Networking - TCP/IP stack introduction and IPv6

IPv6 and Mobile IPv6
Fundamentals, new services, and
applications
Rodolfo Kohn

10/17/2006 IPv6 and MIPv6 - Fundamentals,
new services, and applications
2
Agenda
• Introduction
• Reference Model
– Physical Layer
– Data Link Layer
– Network Layer
– Transport Layer
– Application Layer
• Data Link Layer: IEEE 802.3
• Network and Transport Layers
– IP
– Routing
– UDP
– TCP
• IPv6
– Introduction
– Main Features
– Headers
– MTU and Fragmentation
– QoS Support
– Addressing Architecture
– Autoconfiguration
– Neighbor Discovery Protocol
– Comments on Routing Protocols
– Why do I need IPv6?
– New Challenges and Applications
• P2P
• Embedded Systems
• Telecommunications – IMS
• Low-power devices
• The Long Transition from IPv4 to IPv6
– IPv6 Current Status and Trends
• Mobile IPv6
– Mobility in Distributed Systems
– Mobile IPv6 Fundamentals
– New applications
– Existing MIPv6 stacks

3
Introduction
• ARPANET
• TCP/IP in universities since early eighties.
• Then to the rest of the public
• Internet in the nineties: The network of networks
where any host can contact any other host, no
matter where it is physically set up.
• People connected: chat, B2B, news, FTP, HTTP,
e-mail, low-quality voice and video.

4
Introduction
Internet
Ethernet
Router
Workstation
Token-ring
Router
Server
Computer
Router
Router
FDDI Ring
Computer
Printer
Bridge
Ethernet
Computer
Laptop

5
OSI and TCP/IP Models
Application
Presentation
Session
Transport
Network
Data Link
Physical
Application
Transport
Network
Network Access
• Open Systems Interconnection Reference Model. It divides the
communication functionalities into different layers.
• Its protocols are rarely used. The TCP/IP model is not used but
its protocols are widely used.

6
Physical Layer
• It defines the media as well as the electrical and
physical specifications of transceivers and
receivers. It includes cabling specifications,
coding schemes, voltages, RF bandwidths, etc.
• Some examples are the physical specifications
for Ethernet (e.g. 10Base-T, 100Base-TX) and
for 802.11 (e.g. RF bands)

7
Data Link Layer
• It is responsible for creating the frames and sending them to the destination
using the physical layer. It deals with access contention, framing, bit
stuffing, error detection, error correction, flow control as the main
functionalities.
• It provides a well-defined service interface to the network layer.
• IEEE has standardized the different DLLs under the 802 standards. It has
divided this layer into two sub-layers: Medium Access Control (MAC), and
Logical Link Control (LLC).
• Some examples:
– 802.11, 802.3, 802.5, 802.15, 802.16 for MAC
– 802.2 for LLC
– HDLC (High-Level Data Link Control), defined as ISO 13239, for DLL (with no
sub-layers)
• The MAC 802 standards also specify physical layer characteristics.

8
Medium Access Control
• A challenging problem in communications is how many
devices share the same medium to transmit (fiber optics,
twisted pair, coaxial cable, RF, etc.). Collisions are
possible.
– There are different techniques depending on the medium. Two
very popular are (NOT the only ones):
• Carrier Sense Multiple Access with Collision Detection (CSMA/CD)
used in 802.3 (and Ethernet).
• CSMA/CA (Collision Avoidance) used in 802.11
• MAC’s main objective is to allocate a channel to transmit
in a medium where there are competing users.
• Error detection is performed in this layer.

9
Logical Link Control
• IEEE specifies LLC in the standard 802.2
• Thus different 802 specifications (802.3, 802.11, etc.)
can provide a common interface to the network layer.
• It can provide 3 different type of services:
– Unreliable Datagram Service (could be used with 802.3, in very
reliable media, for streaming, etc.).
– Acknowledged Datagram Service (useful for wireless
technologies like 802.11)
– Reliable Connection-Oriented Service (frames are received in
the order they were sent and are not duplicated).
• It provides multiplexing for applications in the layer
above (with source and destination service access
points).

