • Early Majority, Launch June 2012 (Transitioning)


§  Standards - Leadership in IP protocols within IETF & IPv6 development

§  Experience - Professional Services offering years of experience in IPv6

§  Solutions - Innovation, Feature Acceleration




340,282,366,920,938,463,374,607,432,768,211,456 (IPv6 Address Space - 340 undecillion, 282 decillion, 366 nonillion, 920 octillion, 938 septillion, 463 sextillion, 463 quintillion, 374 quadrillion, 607 trillion, 431 billion, 768 million, 211 thousand and 456

vs 4,294,967,296 (IPv4 Address Space - 4 Billion)

•  Lot’s of talk about how big, it’s BIG, do NOT worry about waste

•  Each /64 prefix contains 18 Quintillion host address’s (18,446,744,073,709,551,616)

•  Theoretical vs. Practical deployment, still not an issue

Antares 15th Brightest star in the sky

Our Sun


•  IPv6 has a specific Ethernet Protocol ID •  IPv6 relies heavily on Multicast

Destination Ethernet Address!

Source Ethernet Address!

0x0800!!

IPv4 Header and Payload!

Destination Ethernet Address!

Source Ethernet Address!

0x86DD!!

IPv6 Header and Payload!

xx 33 33 xx xx xx

I bit = Local Admin, L bit = Multicast/Broadcast

0000 00IL


Fragment Offset Flags

Total Length Type of Service IHL

Padding Options Destination Address

Source Address

Header Checksum Protocol Time to Live

Identification

Version

IPv4 Header (20-60)

Next Header Hop Limit

Flow Label Traffic Class

Destination Address

Source Address

Payload Length

Version

IPv6 Header (40)

•  Length is constant in IPv6 •  Fragmentation occurs in (EH)

•  Option’s occur in (EH) •  UDP must have valid Checksum, unlike v4.

•  Upper layer checksums use the Pseudo Header format: SRC/DST Addr + Next Header


Extension Header * Type Hop-by-Hop Options 0

Routing Header 43

Destination Options 60

Fragment Header 44

ESP Header 50

Authentication Header 51

Mobility Header 135

Shim6 140

Experimental 253,254

No Next Header 59

IPv6 Header Hop-by-Hop Destination Opt TCP Header Payload

•  EH are daisy chained, processed in order •  Length is variable, must be on 8 byte boundary, typically 24 bytes •  If HbH is present, must be first, must be processed, likely in SW


Type Code Data

Checksum

•  Neighbor Discovery, Router Discovery, Path MTU Discovery and (MLD) Type – (1-127) = Error Messages, (128-255) = Informational Messages Code – More Granularity within the Type Checksum – computed over the entire ICMPv6 Data - Original Header Return (8 bytes), then fill to Min MTU (1280)

58

1 Destination Unreachable

2 Packet Too Big

3 Time Exceeded

4 Parameter Problem ICMPv6 Header

Next Header *58, not 1 (ICMP)


•  Will break ICMP error message rule (by responding for Multicast)

•  Must set ALL interfaces to 1280 MTU if you disallow PTB at your FW Source Destination Link

MTU 1500 MTU 1500 MTU 1400 MTU 1300

Packet, MTU=1500

ICMPv6 Packet Too Big, Use MTU=1400

Packet, MTU=1400

ICMPv6 Type 2 PTB, Use MTU=1300

Packet, MTU=1300


IPv6 Address Family

Multicast Anycast Unicast

Assigned Solicited Node

Unique Local Link Local Global Special Embedded

*IPv6 does not use broadcast addressing

Well Known Temp


•  IPv6 addresses are 128 bits long Segmented into 8 groups of 16 bits separated by (:) 32 HEX characters – CAsE DoEs not mAttEr •  It’s a Prefix, not a mask, •  Word, Group or Quad..

Host Id

:HHHH:HHHH:HHHH:HHHH Host Portion

Subnet Id

SSSS Global Routing Prefix

NNNN:NNNN:NNNN:

Network Portion

2001:0DB8:1010: : :0000:0000:0000:1E2A 00A4


•  Recommended Alloca,ons •  Consumer, SMB /56 /60 /64 •  Municipal Government, Enterprise, Single AS /48 •  State Governments, Universi,es (LIR) /32 /36 /40 /44 /48

•  Addressing Plan, Site Count •  IPv4 Allocation, Multi-homed ISP •  Point of Contact, Org ID •  Submit, Verify •  Review, Officer Certification •  Approval, Fees Paid, Assignment

Registries

Level Four Entity

IANA

ISP Org

PA

/48

2000::/3

/12

/32

2000::/3

/48

/12

PI

/32 ARIN


•  Leading 0’s can be omitted

•  The double colon (::) can appear only once

2001:0DB8:0000: :0000:0000:0000:1E2A 00A4 Full Format

2001:DB8:0: :0:0:0:1E2A A4 Abbreviated Formats

2001:DB8:0: ::1E2A A4


Link-Local – Non routable exists on single layer 2 domain (FE80::/64) FE80:0000:0000:0000

:: xxxx:xxxx:xxxx:xxxx

FC0g:gggg:gggg: xxxx:xxxx:xxxx:xxxx ssss:

FD0g:gggg:gggg: xxxx:xxxx:xxxx:xxxx ssss:

Unique-Local – Routable within administrative domain (FC00::/7)

2000:NNNN:NNNN HHHH:HHHH:HHHH:HHHH Global – Routable across the Internet (2000::/3)

:SSSS:

3FFF:NNNN:NNNN HHHH:HHHH:HHHH:HHHH :SSSS:


Similar to IPv4 New in IPv6

Manually configured StateLess Address AutoConfiguration SLAAC EUI64

SLAAC Privacy Addressing

Assigned via DHCPv6

*Secure Neighbor Discovery SeND


00 90 27 FF FE 17 FC 0F

OUI Device Identifier

00 90 27 17 FC 0F

02 90 27 FF FE 17 FC 0F

0000 00U0 U= 1 = Universel/unique

0 = Local/not unique U bit must be flipped

FF FE 00 90 27 17 FC 0F


•  Generated on unique 802 using MD5, then stored for next iteration •  Enabled by default in Windows, Android, iOS, Mac OS/X, Linux •  Temporary or Ephemeral addresses for client application (web browser)

Recommendation: Good for the mobile user, but not for your organization/corporate networks (Troubleshooting and accountability)

23

2001 DB8

/32 /48 /64

Random Generated Interface ID 0000 1234


DHCP Messages IPv4 IPv6

Initial Message Exchange 4-way handshake 4-way handshake

Message Types Broadcast, Unicast Multicast, Unicast

Client à Server (1) DISCOVER SOLICIT (any servers)

