communication systems 7 th lecture
DESCRIPTION
Chair of Communication Systems Department of Applied Sciences University of Freiburg 2006. Communication Systems 7 th lecture. 1 | 56. Communication Systems Last lecture and practical course. Thursday (25.05) is a holiday, no lecture, no practical. - PowerPoint PPT PresentationTRANSCRIPT
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
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
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
20 | 56
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 (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
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
24 | 56
Communication SystemsIP v6 – address space assignment
● 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
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
29 | 56
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
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
Communication SystemsIP v6 – header fields
● IPv6 header in ethereal (example of specific ICMP message)
32 | 56
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
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
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
36 | 56
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
37 | 56
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
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
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
40 | 56
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
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
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
43 | 56
Communication SystemsIP v4 to IP v6 transition
● Step 2: Add more IPv6 capable nodes● Add separate IPv6 infrastructure
44 | 56
Communication SystemsIP v4 to IP v6 transition
● Step 3: IPv6 dominates. Remove IPv4 infrastructure and tunnel IPv4 traffic in IPv6 traffic.● Transition finishes
45 | 56
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
46 | 56
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
47 | 56
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
48 | 56
Communication SystemsIP v4 to IP v6 transition
● Several tunneling mechanisms Configured tunnels
● 6to4● Tunnel broker● TSP● ISATAP● DSTM
Automatic tunnels● 6over4● Teredo ● BGP-tunnel
49 | 56
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
50 | 56
Communication SystemsIP v4 to IP v6 transition
● 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
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.
53 | 56
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.
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.potaroo.net/presentations/2003-09-04-V4-AddressLifetime.pdf
● http:www.ipv6.org
● Paper in English: IPv4-IPv6-Migration
Http://www.ks.uni-freiburg.de/download/studienarbeit/WS03/IPv4-IPv6-Migration.pdf.