10
Network Layer
• DLL is enough to provide an application reliable
communication between different nodes in the same
LAN or a few LANs connected by bridges (or switches),
usually (not always, think of 802.3-802.11b/g bridges) of
the same technology.
• In a more complex network encompassing a large
number of LANs, MANs, and WANs of many different
technologies, the network layer is necessary to provide
end-to-end communication.

11
Network Layer
• The network layer basically routes a datagram from one node to
other in a network of networks.
• It hides different technologies from upper layers, provides a common
addressing architecture, and involves devices called routers which
connect different networks.
• A network layer could offer a connection-oriented service (mainly for
QoS) or a connectionless service (Internet).
• IP (Internet Protocol) and IPX are examples of network protocols. IP
is widely used. It is used by the Internet.
• It relies on Routing Protocols which updates the routes a datagram
must wind to reach a destination node. Examples are: RIP, OSPF,
BGP.

12
Transport Layer
• This layer basically provides end-to-end service between entities
(usually processes) in the application layer. It multiplexes network
services between different TSAPs (ports in TCP and UDP).
• This service can also provide efficiency and reliability to the
communication between application layer entities.
• It provides a set of service primitives to be used at the application
layer.
• The service can be connectionless or connection oriented.
• Examples: TCP, UDP, SCTP.

13
Application Layer
• This layer involves protocols implemented for
specific applications.
• Name Services turn essential in this layer: DNS,
OSI X.500 directory service, LDAP.
• Examples: SMTP and POP3 for e-mail, HTTP,
WAP, SNMP, DNS protocol, LDAP, RTP, RTSP,
H.323, SIP.

14
General Model
LLC
Physical
Application
Transport
Network
• Neither the OSI model nor the TCP/IP model are perfectly
implemented. Rather a mixed model is used.
• Each layer in one node communicates with the corresponding
layer in other node.
MAC
Physical
MAC
LLC
Physical
Network
MAC
LLC
Physical
Application
Transport
Network
MAC

15
How does a packet travel from A to B?
Internet
Ethernet
Router
Workstation
Token-ring
Router
Server
Computer
Router
Router
FDDI Ring
Computer
Printer
Bridge
Ethernet
Computer
Laptop
A
B

16
IEEE 802.3 LANs
• In order to understand how a packet is routed
from a source node to a destination node it is
necessary to understand how a frame is
transmitted in a LAN.
• IEEE 802.3 LANs will be used as a reference.
• We will only focus on how a frame is transmitted
from a source to a destination node.

17
Ethernet II and IEEE 802.3
• Ethernet II is a standard created by DEC, Intel,
and Xerox in 1978 based on Xerox Ethernet.
The transmission speed was 10 Mbps.
• The IEEE created the standard 802.3 based on
Ethernet II with two minor changes.

18
IEEE 802.3 Frames
S
F
D
Preamble Destination
Address
Source
Address
Length/
Type
802.2 Header
and/or Data
FCS
17 6 6 2 46 - 1500 4
• Type/Length: if <= 1500 then it is length. If it is >= 1536 (0x600)then it
is type.
• If Type/Length contains a length, the next field is an 802.2 Header
followed by the data.
• The frame contains a minimum of 64 Bytes.
• Common Type values: 0x0800 for IPv4, 0x86DD for IPv6, 0x806, etc.
• Ethernet II does not distinguish SFD and after the Source Address only
has the “Type” field. It does not distinguish an LLC sub-layer.

19
MAC Address
• MAC Addresses are 48-bit long:
– 3 Less significant octets: Vendor ID.
– 3 Most significant octets: Unique Identifier.
• IEEE defines 3 general categories:
– Unicast: identifies a single NIC.
– Broadcast: refers to all NICs in a LAN. It is the value FF-FF-FF-
FF-FF-FF.
– Multicast: refers to a group of NICs. Multicast addresses have
the least significant bit of the Most significant octet set to 1.