Server à Client (2) OFFER ADVERTISE (want this address)

Client à Server (3) REQUEST REQUEST (I want that address)

Server à Client (4) ACK REPLY (It’s yours)

•  Digital Millennium Copyright Act (DMCA), HIPAA (health), PCI (credit card) •  FF02::1:2 = All DHCP Agents (servers or relays, Link-local scope) •  FF05::1:3 = All DHCP Servers (Site-local scope) •  Clients listen on UDP port 546; Servers/relays on UDP port 547 •  Rapid Commit, 2 packet exchange. Solicit/Reply, client sets for options •  ipv6 dhcp relay destination replaces ip helper address


•  Always uses Link Local (FE80::/64) as its source

•  Maps Layer 3 IPv6 address to Layer 2 MAC address

•  LINK OPERATIONS (control plane) •  Neighbor discovery messages

•  Router solicitation (ICMPv6 type 133) •  Router advertisement (ICMPv6 type 134) •  Neighbor solicitation (ICMPv6 type 135) •  Neighbor advertisement (ICMPv6 type 136) •  Redirect (ICMPV6 type 137)

IPv4 IPv6 ARP Request Neighbor Solicitation

Broadcast Solicited Node Multicast

ARP Reply Neighbor Advertisement

Unicast Unicast

NDP

RA RS

NS NA Redirects

NUD DAD

IPv6


•  Router solicitations (RS) are sent by nodes at bootup

•  Host needs an RA to finish building it’s Address’s

RS

ICMP Type 133 IPv6 Source FE80::A IPv6 Destination FF02::2 Option 1 SRC Link Layer Address

RA

ICMP Type 134 IPv6 Source FE80::2

IPv6 Destination FE80::A Data Options, subnet prefix,

lifetime, autoconfig flag

RS RA

A


•  M-Flag – Stateful DHCPv6 to acquire IPv6 address

•  O-Flag – Stateless DHCPv6 in addition to SLAAC

•  H-Flag – Mobile IP home agent

•  Preference Bits – Low, Med, High

•  Router Lifetime – Must be >0 for Default

•  Options - Prefix Information, Length, Flags

•  L bit – Only way a host get a On Link Prefix

•  A bit – Set to 0 for DHCP to work properly

Type: 134 (RA) Code: 0 Checksum: 0xff78 [correct] Cur hop limit: 64 ∞ Flags: 0x84 1… …. = Managed (M flag) .0.. …. = Not other (O flag) ..0. …. = Not Home (H flag) …0 1… = Router pref: High Router lifetime: (s)1800 Reachable time: (ms) 3600000 Retrans timer: 1000 ICMPv6 Option 3 (Prefix Info) Prefix length: 64 ∞ Flags: 0x80 1… …. = On link (L Bit) .1.. …. = No Auto (A Bit) Prefix: 2001:0db8:4646:1234::/64

RA

type = 134 code = 0 checksum hop limit M|O|H|pref router lifetime

reachable time retransmit timer options (variable)


•  Enable DHCPv6 via the M flag •  Disable auto configuration via the A bit in option 3 •  Enable Router preference to high •  Enable DHCPv6 relay

interface fastEthernet 0/0 ipv6 address 2001:db8:1122:acc1::/64 eui-64 ipv6 nd managed-config-flag ipv6 nd prefix default no-autoconfig ipv6 nd router-preference high ipv6 dhcp relay destination 2001:db8:add:café::1


IPv4 IPv6

A record:

Function IPv4 IPv6

Hostname to

IP Address

A Record www.abc.test. A 192.168.30.1

AAAA Record (Quad A) www.abc.test AAAA 2001:db8:C18:1::2

IP Address To

Hostname

PTR Record 1.30.168.192.in-addr.arpa. PTR www.abc.test.

PTR Record 2.0.0.0.0.0.0.0.0.0.0.0.0.0.0.0.1.0.0.0.8.1.c.0.8.b.d.0.1.0.0.2.ip6.arpa PTR www.abc.test.

DNSServer!

2001:db8:1::1!

IPv4 IPv6

IPv4

IPv6

192.168.0.3!

www IN A 192.168.0.3 www IN AAAA 2001:db8:1::1

•  AAAA = easy, PTR = messy!•  Draft RFC 6106, RA Option 25!•  We need content..!


Node A can start using address A

B A C

•  ICMPv6 runs on top of IPv6, etype = 86DD, Layer 3.14 :’)

•  Probe neighbors to verify address uniqueness

ICMP Type 135 NS IPv6 Source UNSPEC = ::

IPv6 Dest. A Solicited Node Multicast FF02::1:FF00:A

Data FE80::A Query Anyone using A?

NS


•  For each Unicast and Anycast address configured there is a corresponding solicited-node multicast

•  Solicited-node multicast consists of FF02::1:FF/104 {lower 24 bits from IPv6 Unicast interface ID}

FF02 0000 0000 0000 0000 0001 FFBC FC0F

FE80 0000 0000 0000 1234 5678 9ABC FC0F

33 33 BC FC 0F FF Ethernet Multicast Uses last 32 bits


R1#sh ipv6 int e0 Ethernet0 is up, line protocol is up

IPv6 is enabled, link-local address is FE80::200:CFF:FE3A:8B18 Global unicast address(es):

2001:DB8:0:1234::1 subnet is 2001:DB8:0:1234::/64 Joined group address(es): FF02::1 FF02::2 FF02::1:FF00:1 FF02::1:FF3A:8B18 MTU is 1500 bytes ICMP error messages limited to one every 100 milliseconds ICMP redirects are enabled ND DAD is enabled, number of DAD attempts: 1 ND reachable time is 30000 milliseconds ND router advertisements are sent every 200 seconds *If EUI format is used then the 1rst solicited node mcast addr is used for both the LL & GU

Solicited-Node Multicast Address*


A! B!

ICMP Type 135 NS IPv6 Source FE80::A

IPv6 Destination B Solicited Node Multicast FF02::1:FF00:B

Target Address 2001:db8:1:46::B Code 0 (need link layer) Query What is B link layer

address?

ICMP Type 136 NA IPv6 Source FE80::B

IPv6 Destination FE80::A Target Type 2

Data Link Layer address of B *Flags R = Router

S = Response to Solicitation O = Override cache information

NS NA

•  ARP replacement, Map’s L3 to L2.

•  Node B will add node A to it’s neighbor cache during this process w/o sending NS

•  Multicast for resolution (new), Unicast for reachability (cache)


Neighbors are only considered “reachable” for 30-seconds. “Stale” indicates that, we MAY need to send a NS packet.


