Date post: | 03-Jan-2016 |
Category: |
Documents |
Upload: | mervin-cook |
View: | 219 times |
Download: | 0 times |
1 | 56
Communication Systems8th lecture
Chair of Communication SystemsDepartment of Applied Sciences
University of Freiburg2008
2 | 56
Communication SystemsLast lecture and practical course
Next Friday (06.06) start of practical course block in the comp dept. -114 9pm - 12pm
Next lecture: Friday, 13th June (here in this lecture room again)
Idea of the Internet Domain Name Ssystem 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
After dealing with an network application layer protocol (DNS using UDP/IP in most client server interaction) we will switch back again to the network layer
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 (how to integrate the new network layer protocol into
the traditional service)
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 now 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 towards the Internet infrastructure (streaming, real-time services like VoIP, 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 25+ years old (rather old in the computer world)- 32 bits address range is too small (less max. number of addresses
than inhabitants of earth, without counting the loss of addresses because of rather generous assignments)
- routing is inefficient (long routing tables, problems with aggregation)
- bad support for mobile (roaming) devices
- security needs grew
• But some of the problems are of the late nineties and mostly solved or not as important any more ... thus postponed the switch over to the new scheme
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
completely 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 2010 to 2015 (if we believe some forecasters) 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.pdf
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 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 for quite a while :)
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 representation as we know it, e.g. 132.230.4.44 (4
byte IP v4 address) would not really be human readable
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 throughout 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 in IPv4 was 576 for IPv6 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 at the moment (connectivity often worse than for IPv4)
End user systems are capable of IP v6? Linux seems to work with it for quite a while
WinXP was incompatible to itself with different patch levels, but working implementation since SP2
Vista has IPv6 fully integrated
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 (and services) 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
IP v6 delivered ideas for IP v4 network operation IPsec standard is derived from it
Auto-IP, ... (“class-B” network with prefix: 169.254)
56 | 56
Communication SystemsLiterature
The exercise sheet #4 is due to the 13th June (next lecture after exercise block) – available from the website (download section) of this course
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
IPv6 on Heise, see referenced article at the end too ...
http://www.heise.de/newsticker/IPv6-Nie-war-es-so-wichtig-wie-heute--/meldung/108754
http:www.ipv6.org
Paper on IPv4-IPv6-Migration
Http://www.ks.uni-freiburg.de/download/studienarbeit/WS03/IPv4-IPv6-Migration.pdf.