20
Other features
• Although we will not focus on them it is worth
mentioning other two 802.3 important features:
– CSMA/CD
– Back off exponential algorithm

21
Typical 802.3 LANs
• The most common 802.3 LANs today are:
– A multidrop configuration with a Hub.
– A point-to-point configuration with an Ethernet switch.
• Both LANs have a star physical topology.
• A LAN with a hub has a bus logical topology.

22
Hubs and Switches
• A Hub is a L1 device:
– Receives a signal in one port and transmits it through all other
ports.
– All connected nodes receive a sent frame.
• A switch works at the MAC layer:
– Based on the destination address it looks for the appropriate port
in its MAC table and sends the frame through indicated port.
– If an entry is not found in the MAC table, it performs flooding (the
frame is sent out all ports except the receiving one).

23
A Frame in the LAN
• When a node wants to transmit message to
other node in the same LAN over layer two, it
creates an IEEE 802.3 frame setting the
destination address, the source address with its
own MAC address, the Type/Length, and the
message in the data field. Then it sends the
frame.

24
A Frame in the LAN (2)
• The frame is transmitted through the cable:
– A switch relays it to the appropriate output port or generates
flooding.
• The destination node NIC receives the frame from the
cable, checks correctness, and then discovers it is
destined to itself by looking at the destination MAC
address.
• Then the data is grabbed and passed to the upper layer.

25
Fast Ethernet
• Standard Addendum IEEE 802.3u (1995).
• It only based on 10Base-T (100Base-T4/TX/FX). No
multidrop cables.
• Different coding scheme from 802.3
• 100Base-TX is Full Duplex unless hubs are used
(which require half-duplex communication).
• To make upgrading easier, all switches and almost
all 100-Mbps NICs support automatic negotiation of
optimum speed and duplexity.

26
Gigabit Ethernet
• Standard Addendum IEEE 802.3z (1998).
• Backward Compatibility is maintained.
• All configurations are point-to-point.
• The normal mode is full duplex:
• Since there is no CSMA/CD, there is no limit for the
cable length (only determined by signal strength
issues).

27
Point-to-point connections
`
`
` `
switch switch

28
Gigabit Ethernet with hubs
• But Hubs are still allowed for compatibility:
• CSMA/CD protocol must still be used.
• The maximum cable length would be 25 meters.
• To extend the cable length to 100 m, two variants
were created:
• Carrier Extension: the sending HW adds padding to
extend the frame to 512 Bytes and the receiving HW
removes it (Low Efficiency).
• Frame Bursting: the sender transmits a concatenated
sequence of multiple frames up to 512 Bytes, padding if
necessary.

29
Gigabit Ethernet more
• Types: 1000Base-SX, 1000Base-Lx, 1000Base-CX,
1000Base-T.
• New coding schemes.
• Buffer overflows are pretty probable when transmitting at 1
Gbps. So, flow control is supported:
• The receiving end can send back a control frame (Type 0x8808)
telling the sender to pause for some period of time.
• First 2 Bytes of the payload convey the command and then follow
the parameters.
• PAUSE frames are used for flow control. The parameter tells how
long to pause.

30
10-Gigabit Ethernet
• Standard approved as addendum 802.3ae in 2002.
• No official effort to achieve 100-Gbps Ethernet yet
but some have begun to talk about it
(http://www.ethernetalliance.org/technology/presenta
tions/ )
• Anyway, a 100-Gbps Ethernet standard would not
be ready before 2010.

31
Introduction to IP
• IP: Internet Protocol.
• Described by RFC 791.
• Network Layer: it provides the routing of
datagrams (or packets) from one node (host) to
another one, hiding the existence of possibly
different networks (and different technologies)
between them.
• Internet: the network of networks where any host
can contact any other host, no matter where it is
physically set up.