2001:db8:C18:2::/64

R1 R2 A B

Packet

IPv6 Source B IPv6 Dest. 2001:db8:c18:2::1 ULP variable

Redirect

ICMP Type 137 IPv6 Source Link Local (R2) IPv6 Dest. B Data Use Link Local (R1)

Redirect Packet

•  Cannot be used if destination is multicast

•  Hosts should not send redirects

•  Should be turned off on routed links


•  Prefix FF00::/8 8-bit 4-bit 4-bit 112-bit

1111 1111 0 R P T Scope Variable format

Flags

O Reserved

R = 0 R = 1

No embedded RP Embedded RP

P = 0 P = 1

Without Prefix Address based on Prefix

T = 0 T = 1

Well Known Address (IANA assigned) Temporary address (local assigned)

Scope 1 Node

2 Link

3 Subnet

4 Admin

5 Site

8 Organization

E Global


Address Scope Meaning FF01::1 Node-Local This Node

FF05::2 Site-Local All Routers

FF02::1 Link-Local All Nodes

FF02::2 Link-Local All Routers

FF02::5 Link-Local OSPFv3 Routers

FF02::6 Link-Local OSPFv3 DR Routers

FF02::9 Link-Local RIPng

•  FF02, is a permanent address and has link scope

•  Link Operations, Routing Protocols, Streaming Services


•  MLD uses LL source addresses

•  3 msg types: Query, Report, Done

•  MLD packets use “Router Alert” in HBH

•  MLDv1 = (*,G) shared, MLDv2 = (S,G) source

MLD snooping

MLD IGMP Message Type

ICMPv6 Type Function

MLDv1 (RFC2710) IGMPv2 (RFC 2236) Listener Query

Listener Report

Listener Done

130

131

132

Used to find out if there are any multicast listeners

Response to a query, joins a group

Sent by node to report it has stopped listening

MLDv2 (RFC 3810) IGMPv3 (RFC 3376) Listener Query

Listener Report

130

143

Used to find out if there are any multicast listeners

Enhanced reporting, multiple groups and sources


•  Hosts send MLD report to alert router they wish to join a multicast group

•  Router then joins the tree to the source or RP

MLD Report (A)

ICMP Type 131

IPv6 Source fe80::209:5bff:fe08:a674

IPv6 Destination FF38::276

Hop Limit 1

Group Address ff38::276

Hop-by-Hop Header

Router Alert Yes

MLD Report

A MLD Report

B I wish to receive

ff38::276 I wish to receive

ff38::276

MLD Report (B)

ICMP Type 131

IPv6 Source fe80::250:8bff:fE55:78de


Hop Limit 1


Hop-by-Hop Header

Router Alert Yes

(S, G)

Source for multicast ff38::276

fe80::209:5bff:fe08:a674 fe80::250:8bff:fE55:78de fe80::207:85ff:fe80:692


MLD Done (A)

ICMP Type 132


IPv6 Destination FF02::2 (All routers)

Hop Limit 1


Hop-by-Hop Header

Router Alert Yes

MLD Done (A)

A

fe80::209:5bff:fe08:a674 MLD Report (B)

B

fe80::250:8bff:fE55:78de

I wish to leave ff38::276

I am watching ff38::276

MLD Query (C)

ICMP Type 130

IPv6 Source fe80::207:85ff:fe80:692


Hop Limit 1

Hop-by-Hop Header

Router Alert Yes

Query (C

)

fe80::207:85ff:fe80:692

C MLD Report (B)

ICMP Type 131

IPv6 Source fe80::250:8bff:fE55:78de


Hop Limit 1


Hop-by-Hop Header

Router Alert Yes


MLD Report (A)

ICMP Type 143



Hop Limit 1

# of Records Include/exclude

Group Address FF38::4000:BA11

Hop-by-Hop Header

Router Alert Yes

MLD Report

A I wish to receive FF38:4000:BA11

(S, G)

Source for multicast FF38::4000:BA11

fe80::209:5bff:fe08:a674


•  General Query FF02::1 Group list empty, who’s listening?

•  Group Specific Query FF38::4000:BA11 Anyone still interested in this stream?

•  Group & Source Specific Query 2001:DB8:CAFÉ::1, FF38::4000:BA11

•  Filter Mode, Change Record

•  Multiple routers on link Lowest address value assumes Querier role

A Q

uery

Source for multicast FF38::4000:BA11


•  Loopback 0:0:0:0:0:0:0:1=> ::1

•  Unspecified address 0:0:0:0:0:0:0:0=> 0::0 => :: => ::/128

•  Documentation Prefix 2001:0DB8::/32

•  Discard Prefix 0100::/64

•  6to4 Automatic Tunneling 2002::/16

•  Default Route ::/0


•  IPv4 Compatible 0:0:0:0:0:0.A.B.C.D/96 0:0:0:0:0:0.192.168.30.1 ::C0A8:1E01 Used by IPv6 aware devices, now deprecated

•  IPv4 Mapped 0:0:0:0:0:FFFF.A.B.C.D/96 0:0:0:0:0:FFFF.192.168.30.1 ::FFFF:C0A8:1E01

Used in automatic tunneling by device with no IPv6 knowledge

IPv4

IPv6 Internet

IPv6 Network


Windows 7, Mac OSX use pseudo random by default. iPad & iPhone generate a new temporary address per association

C:\Documents and Settings\>netsh netsh>interface ipv6 netsh interface ipv6>show address Querying active state... Interface 5: Local Area Connection Addr Type DAD State Valid Life Pref. Life Address --------- ---------- ------------ ------------ ----------------------------- Public Preferred 29d23h58m25s 6d23h58m25s 2001:0db8:2301:1:202:8a49:41ad:a136 Temporary Preferred 6d21h48m47s 21h46m 2001:0db8:2301:1:bd86:eac2:f5f1:39c1 Link Preferred infinite infinite fe80::202:8a49:41ad:a136 netsh interface ipv6>show route Querying active state... Publish Type Met Prefix Idx Gateway/Interface Name ------- -------- ---- ------------------------ --- --------------------- no Autoconf 8 2001:0db8:2301:1::/64 5 Local Area Connection no Autoconf 256 ::/0 5 fe80::20d:bdff:fe87:f6f9



Distribution Layer

Access Layer

Core Layer

Aggregation Layer (DC)

Access Layer (DC)

IPv6/IPv4 Dual-stack

Server

IPv6/IPv4 Dual-stack Hosts

Data Center Block

Access Block

• Leverages existing IPv4 infrastructure • Allows “slower” roll into IPv6 deployment • Poor scalability and overall performance, no Multicast support • Tunneling everywhere, “flattens” the network you have built


