communication systems 7 th lecture

1 | 56

Communication Systems7th lecture

Chair of Communication SystemsDepartment of Applied Sciences

University of Freiburg2006

2 | 56

Communication SystemsLast lecture and practical course

● Thursday (25.05) is a holiday, no lecture, no practical.● Next Tuesday (30.05) is practical course on IPv6 in RZ -113.

● Introduction to the DNS● DNS Components● DNS Structure and Hierarchy● DNS in Context● DNS as an Internet service● ENUM as a DNS extension for Internet telephony

3 | 56

Communication SystemsPlan for this lecture

● Problems and success of IP v4● Introduction to future IP● IP v6 address● IP v6 header and extension headers● IP v6 fragmentation● IP v4 to IPv6 transition● DNS in IP v6

4 | 56

Communication Systemsintroduction to future IP – names and versions

● IP v6 – next generation Internet protocol● Preliminary versions called IP - Next Generation (IPng) ● Several proposals all called IP ng● TUBA – (TCP and UDP over Bigger Addresses) - the idea to use

the OSI connectionless protocol as drop in replacement (but not many people liked it :-))

● SIP – Simple Internet Protocol Plus – predecessor of IP v6 SIP abbreviation for “session initiation protocol”

● IP v5 naming was used for stream protocol version 2● One was selected and uses next available version number (6)

5 | 56

Communication Systemsintroduction to future IP

● Result is IP version 6 (IP v6 – around July 1994)● normally we start with the reasons to switch from a very

successful implementation to a new one

- Rapid, exponential growth of networked computers

- Shortage (limit) of the addresses

- New requirements of the internet (streaming, real-audio, video on demand)

• IP v6 is designed to be an evolutionary step from IP v4. It can be installed as a normal software upgrade in Internet devices and is interoperable with the current IP v4

• Next slide: OSI – IP v6 just replaces IP v4 on network layer ...

6 | 56

Communication Systemsintroduction to future IP – OSI and IPv6

7 | 56

Communication Systemsproblems with IPv4

● Current version of IP - version 4 - is 20 years old (rather old in the computer world)- 32 bits address range is too small

- Routing is inefficient (long routing tables, problems with aggregation)

- Bad support for mobile devices

- Security needs grew

• But some of the problems are of the late nineties and mostly solved or not as important any more ...

8 | 56

Communication Systemssuccess of IPv4

● IPv4 has shown remarkable ability to move to new technologies Other third layer protocols, like AppleTalk, IPX, NetBIOS

nearly vanished

Packet orientated IP services are used even for voice and multimedia services with stricter requirements toward quality of service

IP was open to improvements: e.g. shift from classful to classless interdomain routing

IP was able to operate on every type of new network hardware, e.g. Wireless LAN

9 | 56

Communication Systemscapabilities of IP

● IP has accommodated dramatic changes since original design Basic principles still appropriate today

Many new types of hardware

Scale of Internet and interconnected computers in private LAN

● Scaling Size - from a few tens to a few tens of millions of computers

Speed - from 9,6Kbps over GSM mobile phone networks to 10Gbps over Ethernet or frame delay WAN connections

Increased frame size (MTU) in hardware

10 | 56

● IETF has proposed entirely new version to address some specific problems

● Address space But...most are Class C and too small for many organizations

214 Class B network addresses already almost exhausted (and exhaustion was first predicted to occur a couple of years ago)

Lot of waste within the address space (whole class A network for just the loopback device, no nets starting with 0 and 255)

No geographic orientation within IP number assignment

Next generation mobile phone networks may switch over their addressing scheme

Communication Systemsintroduction to future IP – why IPv6?

11 | 56

Communication Systemsintroduction to future IP – address exhaustion

● Address space exhaustion (main argument for IP v6) Even with the excessive use of private networks, CIDR of the

old Class-A networks, ...

Inefficient routing (very long routing tables)

Think of many households getting connected to the internet, new services and new devices with demand toward addressability over an Internet

Rise of continents beside Northern America and Europe with bigger population than the “new world” and “old europe”

Around 2008 to 2015 (if we believe some forecasters, see link in mass mail) the address space is exhausted

12 | 56

Communication Systemsintroduction to future IP – address exhaustion!?