32
Introduction
Internet
Ethernet
Router
Workstation
Token-ring
Router
Server
Computer
Router
Router
FDDI Ring
Computer
Printer
Bridge
Ethernet
Computer
Laptop

33
Internet Address
• Each host in the internet is assigned a unique 4-
octet address that is used in all communications
with that host.
• Each address is a pair (netid, hostid).
– netid: network in the internet.
– hostid: host in the network.
• One host can have more than one physical
connections, each one with its one IP address:
multi-homed hosts (e.g. routers).

34
Classful Addressing Architecture
• Each address is self-identifying: the prefix and suffix
can be determined from the address alone. RFC 790,
1981. Useful for the next 12 years.
1 2 3
Class 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1
+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
A |0| NETID | HOSTID |
+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
B |1|0| NETID | HOSTID |
+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
C |1|1|0| NETID | HOSTID |
+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
D |1|1|1|0| Multicast Address |
+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
D |1|1|1|1| Reserved |
+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+

35
Classful Addressing
• There are especial prefixes and addresses,
e.g., the loopback address (127.x.x.x,
127.0.0.1).

36
Networks and routers
• All nodes in one network will have the same netid.
• Different L3 networks are connected through routers.
• When a node has to send a packet to a node in a
different network it has to send it to the router, the router
will take care of it.
• One way to do this: the node detects the destination
node is in a different network and sends it to the default
gateway, which is set beforehand.
• Other possibility is the use of proxy ARP.

37
A packet sent (I)
• From host 192.13.13.2 to 192.15.13.201.
C
R3
E
F
G
R1 R2
192.13.13.2
192.13.13.23
192.230.215.7 192.230.215.2
192.230.215.0
192.13.13.0
192.211.7.0
192.211.7.3 192.211.7.2
192.15.13.201
192.15.13.0
192.15.13.76
R4
R1 receives the pkt, it looks at the destination address,
it sends the pkt to 192.230.215.2
R2 sends the pkt to 192.211.7.2
R3 sends the pkt to 192.15.13.201

38
Questions
• Having R1’s IP address, how does node G know the physical address of
router R1 in order to send him the IP datagram?
• How does R2 know R3’s physical address?
• How does R3 know E’s physical address?
C
R3
E
F
G
R1 R2
192.13.13.2
192.13.13.23
192.230.215.7 192.230.215.2
192.230.215.0
192.13.13.0
192.211.7.0
192.211.7.3 192.211.7.2
192.15.13.201
192.15.13.0
192.15.13.76
R4

39
ARP
• ARP: Address Resolution Protocol.
• In order to send an IP datagram through a physical
link, the datagram must be conveyed by a lower layer
frame containing physical addresses.
• An IP address must be mapped into a physical
address. For example, an Ethernet address (48 bits).
• ARP is a protocol that lets a node obtain the physical
address of a destination node, given the destination IP
address.

40
ARP Procedure
• Host A wants to send a packet to Host B:
• Host A only knows the IP address of the destination host B. A
needs to know B’s physical address.
• Host A broadcasts an ARP frame asking for the physical address
of the IP address owner (B).
• Host B receives the frame, it recognizes itself as the destination
and it sends a unicast reply with its physical address to A. Every
other node drops the broadcast frame.
• A receives the reply, it updates an ARP cache with mappings IP
address-Physical address. Afterwards, it sends out the frame
within the physical network.
• Next time A will obtain B’s physical address from its ARP cache.

41
Questions (II)
• Having E’s address, how does G know that it has to send the packet to
R1?
• How does R1 send the packet to the next hop?
• How is the packet routed from G to E?
C
R3
E
F
G
R1 R2
192.13.13.2
192.13.13.23
192.230.215.7 192.230.215.2
192.230.215.0
192.13.13.0
192.211.7.0
192.211.7.3 192.211.7.2
192.15.13.201
192.15.13.0
192.15.13.76
R4

42
Routing I
The main responsibility of the network layer is to make a
packet reach a destination interface from a source interface.
IP provides a Connectionless Delivery Service.
IP implements the mechanism to route one packet through
possibly different physical networks.
This function is implemented by routers. Multi-homed hosts
can provide this service as well.