ISATAP IPv6 Service Block

DA

Data Center Block

WAN/ISP Block

Access Layer

Dist. Layer

Core Layer

IPv4-only Campus Block

Server Internet

• Provides tighter control of where IPv6 is deployed • Allows for reduced time to deliver IPv6 services • Cost of SB equipment and it’s reuse in the network • Eventually hits scalability and overall performance, no Multicast support


Distribution Layer

Access Layer

Core Layer

Aggregation Layer (DC)

Access Layer (DC)

IPv6/IPv4 Dual-stack

Server

IPv6/IPv4 Dual-stack Hosts

Data Center Block

Access Block

49

• Preferred Method, Versatile, Scalable and Highest Performance • No Dependency on IPv4, runs in parallel on dedicated HW • No tunneling, NAT or other performance degrading technologies • Does require IPv6 support on all devices


• Should we use both on the same link at Layer 3? • Separate links, possibly to collect protocol specific statistics • Routing protocols OSPFv3, EIGRP combined or separate? • Fate sharing between the data and control planes per protocol

OSPFv3

EIGRP

Internet

2001:db8:1:1::/64 198.51.100.0/24

IPv4 & IPv6

IPv4 & IPv6

2001:db8:6:6::/64 192.168.4.0/24


• Topology hiding, Interfaces cannot be seen by off link devices • Reduces routing table prefix count, Less configuration • Need to use ULA or GUA for management and troubleshooting • What about DNS?, WAN connections and more

WAN/MAN

Internet FE80::/64

FE80::/64

ULA/GUA

FE80::/64

ULA/GUA

ULA/GUA

ULA/GUA

ULA/GUA


Corporate Backbone Branch 2

ULA Space FD9C:58ED:7D73::/48 Global – 2001:DB8:CAFE::/48

Internet

FD9C:58ED:7D73:3000::/64 2001:DB8:CAFE:3000::/64

FD9C:58ED:7D73::2::/64 2001:DB8:CAFE:2::/64

Global

2001:DB8:CAFE::/48

• Both ULA and Global are used except for Internal only hosts • Semi random generator requires non sequential /48’s, avoid M&A challenges • Need to use Global for troubleshooting beyond the internal network • Multiple policies to maintain (ACL, QoS, Routing, etc..)


•  Today, NAT44 & RFC1918 •  All PA or all PI and peering in multiple regions

PI from one region and run it everywhere? ISP in one region reject PI block from another? What about translation?

•  IETF does NOT recommend the use of NAT66 w/IPv6

•  NAT ≠ Firewall – RFC 4864 (Local Network Protection)

•  NAT ≠ Firewall – RFC 7021 (Impact of CGN on Applications)

Firewall+NAT Internet

Some enterprises are getting a prefix per RIR and only deploying one.

Building backup plans with the others


•  /64 everywhere a host

•  /127 Point to Point Should not use all 0’s or 1’s in the host portion Nodes 1&2 are not in the same subnet

•  /128 Loopback out of a single /64 as a design option

•  /64, /64, /64

Pt 2 Pt /127

WAN

Core /64 or /127

Servers /64

Hosts /64

Loopback /128


•  Scope, Preferred over Deprecated, Native over Transitional, Temporary over Public •  Must support application override API, Choice of v6 over v4 is application dependent •  Give IPv6 300ms Head Start Pv6/IPv4 Lookup & Connect Retrieve and Display

Application Layer

TCP/UDP

IPv6 IPv4

Network Interface Card

NCSI – Network Connection Status Indicator

Temporary Preferred 2001:0db8:2301:1:bd86:eac2:f5f1:39c1 Public Preferred 2001:0db8:2301:1:202:8a34:bead:a136 Link Preferred fe80::202:8a34:bead:a136

RFC 6555



•  Many similarities with HSRP for IPv4 from design perspective (odd/even prefix) •  Changes occur in Neighbor Advertisement, Router Advertisement, and ICMPv6 redirects •  Virtual MAC derived from HSRP group number and virtual IPv6 link-local address •  Layer 2 - 0005.73A0.0000-0F0F à 3333.0000.0066 •  Layer 3 – FE80::/64 à FF02::66 •  Layer 4 – UDP port 2029 (HSRPv6)

interface FastEthernet0/1

ipv6 address 2001:DB8:66:67::2/64

standby version 2

standby 2 ipv6 autoconfig

standby 2 timers msec 250 msec 800

standby 2 preempt

standby 2 preempt delay minimum 180

standby 2 authentication cisco

HSRP Standby

HSRP Active

Unix Host with GW of VIP unixhost# route -A inet6 | grep ::/0 | grep eth2 ::/0 fe80::5:73ff:fea0:1


•  Provides weighted or ~= load balancing across resources •  Modification to NA, default GW is announced via RA using vmac •  AVG assigns Vmac’s and responds to NDP, directing hosts to the AVF’s •  Layer 2 – 0007.B4xx.xx08 à 3333.0000.0066 •  Layer 3 – FE80::/64 à FF02::0100:5E00:66 •  Layer 4 – UDP port 3222 (GLBPv6)

interface fastethernet0/0

ipv6 address 2001:db8::/64 eui-64

glbp 8 ipv6 2001:db8::D38:C677:2925:8

glbp 8 priority 110

glbp 8 preempt

glbp 8 load-balancing weighted

glbp 8 weighting 110 lower 95 upper 105

glbp 8 authentication md5 key-string 7 cisco

58

GLBP 8 AVF

GLBP 8 AVG AVF


•  IPv4 syntax has used “ip” following match/set statements Example: match ip dscp, set ip dscp

•  New match criteria match dscp match precedence

•  New set criteria set dscp set precedence

•  Modification to support IPv6 and IPv4

Data Voice

Video Internet


•  Apple (Bonjour) has a light weight approach, adopted quicker

•  FF02::FB – Multicast DNS – mDNS

•  Microsoft (Rally) has a more robust, heavier implementation, has moved slower

•  FF02::C – Simple Service Discovery Protocol – SSDP, UPnP

•  FF02::1:3 – Link Local Multicast Name Resolution – LLMNR (File Sharing enabled)

Personal Computer Operating Systems •  Windows •  Mac OS X •  Linux

Appliances & Networking •  Printers •  Access Points •  Switches •  Routers

AV Equipment •  Speakers •  Cameras •  Displays •  AV Receivers


IPv6 Snooping

IPv6 FHS RA

Guard DHCPv6 Guard

Source/Prefix Guard

Destination Guard

Protection: •  Rouge or

malicious RA •  MiM attacks

Protection: •  Invalid DHCP

Offers •  DoS attacks •  MiM attacks

Protection: •  Invalid source

address •  Invalid prefix •  Source address

spoofing

Protection: •  DoS attacks •  Scanning •  Invalid

destination address

RA Throttler

ND Multicast Suppress

Reduces: •  Control traffic

necessary for proper link operations to improve performance

Core Features Advance Features Scalability & Performance

Facilitates: •  Scale

converting multicast traffic to unicast


•  Admin/Intern sends RA’s with false prefix •  Enthusiast who has a tunnel broker account •  The most frequent threat by non-malicious user

B

Src = C link-local address Dst = All-nodes Options = prefix BAD

RA

A C


•  Attacker hacks any victim's DAD attempts

•  Victim will need manual intervention to configure IP address

Src = UNSPEC Dst = Solicited-node multicast A Data = A Query = Does anybody use A?