Geoff Huston● if main focus of applications stays to client/server principle● and number of peer-to-peer applications does not increase significantly● article of July 2003: exhaustion expected in 2022● http://www.potaroo.net/presentations/2003-09-04-V4-AddressLifetime.pd

f● article of september 2003: expectation even of 2045● http://www.potaroo.net/presentations/2003-09-04-V4-

AddressLifetime.pdf

http://www.potaroo.net/presentations/2003-09-04-V4-AddressLifetime.pdf


13 | 56

Communication Systemsintroduction to future IP – further reasons

● Type of service Different applications have different requirements for delivery

reliability and speed

Current IP has type of service that's not often implemented

Helper protocols for multimedia QoS seldom used

QoS routing only works hop-by-hop

more on IPv4 QoS in later lectures● Multicast

Expermental only within IP v4, not really used in production

Waste of IP numbers from 224.0.0.0 up to 254.255.255.255 for just experimental use

14 | 56

Communication Systemsintroduction to future IP – addresses

● 2128 is around 3,4*1038 possible IP addresses Should be enough :)

6,4*1028 for every human on earth

6,6*1014 for every square millimeter on earth (sea, continents and ice caps)

Opens lots of space for waste● IP v6 16 byte long addresses● So classical format as we know it, e.g. 132.230.4.44 (4 byte IP v4

address) is not really usable

15 | 56

Communication Systemsintroduction to future IP – address format

● IP v6 addresses are given in hexadecimal notation, with 2 bytes grouped together as known from ethernet MAC addresses

● Example:

2822:0000:0000:0000:0000:0005:EBD2:7008

2001:: (GEANT address prefix)

2001:07C0:0100::/48 (BelWue address prefix)

2001:07C0:0100::/64 (Freiburg university address prefix)● Try to write that address in dotted quad notation, so ...● Domain Name System becomes even more important● For better handling compression is introduced

16 | 56

Communication Systemsintroduction to future IP – address format

● Compression is achieved by Replace groups of zeros by a second colon directly following

the first

Delete leading zeros in each double byte● The address

0000:0000:0000:0000:00A5:B8C1:009C:0018 is reduced to

::A5:B6C1:9C:18

1000:0000:0000:0000:20A5:B8C1:0001:00A3 could be compressed

1000:0:0:0:20A5:B8C1:1:A3 and finally 1000::20A5:B8C1:1:A3

17 | 56

Communication SystemsIP v6 – address types

● IP v6 knows three types of addresses Classical unicast address

Multicast address

New type of address: anycast or cluster

18 | 56

Communication SystemsIP v6 – address composition

● Addresses are split into prefix and suffix as known from IPv4● No address classes - prefix/suffix boundary can fall anywhere● IPv4 broadcast flavors are subsets of multicast● Unicast addresses are distinguishable by their format prefix● The new aggregatable global address format splits address

into Global, public part

Location specific part

End system identificator

19 | 56


● Addresses split into prefix and suffix as known from IP v4● Unicast addresses are distinguishable by their format prefix● The new aggregatable global address format splits address

into Global, public part

Location specific part

End system identificator

● Global part consists of prefix, Top Level Aggregator (TLA) and Next Level Aggregator (NLA)

● Describes a site (group of machines) within the global internet

20 | 56


● TLA are only available for service providers who provide internet transit services, e.g. GEANT (2001::)

● NLAs for smaller service providers / organizations / firms which use a TLA provider, e.g. BelWue (2001:07C0:0100::)

● NLA could be split in several hierachy layers● Location specific part of the address the Site Level

Aggregator (SLA) describes subnet structure of a site and the interface ID of connected hosts

● Interface ID consists of 64bit and can contain the MAC address of the interface card for global uniqueness

21 | 56

Communication SystemsIP v6 – address space assignment

22 | 56

Communication SystemsIP v6 – address assignment example (under linux OS)