43
Routing II
The routing of a packet is based on the netid part of the destination address.
A Router maintains a routing table in which, for each network known to this
router, there is an entry indicating the next hop (maybe next router) through
which the destination is to be reached.
Network
192.13.13.0
Network
192.230.215.0
Network
192.211.7.0
Network
192.15.13.0R1 R2 R3
192.13.13.0 192.230.215.7
192.230.215.0 Interface 0
192.211.7.0 Interface 1
192.15.13.0 192.211.7.2
192.13.13.0 Interface 0
192.230.215.0 Interface 1
192.211.7.0 192.230.215.2
192.15.13.0 192.230.215.2

44
Routing III
To determine which is the next hop to access a destination
network different criteria can be used. It can be based on the
number of hops, delay, etc.
An algorithm is used to choose one among different paths.
Most used routing algorithms:
• Distance Vector Routing: RIP (Bellman-Ford)
• Link State Routing: OSPF
• Between Autonomous Systems: BGP
Routers have to exchange information with neighbors in order
to let them know the networks it can reach and to know the
networks it can access through the neighbors.

45
Routing IV
RIP
Each router advertises the networks that it can reach and the number of
hops to access it.
These advertisements are broadcasted every 30 seconds.
A router receives the advertisement from its networks and it updates its
routing tables:
• An entry for that network does not exist- It adds the entry
• There is an entry with a higher number of hops - It is updated.
Every entry has a timeout of 180 seconds.
network next-hop cost

46
A packet sent II
A packet is sent from host 192.13.13.2 to 192.15.13.201
C
R3
E
F
G
R1 R2
192.13.13.2
192.13.13.23
192.230.215.7 192.230.215.2
192.230.215.0
192.13.13.0
192.211.7.0
192.211.7.3 192.211.7.2
192.15.13.201
192.15.13.0
192.15.13.76
R4
• Before sending the packet in the physical network the destination physical address has to be
obtained.
• Each router looks into its routing table in order to get the next hop for the packet.
• R1 knows it can reach network 192.15.13 in 3 hops by R2 (192.230.215.2).

47
IP characteristics
The service characteristics are:
• It is connectionless
• It is unreliable
• It provides a best-effort delivery.

49
Fragmentation
A datagram may have to travel through different networks to reach
the destination.
Each physical network has its MTU: Maximum Transfer Unit.
The datagram size may be too long for transmission in one physical
network: it has to be fragmented.
Each fragment is sent independently.
In the destination host, fragments are reassembled.
Fragment offset indicates the offset of the data inside the non
fragmented datagram.
Identification field is equal in all fragment.
The Flags field is used: do not fragment bit, more fragments bit.

50
Transport Layer
It provides:
The interface for applications to access the network services. TSAPs.
It permits a communication between one entity in the source host
and another entity in the destination host.
A message is received from any application, through the TSAP and
sent to the network layer: Multiplexing.
When a message is received from the network layer it is delivered to
the appropriate application (TSAP): Demultiplexing.

51
UDP
It provides an unreliable, connectionless delivery service over IP to
transport a message from one application in one machine to other
application in the destiantion host.
It uses the port concept to multiplex and demultiplex messages
between the applications and the UDP layer.
In a UDP message a source port and a destination port are specified.

52
TCP I
It provides a connection oriented and reliable service for
messages transmission between two applications in possibly
different hosts. Described by RFC 793.
Its features:
• Connection oriented.
• Stream oriented.
• Buffered transfer.
• Unstructured stream.
• Full duplex.