Src = any C’s IF address Dst = A Option = link-layer address of C

A B

NS

NA

C


•  Flooding RA’s overwhelms the system, OSX, MSFT, ipad/phone, Android

B RA, prefix BAD1

A 2 3 5

RA, prefix BAD2 RA, prefix BAD3 RA, prefix BAD4 RA, prefix BAD5 RA, prefix BAD6

C

Update: MSFT Addresses Vulnerability in IPv6 Could Allow Denial of Service (2904659) Published: Tuesday, February 11, 2014


•  Attacker spoofs Router Advertisement with false on-link prefix •  MITM, Splash Screen, Capture

B

Src = B’s link-local address Dst = All-nodes Options = prefix BAD

RA

A C


•  Port ACL blocks all ICMPv6 RA from hosts interface FastEthernet0/2

ipv6 traffic-filter ACCESS_PORT in

deny icmp any any router-advertisement

•  RA-guard lite (12.2(33)SXI4 & 12.2(54)SG ): also dropping all RA received on this port


ipv6 nd raguard

access-group mode prefer port

•  RA-guard (12.2(50)SY, 15.0(2)SE)

ipv6 nd raguard policy HOST device-role host

ipv6 nd raguard policy ROUTER device-role router

ipv6 nd raguard attach-policy HOST vlan 100


ipv6 nd raguard attach-policy ROUTER

HOST Device-role

RA

RA

RA

RA

RA

ROUTER Device-role


Prevent Rogue DHCP responses from misleading the client

Before DHCP Guard After DHCP Guard

Host First Hop Switch Host First Hop Switch

DHCP Server DHCP Server

I am a DHCP Server

DHCP Req. DHCP Req.

I am a DHCP Server


•  Deep control packet Inspection •  Address Glean (ND , DHCP, data) •  Address watch •  Binding Guard

Instrumental link-operation security feature that analyzes control/data switch traffic, detect IP address, and store/update them in Binding Table to ensure rogue users cannot spoof or steal addresses.

Intf IPv6 MAC VLAN State

g1/0/10 ::000A 001A 110 Active

g1/0/11 ::001C 001C 110 Stale

g1/0/16 ::001E 001E 200 Verifying

IPv6 Binding Table (RFC6620)

IPv6 Source Guard

IPv6 Destination

Guard Device Tracking


Host A First Hop Switch Host A First Hop Switch

Mitigates Address High Jacking, Ensures Proper Prefix


g1/0/10 ::000A 001A 110 Active

g1/0/11 ::001C 001C 110 Stale

g1/0/16 ::001E 001E 200 Verifying

g1/0/21 ::0021 0021 200 Active


g1/0/10 ::000A 001A 110 Active

g1/0/11 ::001C 001C 110 Stale

g1/0/16 ::001E 001E 200 Verifying

g1/0/21 ::0021 0021 200 Active

NA NA

NA

NA

~Host A ~Host A


•  Mitigate prefix-scanning attacks and Protect ND cache •  Drops packets for destinations without a binding entry


g1/0/10 ::0001 001A 110 Active

g1/0/11 ::001C 001C 110 Stale

g1/0/16 ::001E 001E 200 Verifying

Forward packet

Lookup Table

found No

Yes

NS 2001:db8::1

Ping 2001:db8::1

Ping 2001:db8::4 Ping 2001:db8::3

Ping 2001:db8::2


•  Scaling the 802.11 multicast reliability issues •  NDP process is multicast “chatty”, consumes airtime •  Controller rate limits the period RA’s, while allowing RS to flow •  Caching allows the Controller to “proxy” the NA, based on gleaning

(NS)

00:24:56:75:44:33 2001:db8:0:20::2 00:24:56:11:93:28 2001:db8:0:20::4

(Unicast NA)

(NS) (Unicast NA)

2

4 Periodic (RA’s)


Anchor

Foreign

Mobility Tunnel

Unicast RA

Mcast RA

Roaming Client

•  Roaming client must be able to receive the original router advertisement •  Controllers must be part of the same mobility group domain •  The anchor controller sends the RA to the foreign in the mobility tunnel •  AP convert’s multicast RA to an L2 unicast (MC2UC)


ipv6 general-prefix foo-core 2001:0db8:4646:6000::/52 ipv6 general-prefix foo-acc 2001:0db8:4646:6acc::/56 ipv6 unicast-routing ! interface GigabitEthernet1/0/1 description To 6k-core-right ipv6 address foo-core ::3:0:0:0:d63/64 ipv6 eigrp 10 ipv6 hello-interval eigrp 10 1 ipv6 hold-time eigrp 10 3 ipv6 authentication mode eigrp 10 md5 ipv6 authentication key-chain eigrp 10 eigrp ipv6 summary-address eigrp 10 2001:0db8:4646:6000::/52 ! interface GigabitEthernet1/0/2 description To 6k-core-left ipv6 address foo-core ::C:0:0:0:d63/64 ipv6 eigrp 10 ipv6 summary-address eigrp 10 2001:0db8:4646:6000::/52

interface Vlan4 description Data VLAN for Access ipv6 address foo-acc ::4:d63/64 ipv6 eigrp 10 ! interface Vlan6 description Data VLAN for Access ipv6 address foo-acc ::6:d63/64 ipv6 eigrp 10 ! ipv6 router eigrp 10 no shutdown router-id 10.122.10.10 passive-interface Vlan4 passive-interface Vlan6 passive-interface Loopback0


•  RIPng – UDP 521, 15 hops •  FE80::/64 Source à FF02::9 Destination

•  Distance Vector, Hop Count (1-15)

•  Split Horizon, Poison Reverse

•  Beuller, Bueller, anyone?