1: lo: <LOOPBACK,UP> mtu 16436 qdisc noqueue

link/loopback 00:00:00:00:00:00 brd 00:00:00:00:00:00 inet 127.0.0.1/8 scope host lo inet6 ::1/128 scope host valid_lft forever preferred_lft forever2: eth0: <BROADCAST,MULTICAST,UP> mtu 1500 qdisc pfifo_fast qlen 1000 link/ether 00:10:a4:8d:56:0a brd ff:ff:ff:ff:ff:ff inet 192.168.1.2/24 scope global eth0 inet6 fe80::210:a4ff:fe8d:560a/64 scope link valid_lft forever preferred_lft forever3: sit0: <NOARP> mtu 1480 qdisc noop link/sit 0.0.0.0 brd 0.0.0.04: eth1: <BROADCAST,MULTICAST,NOTRAILERS,UP> mtu 1500 qdisc pfifo_fast qlen 1000 link/ether 00:02:2d:09:f6:df brd ff:ff:ff:ff:ff:ff inet 10.100.5.63/16 brd 10.100.255.255 scope global eth1 inet6 fe80::202:2dff:fe09:f6df/64 scope link valid_lft forever preferred_lft forever9: tun0: <POINTOPOINT,MULTICAST,NOARP,UP> mtu 1412 qdisc pfifo_fast qlen 500 link/[65534] inet 134.76.3.40/32 scope global tun0

● Automatically configured IP v6 Addresses (lo, eth0, eth1): ip addr show

23 | 56


● Link local addresses – contain beside the prefix only the interface ID

● Used for automatic configuration or used in networks without router● Position local addresses used for sites which are not connected to

the IP v6 network (aka Internet) yet● The prefix is interchanged with the provider addresses (TLA, NLA)

in case of connection to the net● Anycast – new type of address, introduced with IP v6

24 | 56


● Special addresses:● Loopback: 0:0:0:0:0:0:0:1 = ::1● for use in tunnels: 0::FFFF:a.b.c.d

139.18.38.71 (IP v4) = ::FFFF:139.18.38.71 (IPv6) = ::FFFF:8b12:2647 (IP v6)

● IP v4-compatible-addresses ::a.b.c.d

= 0.0.0.0.0.0.139.18.38.71● Link local

● Interface address auto assignment (like 169.254.X.Y)● Start with FE80:: local MAC is last part

25 | 56

Communication SystemsIP v6 – anycast addresses

● Type of address used for number of interfaces connected to different end systems

● An anycast packet is routed to the next interface of that group● Anycast addresses are allocated within unicast address space● Idea: route packets over a subnet of a specific provider● Cluster / anycast addressing allows for duplication of services● Implementation: do not use them as source address and identify

only routers with them

26 | 56

Communication SystemsIP v6 – multicast addresses

● Now fixed part of the specification● One sender could generate packets which are routed to a number

of hosts througout the net● Multicast addresses consists of a prefix (11111111), flag and scope

field and group ID● Flag for marking group as transient or permanent (registered with

IANA)● Scope defines the coverage of address (subnet, link, location or

global)

27 | 56

Communication SystemsIP v6 – header format

● Some important changes within header format – faster processing within routers

● Header length, type of service and header checksum were removed

28 | 56


● Other header parts moved to so called extension headers (light gray)

● IP v6 header contains less information than IP v4 header● Less header information for routing speed up and avoiding of

duplication of standard information

● Other header parts moved to so called extension headers (light gray)

● IP v6 header contains less information than IP v4 header● Less header information for routing speed up and avoiding of

duplication of standard information

29 | 56


● Concept of on-the-way packet fragmentation dropped Slow down of routers

Reassembly was possible at destination only● Fragmentation is done by source and destination only (explained

later this lecture)● If packet is too big for transit intermediate routers send special

“packet too big” ICMP message Minimum MTU is 576 or 1280 byte (?)

Host has to do MTU path discovery● No header checksum – left to UDP/TCP or layer 2 protocols, like

Ethernet

30 | 56

Communication SystemsIP v6 – header fields

● Precedence, total length, time to live and protocol are replaced with traffic class, payload length, hop limit and next header (type)

31 | 56


● IPv6 header in ethereal (example of specific ICMP message)

32 | 56


● NEXT HEADER points to first extension header ● FLOW LABEL used to associate datagrams belonging to a flow

or communication between two applications Traffic class for Quality of Service routing

Specific path

Routers use FLOW LABEL to forward datagrams along prearranged path

● Base header is fixed size (other than IP v4) - 40 octets ● NEXT HEADER field in base header defines type of header

33 | 56

Communication SystemsIP v6 – header fields – traffic classes

000-111 = time insensitive (could be discarded)

1000-1111 = priority (should not be discarded)

0 = uncharacterized

1 = filler (NetNews)

2 = unattended transfer (mail)

4 = bulk (ftp)

6 = interactive (telnet)

7 = Internet control

8 = video

15 = low quality audio

34 | 56

Communication SystemsIP v6 – extension headers

● All optional information moved to extension headers● Put in between IP v6 header and payload header (e.g. TCP

header)● Extension headers (mostly) not interpreted by routers● Each header is tagged with special mark

Hop-by-hop options

Destination options header

Routing header

Fragment header

Authentication header

35 | 56


Encapsulated security payload header

Destination options header

Next header: transportation (TCP, UDP, ...)