53
TCP II
A connection is established between two end points
characterized by:
• Host id.
• Port.

54
TCP-Segment Format
The TCP segment format:
0 1 2 3
0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1
+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
| Source Port | Destination Port |
+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
| Sequence Number |
+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
| Acknowledgment Number |
+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
| Data | |U|A|E|R|S|F| |
| Offset| Reserved |R|C|O|S|Y|I| Window |
| | |G|K|L|T|N|N| |
+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
| Checksum | Urgent Pointer |
+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
| Options | Padding |
+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
| data |
+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+

55
TCP-Reliability and Window Size
Reliability is achieved by the acknowledgement of the data
sent and retransmission.
Sequence number: sequence number of the first byte sent in the
segment.
Acknowledgement number: sequence number of the last byte
being acknowledged.
Flow control and window size: the field Window let one end-
point control the flow of data sent by the other end-point by
varying the size of the window that can be sent. If it wants the
other application to stop sending data it sends an
acknowledgement segment with Window size = 0.

56
TCP-Connection establishment
The connection is established by a three-way handshake.
To initiate a connection a segment is sent with the SYN flag
set.
In the destination host one application has to be listening.

57
TCP-Connection establishment
The connection is established by a three-way handshake.
To initiate a connection a segment is sent with the SYN flag
set.
In the destination host one application has to be listening.
Host 1 - Connects Host 2 - ListeningMessages
Send SYN seq=x
Receive SYN
Send SYN seq=y, ACK=x+1
Receive SYN + ACK segments
Send ACK=y+1
Receive ACK segment

58
TCP– The state machine
PICTURE SOURCE: http://en.wikipedia.org/wiki/Image:TCP_state_diagram.png

59
TCP-Congestion
Whenever there is a timeout of a segment that did not receive
its corresponding ACK, TCP considers there is congestion in
the network and it tries to avoid it.
It lowers the sending rate. Algorithms slow start and
congestion avoidance.
It sends one segments and waits for the ACK.
When the ACK arrives it sends two segments and as long as
there is no timeout expiration the rate keeps on increasing.

60
A message sent
A segment is sent from host 192.13.13.2 to 192.15.13.201
C
R3
E
F
G
R1 R2
192.13.13.2
192.13.13.23
192.230.215.7 192.230.215.2
192.230.215.0
192.13.13.0
192.211.7.0
192.211.7.3 192.211.7.2
192.15.13.201
192.15.13.0
192.15.13.76
R4
• Application A1, in G, wants to send a message to application A2 that is listening in a port P2 of E.
• It connects to the port (E,P2) a port P1 is assigned by the OS to A1 in G. A TCP connection is
Established: two sockets are created.
• A1 sends a message to A2 through its socket. A TCP segment is sent. TCP in E sends an
acknowledgement.
• The TCP connection is closed.

61
Address shortage
Subnetting, CIDR, and NAT
• During its whole evolution , as it expanded, IP had to
deal with address shortage.
• In 1984, RFC 917 described subnetting. This helped the
distribution of aggregated addresses.
• In 1993, to solve the problem of Class-B Address
Exhaustion and routing information overload, Classless
Inter-Domain Routing was standardized (RFC 1519).
• Also NAT has collaborated to provide connectivity to
many nodes with only one public address.

62
Internet scales up
Different devices connected
• From the academic world, Internet has expanded to connect the rest
of the public. It is already widely used in the business world and new
applications are continuously dreamt up.
• Now many devices are offered a limited connectivity.
• Cellular phones might have Internet connectivity as long as they are
within the Carrier Provider Radio Area Network.
• Other type of devices can be connected (Camera, Camcorder,
robots, cars, etc.). Today some have limited PAN connectivity with
Bluetooth (802.15.1) or Zigbee (802.15.4) but they rarely have
Internet access.
• As the necessity to connect all type of devices increases, it is more
obvious that the current IP address space is not enough.

63
Internet scales up
More people connected
• As a result of many societies development, more
people gets access to the Internet.
• This is another challenge the current IP address
space could hardly cope with.

64
IPv6 - Agenda
• A packet in the Internet.
• IPv6 main features.
• IPv6 Headers.
• MTU.
• QoS Support.
• IPv6 Addresses.
• Autoconfiguration.
• Neighbor Discovery Protocol.
• Why do I need IPv6?
• IPv6 new challenges.