Ipv6 unicast-routing ! Interface loopback 0 Ipv6 address 2001:db8:1000::1/128 Ipv6 rip CISCO enable ! Interface ethernet 0/0 Ipv6 address 2001:db8:5000:31::1/64 Ipv6 rip CISCO enable ! Ipv6 router rip CISCO


Ipv6 unicast-routing ! Interface ethernet 0/0 Ipv6 address 2001:db8:5000:31::1/64 Ipv6 router isis CISCO Isis circuit-type level-1 Isis ipv6 metric 10000 ! Router isis CISCO Net 49.0001.2222.2222.222.00 Metric style wide ! Address-family ipv6 Multi-topology

•  IS-IS – (RFC5308) CLNS •  IPv4 & IPv6

•  Link State •  2 New TLV’s Added

•  Topology Supoprt •  Single Topology •  Multi Topology •  Multi Instance


Ipv6 unicast-routing ! Interface loopback0 Ipv6 address 2001:db8:1000::1/128 Ipv6 eigrp 11 ! Interface ethernet 0/0 Ipv6 address 2001:db8:5000:31::1/64 Ipv6 eigrp 11 ! Ipv6 router eigrp 11 Passive-interface loopback0 Eigrp router-id 10.10.10.10

•  EIGRP – IP 88

•  FE80::/64 Source à FF02::A Destination

•  2 New TLV’s – internal-type & external-type

•  No Split Horizon, Auto Summary Disabled

•  Stub reduces topology & queries •  EIGRP can perform better in large scale

hub and spoke environments


Ipv6 unicast-routing ! Interface loopback0 Ipv6 address 2001:db8:1000::1/128 Ipv6 ospf 3 area 0 instance 13 ! Interface ethernet 0/0 Ipv6 address 2001:db8:5000:31::1/64 Ipv6 ospf 3 area 0 instance 13 ! Ipv6 router ospf 3 router-id 10.10.10.10 passive-interface loopback0

•  OSPFv3 – IP 89 •  FE80::/64 Source à FF02::5, FF02::6 (DR’s) •  Link-LSA (8) – Local Scope, NH •  Intra-Area-LSA (9) – Routers Prefix’s •  Use Inter-Area-Prefix (3) – Between ABR’s

•  Can converge quickly to a point of scale, initial database build and discovery takes some time

•  Link state protocols perform better in full mesh environments, if tuned correctly


•  RFC 4798 ‒ Utilizes Existing IPv4 Core ‒ Dual Stack on the PE

•  MP-BGP Next Hop ::ffff:192.0.2.210 •  LDP Next Hop 192.0.2.210 •  Control Plane uses IPv4, for now

2001:db8:babe::/48 2001:db8:d00d::/48

R1 R4

IPv4 core, LDP, IGP, TE, etc.

IPv6 MP-BGP LDPv4

router bgp 192 no bgp default ipv4-unicast bgp log-neighbor-changes neighbor 192.0.2.3 remote-as 192 neighbor 129.0.2.3 update-source Loopback0 ! address-family ipv6 redistribute connected redistribute static no synchronization neighbor 192.0.2.3 activate neighbor 192.0.2.3 send-label exit-address-family ! ipv6 route 2001:db8:babe::/48 2001:db8:1:1::2


•  RFC 4659 ‒ Utilizes Address Family (AF) in VRF Context ‒ Allows for VPN Functionality

•  Subsequent Address Family Identifiers (SAFI) ‒ MP-BGP Address-family SAFI = 2 (IPv6) ‒ VRF Address-family SAFI = 128 (VPN)

2001:db8:café:1::/64

2001:db8:babe:1::/64

2001:db8:d00d:1::/64

2001db8:café:4::/64

2001:db8:babe:4::/64

2001:db8:dood:4::/64

R1

R4

IPv4 core, LDP, IGP, TE, etc.

vrf definition 6vpe rd 192:1 route-target export 192:1 route-target import 192:3 ! address-family ipv6 exit-address-family ! router bgp 192 address-family vpnv6 neighbor 192.0.2.3 activate neighbor 192.0.2.3 send-community both exit-address-family ! address-family ipv6 vrf 6vpe redistribute connected redistribute static no synchronization exit-address-family



•  IPv4 Only Data Center – IPv6 Translation on the Front End

•  Dual Stack – Both IPv4 & IPv6 Into the Data Center

•  IPv6 Only Data Center – IPv4 Translation on the Front End

•  What is the Cost of Each Stage?


•  IPv4

•  Load Balancer inline

•  No translation in this design

•  Services are Firewalled

Internet Firewall Edge Router Load Balancer Switch Web, Email, Etc.

IPv4


•  Dual Stack Front End

•  Translation via NAT/Proxy/SLB

•  Easy to Turn Up

•  Hard to Move Forward

•  False Sense of Accomplishment

Firewall Edge Router Load Balancer Switch Web, Email, Etc.

NAT/Proxy/SLB

IPv4/IPv6 IPv4

Internet


•  IPv4 & IPv6 Addressing on All Devices

•  Incremental Operational Cost (~20%)

•  Double Everything (ACL’s, SLA’s, etc.)

•  Two Data Planes, Two Control Planes

•  Recommended Approach


IPv4/IPv6

Internet


•  Dual Stack Front End

•  Translation via NAT/Proxy/SLB

•  Forces Developers to use IPv6

•  Reduces Operational Costs

•  Eliminates Complexity within the DC

Load Balancer Switch Web, Email, Etc.

NAT/Proxy/SLB

IPv6 IPv4/IPv6


•  IPv4 Sailed Toward Sunset

•  Not Likely for a Very Long Time

•  Someday, Maybe Someday


IPv6

Internet


•  Large DCs with very dense hosts populations can cause severe performance problems on the control plane of switches due to IPv4 and IPv6 ‘control’ traffic

•  One size will not fit all, tuning will require experimentation

87

• NUD Reachable Time: ipv6 nd reachable-time time-in-milliseconds

• NUD Retry Interval: ipv6 nd nud retry base interval-in-milliseconds max-attempts

• Scavenge and Refresh Timer: ipv6 nd cache expire time-in-seconds

• Unsolicited NA Glean: ipv6 nd na glean

• Glean rate limiter: mls rate-limit unicast cef glean <pps> <burst>


•  Same configuration requirements and operation as with IPv4

•  Configure VRRP address to be the same as physical interface of “primary”


• Tunnel Protocol for Fiber Channel over an IP infrastructure • RFC 4404 – Entity Address Size IPv4 (4) or IPv6 (16) • MDS 9x00 Series

– out-of-order delivery, jumbo frames, traffic shaping, TCP optimization


• Replication Services, Disaster Recovery • Shared Assets and Resources • Overlay Transport Virtualization (OTV) – L2 Technologies Encapsulated in L3 – IPv6 Within OTV over IPv4 – Disable Optimized Multicast Forwarding (OMF) in IGMP snooping on OTV edge devices for IPv6 Solicited Node Multicast traffic to flow.