● Extension headers have task specific format● Each header is of multiple of 8 byte ● Some extensions headers are variable sized

NEXT HEADER field in extension header defines type

HEADER LEN field gives size of extension header

36 | 56


● Special hop-by-hop option is header for so called jumbograms● Normal packet length is 65535 byte - but can be extended with

jumbo payload length of a 4 byte length indicator● But problems with UDP and TCP specification

UDP contains 16bit packet length field

TCP contains MSS (max. segment size) field set with the start of every TCP connection, could be omitted but then problems with urgent pointer

37 | 56


● Use of multiple headers:● Efficiency - header only as large as necessary ● Flexibility - can add new headers for new features ● Incremental development - can add processing for new features

to testbed; other routers will skip those headers● Conclusion: streamlined 40 byte IP header

Size is fixed

Information is reduced and mostly fix

Allows much faster processing

38 | 56

Communication SystemsIP v6 – new concept of fragmentation

● Fragmentation information kept in separate extension header ● Each fragment has base header and (inserted) fragmentation

header

39 | 56


● Entire datagram, including original header may be fragmented● IPv6 source (not intermediate routers) responsible for

fragmentation Routers simply drop datagrams larger than network MTU

Source must fragment datagram to reach destination ● Source determines path MTU

Smallest MTU on any network between source and destination

Fragments datagram to fit within that MTU

40 | 56


● Uses path MTU discovery (as discussed with IP v4 / ICMP) Source sends probe message of various sizes until destination

reached

Must be dynamic - path may change during transmission of datagrams

● Standard MTU is about 1300 octets (ethernet MTU minus special headers like PPPoE, tunnels, ...)

● New ICMP for IP v6 introduced

41 | 56

Communication SystemsIP v4 to IP v6 transition

● Typical problem – who should start with it?● IP v6 implemented in some backbones (e.g. German Telekom)● DFN is talking about testbeds, university of Münster is conducting

test installations and networks● IP v6 address space assigned for GEANT, BelWue, Uni FR

But nobody really using it● End user systems are capable of IP v6?

Linux seems to work with it for a while

WinXP was incompatible to itself with different patch levels

Not all features are implemented

42 | 56


● Step 1: Add IPv6 capable nodes into the current IP v4 infrastructure● IPv6 traffic is tunnelled in IPv4 traffic

43 | 56


● Step 2: Add more IPv6 capable nodes● Add separate IPv6 infrastructure

44 | 56


● Step 3: IPv6 dominates. Remove IPv4 infrastructure and tunnel IPv4 traffic in IPv6 traffic.● Transition finishes

45 | 56


● Several transition mechanisms proposed● IETF ngtrans working group has proposed many transition

mechanisms:● Dual Stack● Tunnelling● Translation

● Every mechanism has pros and cons choose one or more of them, depending on specific transition

scenarios

no one suits for all

46 | 56


• Dual Stack

• Both of IPv4 and IPv6 are implemented;

• IPv4 address and IPv6 address;

• DNS must be upgraded to deal with the IPv4 A records as well as the IPv6 AAAA records

47 | 56


● Tunnelling is a process whereby one type of packet in this case IP v6 - is encapsulated inside another type of packet - in

this case IP v4.

● This enables IPv4 infrastructure to carry IPv6 traffic● Most tunnelling techniques cannot work if an IPv4 address translation

(NAT) happens between the two end-points of the tunnel. ● When firewalls are used, IP protocol 4 must be allowed to go through

48 | 56


● Several tunneling mechanisms Configured tunnels

● 6to4● Tunnel broker● TSP● ISATAP● DSTM

Automatic tunnels● 6over4● Teredo ● BGP-tunnel

49 | 56


● Translation With tunnelling, communication between IP v6 nodes is

established

How about communication between IP v4-only node and IP v6-only node?