65
The Internet
Internet
Ethernet
Router
Workstation
Token-ring
Router
Server
Computer
Router
Router
FDDI Ring
Computer
Printer
Bridge
Computer
Laptop
switch
UMTS Core Network & UTRAN
Node-B
UMTS Mobile Node
GGSN
Router
Video
Access Point switch
Router
802.11g
IP routes a packet in the
Internet from a source
node to one or many
destination nodes.

66
How is this task carried out?
• A Protocol is needed that basically indicates each node
which is the next node (next hop) a packet must be sent
to.
• The Protocol must define a packet that conveys data
from a source node to a destination node.
• The protocol must define an addressing mechanism to
be able to locate a destination node and route a packet
to it.
• The Protocol must run on different types of networks.
• The Protocol can provide security.
• The Protocol can provide mechanisms for service
extensions.

67
In the Internet model.
How does a packet sent from G arrive to E?
C
R3
E
F
G
R1 R2
R4
Routers

68
IPv6
• Internet Protocol (IP) version 6 provides all the
required services.
• IPv6 is the next version of the Internet Protocol
after IPv4. Some documents call it IPng (IP next
generation).

69
IPv6 - Main Features
• Address size is 16 Bytes.
• Extended address hierarchy.
• New header format: 1 Base Header + n
Extension Headers.
• Different support for options (comparing to IPv4).
• Support for protocol extensions (e.g. Mobility
support).

70
IPv6 - Main Features
• Support for autoconfiguration and renumbering.
• Support for resource allocation: Flow label and
service type.
• Support for authentication and privacy is
specified: IPsec natively supported

71
IPv6 Packet
• In all protocols, a message usually has a basic
structure made of:
• A Header: contains Protocol information.
• A Payload: contains data relative to the current layer that must be
passed to the layer above (it could be a message belonging to the
layer above).
• In IP the message relayed is called “packet”.

72
IPv6 – Headers
• An IPv6 datagram has 40-octect Base Header
and n Extension Headers.
• Advantages:
– Improves performance on header processing.
– The user can choose which extension headers to
include and which to omit.
– Flexibility for new options: more extension
headers can be added.

73
IPv6 – Base Header
+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
|Version| Traffic Class | Flow Label |
+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
| Payload Length | Next Header | Hop Limit |
+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
| |
+ +
| |
+ Source Address +
| |
+ +
| |
+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
| |
+ +
| |
+ Destination Address +
| |
+ +
| |
+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+

74
IPv6 – Headers
• In order to Extract any header information from an IPv6 datagram a
sequential search from the base header is required. Every header
has a “next header” field.
• Different possibilities for datagrams with base header and n
extension headers:
Base Header
NEXT=TCP
TCP segment
Base Header
NEXT=hbh
hbh Header
NEXT=TCP
TCP segment
Base Header
NEXT=hbh
hbh Header
NEXT=AUTH
AUTH Header
NEXT=TCP
TCP segment

75
IPv6 – Extension Headers
• There are Extension headers useful for each hop from
the source to the destination and extension headers to
be processed at the destination node.
• A full implementation of IPv6 includes implementation
of the following extension headers:
– Hop-by-Hop options: to be processes by routers.
– Routing (Type 0): source routing.
– Fragment: fragmentation header (instead of IPv4 fields).
– Destination Options: to be processed by the destination
node.
– Authentication (AH).
– Encapsulating Security Payload (ESP).