OTV Global Cloud


•  Hosts are ready Windows enabled by default, disabling it = no more support from Microsoft Mac OS X, iOS, Android, Linux, */BSD: enabled by default

•  File & Print No WINS or NetBios over IPv6 SMB on TCP 445

•  SQL Server IPv6 preferred Watch for v4 socket calls

•  Server 2008/R2 Needs Unified Access Server

•  Server 2012 Includes UAS, NAT64/DNS64


• Inconsistent API’s use of IPv6 Addresses •  Some functions expect a URL (must enclose with brackets for IPv6) •  Some functions expect just an IP (no bracket)

• Know whether your app displays or accept an IPv6 address • 198.51.100.44:8080 à [2001:db8:café:64::26]:8080

• Home grown App’s may only support IPv4 • Pressure vendors to move to protocol agnostic framework • RFC 3493 – Open Socket Call, 64 bit structure align to HW • RFC 3542 – Raw Socket, ping, Traceroute, r commands


• May Need Radical Change – Disabling IPv4 Forces IPv6 Development

• Open Stack – Havana – Limited IPv6 Support, WP – Ice House – Neutron L3 Extension

• Application Centric Infrastructure (ACI) – Automates Network Resource Provisioning – On Demand Scale of Applications


•  Server supports IPv4 and IPv6

•  Internal & external

•  Server supports IPv4 & IPv6

•  DNSSEC

DHCP DNS

•  Integrated DNS and DHCP

•  Configuration and reporting

IPAM

•  SNMPv3 over IPv6 and managing IPv6 MIB’s •  Protocol Version Independent (PVI) manage the same OID’s (RFC’s 4292, 4293) •  NetFlow, Deep Packet Inspection, IPSLA, all work with IPv6 •  IPv6 support for Wireshark, Packet analysis, MRTG, Netflow collectors, etc..


•  Anycast Address for Client Access to DHCP/DNS •  Uses the same address in multiple locations •  Simple, Scalable and Reliable Solution •  Global Unicast Address (GUA) for Service Uptime •  DNS server injects /128 via OSPF DDI1

2001:db8:aa::21

2001:db8:aa::21

2001:db8:aa:: Cost 10

I pick DNS1 closest metric

2001:db8:aa:: Cost 30

2001:db8:aa:: Cost 20

DDI2 2001:db8:aa::21

DDI3 2001:db8:aa::21

Command &

Control

GUA

GUA



Application Support

Server Load Balancer

IPv6

IPv4

IPv6 Internet

Stateful NAT64

Client Visibility

IPv4

IPv6

IPv4 Internet

SW = Poor Performance

Proxy

IPv6

IPv4

IPv6 Internet


•  Translation Algorithms RFC 6052 (Implementation Details)

•  Framework for Translation RFC 6144 (Implementation Scenarios)

•  Stateless NAT64 RFC 6145 (IP/ICMP Translation Algorithm) Maps the Entire IPv4 Internet into IPv6 Prefix

•  Stateful NAT64 RFC 6146 (State Table for IPv4/IPv6 Translation) Used mainly where IPv6-only clients need to access IPv4 servers

•  DNS64 RFC 6147 (IPv6 Client to IPv4 Server)

Version IHL Type of Service Total Length

Identification Flags Fragment Offset

Time to Live Protocol Header Checksum

Source Address

Destination Address

Version Traffic Class Flow Label

Payload Length Next Header Hop Limit

Source Address

Destination Address


•  RFC 6144 8 Total Scenarios (4, 7, 8 are NA) 1, 2, 3 Involve Internet Connectivity 5 & 6 Are Focused on Intranet Connectivity

•  Stateless Translation Algorithmic Mapping Address Information & Translator Configuration Initiation from IPv4 or IPv6

•  Stateful Translation Uses a State Table for Translation Based on L3/L4 Tuples Generally Initiation is from IPv6

Scenarios Defined Version IHL Type of Service Total Length



Source Address

Destination Address



Source Address

Destination Address

IPv4 Internet

IPv4 Internet

IPv4 Network

IPv6 Network

IPv6 Network

IPv6 Internet

IPv6 Network

IPv4 Network

IPv4 Network

IPv6 Network

1

2

3

5

6


•  RFC 6145 1:1 Address Mapping

•  IP Header Fields Copy ToS to/from Traffic Class Id, Flags & Offset to/from EH 44 TTL to/from Hop Limit and Decrement Protocol to/from Next Header Checksum Computed IPv6 to IPv4

•  ICMP Header Fields Type Translated (IPv4 8, 0 to IPv6 128, 129) Pseudo Checksum for IPv6 (UDP as well)

Ideal for IPv6 Only Data Center

100




Source Address

Destination Address



Source Address

Destination Address

IPv4 Internet

IPv6 Network

IPv6 Network

IPv6 Network

IPv4 Network

IPv4 Network

IPv6 Network

1

2

5

6

IPv4 Internet


•  RFC 6146 Overload Address Mapping

•  TCP/UDP/ICMP Headers Form L4 Portion of State Tuple Pseudo Checksum for IPv6 Portion No Multicast, IPSec, etc.

•  MUST use DNS64 RFC 6147 Synthesis DNS Records AAAA to A

•  ALG’s May be Required

•  MTU Issues Possible

Ideal for IPv6 Only Networks Accessing IPv4

101




Source Address

Destination Address



Source Address

Destination Address

IPv4 Internet

IPv4 Internet

IPv4 Network

IPv6 Network

IPv6 Network

IPv6 Internet

IPv6 Network

IPv4 Network

IPv4 Network

IPv6 Network

1

2*

3

5

6*

*Use Static IPv6 to IPv4 Mappings


Step 1à IPv6 PC queries AAAA Record for v4 Server

2001:db8:122:344::6 DNS Server

192.168.90.101

192.0.2.0/24 2001:db8:122:344::/64

DNS64

DNS46

IPv6 PC

.1 ::2

ßStep 5 Translates it to a AAAA Record

AAAA Record A Record

AAAA Record

A Record

Network-Specific Prefix 3001::/96

Step 3à Translator Sends A Record for v4Server ßStep 2 DNS responds “empty” AAAA Record

ßStep 4 DNS Server responds A Record for IPv4Server


ßSource IPv6 3001::3001::c000:221 Dest. IPv6 2001:db8:122:344::6

ßSource IPv4 192.0.2.33 Dest. IPv4 192.0.2.1

à Source IPv6 2001:db8:122:344::6 Dest. IPv6 3001::3001::c000:221

Network-Specific Prefix 3001::/96

2001:db8:122:344::6 IPv4 Server 192.0.2.33

2001:db8:122:344::/64

Dynamic NAT64

Static NAT46

IPv6 PC

.1 ::2 192.0.2.0/24

àSource IPv4 192.0.2.1 Dest. IPv4 192.0.2.33


Citrix NetScaler

•  Virtual IP (VIP), SNAT Pool •  Publish Appropriate AAAA Record •  IPv6 to IPv4, Similar to NAT64 •  Translation & SLB are done on same platform

•  OS/App dictate design parameters •  Rapid Time to Deploy

ISP-A

Servers WWW

ISP-B

UCS Servers

Dual Stack

IPv4 Only


•  Web Server Logging for Geo Location, Analytics, Security, etc..