We need translation mechanisms

50 | 56


● Several mechanisms too, just names here SIIT NAT-PT ALG TRT Socks64 BIS BIA

51 | 56

Communication SystemsDNS support in IP v6

● Current DNS records store 32-bits IP v4 addresses. They must be upgraded to support the 128-bits IP v6 addresses.

● A new resource record type ‘AAAA’ is defined, to map a domain name to an IPv6 address.

● Example:www.ipv6.uni-muenster.de. IN CNAME tolot.ipv6.uni-muenster.de.

tolot.ipv6.uni-muenster.de. IN AAAA 2001:638:500:101:2e0:81ff:fe24:37c6

ns.join.uni-muenster.de. IN AAAA 2001:638:500:101::53

ns.join.uni-muenster.de. IN A 128.176.191.10

52 | 56


● New domains IP6.INT and IP6.ARPA are defined, to map an IP v6 address to a domain name.

● An IP v6 address is represented by a sequence of nibbles (nibble string) separated every four bits by dots with the suffix “.IP6.INT” or “.IP6.ARPA”.

● Example:; $ORIGIN 0.0.5.0.8.3.6.0.1.0.0.2.ip6.int.

6.0.8.3.5.b.e.f.f.f.2.0.1.0.2.0.0.0.1.0 IN PTR atlan.ipv6.uni- muenster.de.

5.f.4.7.8.d.e.f.f.f.8.1.0.e.2.0.0.0.2.0 IN PTR lemy.ipv6.uni-muenster.de.

or

; $ORIGIN 0.0.5.0.8.3.6.0.1.0.0.2.ip6.arpa.

6.0.8.3.5.b.e.f.f.f.2.0.1.0.2.0.0.0.1.0 IN PTR atlan.ipv6.uni- muenster.de.

5.f.4.7.8.d.e.f.f.f.8.1.0.e.2.0.0.0.2.0 IN PTR lemy.ipv6.uni-muenster.de.

53 | 56


● Existing queries are extended to support IP v4 and IP v6. ● When both ‘A’ and ‘AAAA’ records are listed in the DNS,

there are three different options: return only IPv6 address

return only IPv4 address

return both IPv4 and IPv6 addresses.

● The selection of which address to return, or in which order to return can affect what type of IP traffic is generated.

● BIND 9.X is fully IPv6 compliant.● Problem: name space fragmentation

Not all operating systems and not all DNS servers offer IPv6 transport lookups.

54 | 56

Communication SystemsIP v6 - conclusion

● IP v4 basic abstractions have been very successful ● IP v6 carries forward many of those abstraction... but, all the

details are changed 128-bit addresses

Base and extension headers

Source does fragmentation

New types of addresses

Address notation● Transportation header format does not needed to be changed

55 | 56

Communication SystemsIP v6 - conclusion

● But: (it is always there :-)) Idea of IP v6 was developed in 1994 (!)

Who really needs it in the moment (near future)

Who invests in new services, replaces all the routers

Who implements the outstanding features● IP v6 delivered ideas for IP v4 network operation

IPsec standard is derived from it

Auto-IP, ...

56 | 56

Communication SystemsLiterature / next lectures

● From next lecture (01.06) we will switch over to telephony networks and start with ISDN.

● Exercise sheet is handed out on next Tuesday and will be collected the Tuesday after the holiday break

● Kurose & Ross: Computer Networking, 3rd edition: Section 4.4.4 IPv6● Tanenbaum: Computer Networks, 4th edition: Section 5.6.8 IPv6● IPv4 - How long have we got?

http://www.potaroo.net/ispcolumn/2003-07-v4-address-lifetime/ale.html

● IPv4 Address Lifetime Expectancy Revisited


● http:www.ipv6.org

● Paper in English: IPv4-IPv6-Migration

Http://www.ks.uni-freiburg.de/download/studienarbeit/WS03/IPv4-IPv6-Migration.pdf.

http://www.potaroo.net/ispcolumn/2003-07-v4-address-lifetime/ale.html


http://www.ipv6.org/

http://www.ks.uni-freiburg.de/download/studienarbeit/WS03/IPv4-IPv6-Migration.pdf

communication systems 7 th lecture

Documents

ip version

osi ip v6

future ip osi

ipv4current version

future ip names

success of ip v4introduction

ip ngtuba tcp

new version