Communication Systems 7th lecture

1
Communication Systems7th lecture

• Chair of Communication Systems
• Department of Applied Sciences
• University of Freiburg
• 2006

1 56
2
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

2 56
3
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

3 56
4
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)

4 56
5
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 ...

5 56
6
Communication Systemsintroduction to future IP
OSI and IPv6
6 56
7
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 ...

7 56
8
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

8 56
9
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

9 56
10
Communication Systemsintroduction to future IP
why IPv6?

• 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

10 56
11
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

11 56
12
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.pdf
• article of september 2003 expectation even of
  2045
• http//www.potaroo.net/presentations/2003-09-04-V4
  -AddressLifetime.pdf

12 56
13
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

13 56
14
Communication Systemsintroduction to future IP
addresses

• 2128 is around 3,41038 possible IP addresses
• Should be enough )
• 6,41028 for every human on earth
• 6,61014 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

14 56
15
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
• 282200000000000000000005EBD27008
• 2001 (GEANT address prefix)
• 200107C00100/48 (BelWue address prefix)
• 200107C00100/64 (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

15 56
16
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
• 000000000000000000A5B8C1009C0018 is
  reduced to
• A5B6C19C18
• 100000000000000020A5B8C1000100A3 could be
  compressed
• 100000020A5B8C11A3 and finally
  100020A5B8C11A3

16 56
17
Communication SystemsIP v6 address types

• IP v6 knows three types of addresses
• Classical unicast address
• Multicast address
• New type of address anycast or cluster

17 56
18
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

18 56
19
Communication SystemsIP v6 address composition

• 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

19 56
20
Communication SystemsIP v6 address composition

• 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 (200107C00100)
• 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

20 56
21
Communication SystemsIP v6 address space
assignment
21 56
22
Communication SystemsIP v6 address assignment
example (under linux OS)

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

1 lo ltLOOPBACK,UPgt mtu 16436 qdisc noqueue
link/loopback 000000000000 brd
000000000000 inet 127.0.0.1/8 scope host
lo inet6 1/128 scope host valid_lft
forever preferred_lft forever 2 eth0
ltBROADCAST,MULTICAST,UPgt mtu 1500 qdisc
pfifo_fast qlen 1000 link/ether
0010a48d560a brd ffffffffffff inet
192.168.1.2/24 scope global eth0 inet6
fe80210a4fffe8d560a/64 scope link
valid_lft forever preferred_lft forever 3 sit0
ltNOARPgt mtu 1480 qdisc noop link/sit 0.0.0.0
brd 0.0.0.0 4 eth1 ltBROADCAST,MULTICAST,NOTRAILE
RS,UPgt mtu 1500 qdisc pfifo_fast qlen 1000
link/ether 00022d09f6df brd
ffffffffffff inet 10.100.5.63/16 brd
10.100.255.255 scope global eth1 inet6
fe802022dfffe09f6df/64 scope link
valid_lft forever preferred_lft forever 9 tun0
ltPOINTOPOINT,MULTICAST,NOARP,UPgt mtu 1412 qdisc
pfifo_fast qlen 500 link/65534 inet
134.76.3.40/32 scope global tun0
22 56
23
Communication SystemsIP v6 address space
assignment

• 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

23 56
24
Communication SystemsIP v6 address space
assignment

• Special addresses
• Loopback 00000001 1
• for use in tunnels 0FFFFa.b.c.d
• 139.18.38.71 (IP v4)
• FFFF139.18.38.71 (IPv6)
• FFFF8b122647 (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

24 56
25
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

25 56
26
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)

26 56
27
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

27 56
28
Communication SystemsIP v6 header format

• 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

28 56
29
Communication SystemsIP v6 header format

• 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

29 56
30
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)

30 56
31
Communication SystemsIP v6 header fields

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

31 56
32
Communication SystemsIP v6 header fields

• 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

32 56
33
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

33 56
34
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

34 56
35
Communication SystemsIP v6 extension headers

• 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

35 56
36
Communication SystemsIP v6 extension headers

• 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

36 56
37
Communication SystemsIP v6 extension headers

• 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

37 56
38
Communication SystemsIP v6 new concept of
fragmentation

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

38 56
39
Communication SystemsIP v6 new concept of
fragmentation

• 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

39 56
40
Communication SystemsIP v6 new concept of
fragmentation

• 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

40 56
41
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

41 56
42
Communication SystemsIP v4 to IP v6 transition

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

42 56
43
Communication SystemsIP v4 to IP v6 transition

• Step 2 Add more IPv6 capable nodes
• Add separate IPv6 infrastructure

43 56
44
Communication SystemsIP v4 to IP v6 transition

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

44 56
45
Communication SystemsIP v4 to IP v6 transition

• 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

45 56
46
Communication SystemsIP v4 to IP v6 transition

• 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

46 56
47
Communication SystemsIP v4 to IP v6 transition

• 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

47 56
48
Communication SystemsIP v4 to IP v6 transition

• Several tunneling mechanisms
• Configured tunnels
• 6to4
• Tunnel broker
• TSP
• ISATAP
• DSTM
• Automatic tunnels
• 6over4
• Teredo
• BGP-tunnel

48 56
49
Communication SystemsIP v4 to IP v6 transition

• 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

49 56
50
Communication SystemsIP v4 to IP v6 transition

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

50 56
51
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
  20016385001012e081fffe2437c6
• ns.join.uni-muenster.de. IN AAAA
  200163850010153
• ns.join.uni-muenster.de. IN A
  128.176.191.10

51 56
52
Communication SystemsDNS support in IP v6

• 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.

52 56
53
Communication SystemsDNS support in IP v6

• 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.

53 56
54
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

54 56
55
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, ...

55 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-addres
  s-lifetime/ale.html
• IPv4 Address Lifetime Expectancy Revisited
• http//www.potaroo.net/presentations/2003-09-04-V4
  -AddressLifetime.pdf
• httpwww.ipv6.org
• Paper in English IPv4-IPv6-Migration
• Http//www.ks.uni-freiburg.de/download/studienarbe
  it/WS03/IPv4-IPv6-Migration.pdf

56 56