•  Source IP of client requests will be logged as the SNAT or other NAT’d address

•  Packet may go through multiple proxies X-Forwarded-For: client, proxy1, proxy2

GET / HTTP/1.1 Host: www.foo.org User-Agent: Mozilla Firefox/3.0.3 Accept: text/html,application/xhtml+xml,application/xml Accept-Language: en-us,en Keep-Alive: 300 x-forward-for: 2001:db8:ea5e:1:49fa:b11a:aaf8:91a5 Connection: keep-alive Servers

WWW

Global IPv6 Address ---Translation--- Source NAT Pool



Single Link Single ISP

Enterprise

ISP 1

Default Route

Dual Links Single ISP

ISP 1 POP1

POP2

Enterprise

Multi-Homed Multi-Prefix

Enterprise

ISP2

USA

ISP4

BGP

ISP3

ISP 1

Europe


•  Do you support dual stack peering? •  Do you have a separate (SLA) for IPv6? •  Do you support BGP peering over IPv6? •  Do you have a FULL IPV6 route table? •  What is the maximum prefix length?

•  What about DNS…

Hosted Cloud Service §  Maximum prefix length offered by the cloud provider? §  Access to provisioning and billing portal over IPv6? §  Global IPv6 addressing for VM’s in your environment?

ISP-A ISP-B

Routing

Switching

Services


•  Challenges Arise ‒ Upstream Address Filters ‒ Asymmetric Routing ‒ Default GW & NH Selection

§  Small to Medium Enterprise ‒ Provider Allocated ‒ NPTv6, LISP

§  Medium to Large Enterprise ‒ Provider Independent ‒ BGP

ISP-A ISP-B


ISP-A ISP-B •  Small to Medium Enterprise ‒ Provider Allocated ‒ Not Topology Hiding

•  Swaps Left Most Bits of Address ‒ Just the Prefix ‒ Length Must be Equal

•  Public on Outside (2001:db8::/48) •  Private on Inside (FD00::/8) •  No Protocol “fixups”, Unless ALG’s are Supported

FD07:18:403e::/48

2001:bd8:11::/48 2001:bd8:55::/48


•  Small to Medium Enterprise •  Tunneling the PA IPv6 over LISP ‒ Provider Allocated /48 ‒ Hosted by PxTR Provider

•  Avoids Multi Prefix PA Issues •  Possibly an ISP that is IPv4 Only •  SHIM6, HIP, ILNP etc. ‒ OS Mods, Code Change

Dual Stack Internet

MR/MS PxTR MR/MS PxTR

Client 172.16.99.100 2001:db8:ea5e:1::/64

2001:db8:cafe::/48

xTRs

192.168.1.x/30

2001:db8:cafe:103::/64

2001:db8:cafe::/48


•  Medium to Large Enterprise

•  Use a /127 on pt-2-pt, /64 on multipoint

•  MD5 shared secret’s, IPSec could be used

•  Controlling TTL, accepting >254 only (allow -1)

•  Path, prefix size limits and filtering

ISP-A

:2

:3

:1

:3

2001:db8:cafe:102::/127

2001:db8:cafe:103::/64

ISP-B

::6 ::7

:4

:5

:2

:4

router&bgp&200&&&

bgp&router,id&2.2.2.2&&

neighbor&2001:DB8:cafe:102::2&remote,as&2112&&

neighbor&2001:DB8:cafe:102::2&ttl,security&hops&1&&&

neighbor&2001:DB8:cafe:102::2&password&cisco123&


•  Bogon filtering (data plane & BGP route-map): http://www.cymru.com/Bogons/ipv6.txt

•  Anti-spoofing (RFC2827, BCP38), Multi homed filtering (RFC3704, BCP 84)

•  uRPF – Unicast Reverse Path Forwarding

Enterprise Internet

B2B


•  Address Range Source of 2000::/3 at minimum vs. “any”, permit assigned space

•  ICMPv6 Error types thru, NDP to, RFC 4890

•  Extension Headers Allow Fragmentation, others as needed. Block HBH & RH type 0

•  IPv6 ACL’s IPv6 traffic-filter – to apply ACL to an interface



•  Address Spoofing (-> uRPF, ACLs) [IOS, ASA]

•  Neighbor Discovery Attacks (NS,NA,RS,RA,REDIR..) (-> First Hop Security, RA Guard, ND Inspection, SeND, ACL, IPv6 Inspects) [Catalyst, ASA]

•  Routing Header (RH0) source routing like attacks (-> no ipv6 source route (before 12.4.(15)T, blocked on ASA by default ) [IOS, ASA]

•  Extention Header (e.g. Fragmentation) Games (-> ACLs e.g. deny ip any any undetermined-transport) [IOS, ASA, Catalyst]

•  DHCP Attacks (-> DHCP Authentication, PACL) [IOS, ASA, Catalyst]

•  Transition Technologies (6to4, Teredo, ISATAP, etc) Attacks (-> ACL, disabled on device, enable IPv6 ) [IOS, ASA, IPS]

•  Smurf Attack (-> uRPF) [IOS, ASA]

•  Routing Protocol Attacks (-> Authentication) [IOS,ASA]

•  … and more. To be continued… have fun defending them. J


© 2012 Cisco and/or its affiliates. All rights reserved. 118 A Phased-Plan Approach for Successful IPv6 Adoption

IPv6 Assessment Service •  Determine how your network needs to change to support your IPv6 strategy

IPv6 Discovery Service •  Guidance in the early stages of considering a transition to IPv6

IPv6 Planning and Design Service •  Designs, transition strategy, and support to enable a smooth migration

IPv6 Implementation Service •  Validation testing and implementation consulting services

Network Optimization Service •  Absorb, manage, and scale IPv6 in your environment


•  Gain Operational Experience now

•  Security enforcement is possible

•  Control IPv6 traffic as you would IPv4

•  “Poke” your Provider’s

•  IPv6 is here now are you?

119


Technology

IPv6-workshop-tm-gdv-f