76
Hop-by-Hop and Destination Options
C
R3
E
F
G
R1 R2
R4
Hop-by-Hop Options
Destination Options

77
IPv6 – Extension Headers
• When there are more than one header there is a
recommended order:
– Base header
– Hop-by-Hop options header
– Destination options header-1
– Source Routing header (Type 0)
– Type 2 routing header (for mobility)
– Fragment header
– Authentication header (AH).
– Encapsulating Security Payload (ESP) header
– Destination Options header-2
– Upper-layer header

78
Different frame sizes in different link
technologies (different MTU)
C
R3
E
F
G
R1 R2
R4
MTU: n MTU: m
MTU: Maximum Transfer Unit.
MTU: k

79
IPv6 – Path MTU Discovery and
Fragmentation
• Unlike IPv4, Fragmentation is only performed by
source nodes, not by routers along the path.
• IPv6 requires that every link in the Internet have an
MTU of 1280 octets or greater.
• Path MTU discovery is recommended for source
nodes in order to discover a path MTU larger than
1280 octets.
• It is not as easy as in IPv4 to change the route from
the source to the destination. If a route turns
unavailable and the alternative route has a smaller
MTU, the packet could be dropped.

80
IPv6 – Quality Of Service
Flow Label
• It is intended to define a flow of packets from a source to
a destination for which special handling is required, for
example for a “real-time” service.
Traffic Class
• This field is intended to define one of different traffic
classes. Used for DiffServ. Along with the flow label it is
going to be used to define QoS.

81
IPv6 –Addresses
• IPv6 addresses are 128-bit “identifiers” for
interfaces and sets of interfaces.
• Length: 16 Bytes or 128 bits.
• IPv6 addresses of all types are assigned to
interfaces, not nodes.
• Since each interface belongs to a single node,
the interface’s unicast addresses can be used as
an “identifier” for that node.
• IPv6 Addressing Architecture defined by RFC
4291 (IETF).

82
IPv6 – Text Representation of Addresses
• 16 Bytes: 128 bits.
• Text representation: Colon Hexadecimal
Notation.
2004:FFED:01:0:0:0:0CC:A1BC
2004:FFED:1::0CC:A1BC
0:0:0:0:128:FA:1234:5678
::128:F0A:1234:5678
::192.168.0.5 (x:x:x:x:x:x:d.d.d.d for embeddedIPv4
addresses)
Note: Zero compression can be applied only once.

83
IPv6 – Text Representation of Addresses
• Prefixes: the leftmost bits of an address
expressing certain information about it.
FC00::/7 => A 7-bit prefix specified as 1111 110b
2004:FFED:1::/64 => A 64-bit prefix
2001:0DB8:0000:CD30:0000:0000:0000:0000/60
2001:0DB8::cd30:0:0:0:0/60
2001:0DB8:0:cd30::/60
• Node address and prefix:
2001:0DB8:0000:CD31:AB01:3201:A00C:1AF8/60
• Zones?

84
IPv6 - Addresses
Address types
• Unicast: specifies a single interface (a single node).
• Anycast: specifies a set of interfaces (typically
belonging to different nodes). They are unicast
addresses assigned to different interfaces. The pkt is
delivered to exactly one of them, the nearest one.
• Multicast: specifies a set of interfaces (typically
belonging to different nodes). The pkt is delivered to all
of them.
Note: broadcast is handled with multicast addresses.

85
IPv6 - Addresses
Types and Scopes
Address Type Binary Prefix
IPv6
Notation
Unspecified 00…0 ::/128
Loopback 00…1 ::1/128
Multicast Addresses 1111 1111 FF00::/8
Link-Local Unicast Addresses 1111 1110 10 FE80::/10
Unique Local IPv6 Unicast Addresses (RFC 4193) 1111 1100 FC00::/7
Site-Local Unicast Addresses (Deprecated RFC 3879) 1111 1110 11 FEC0::/10
Global Unicast Addresses (Aggregatable) everything else
IPv4-Mapped IPv6 Address: ::FFF:d.d.d.d (used in dual-stack nodes)

86
IPv6 – Scopes
Internet
Computer
Computer
Workstation
Server
Router
Computer
Computer
Computer
Server
Link-local
Site-local
Global
(global, unique local)

87
IPv6 – Addresses
Interface Identifier (ID)
• Interface IDs in IPv6 unicast addresses are used to identify
interfaces on a link.
• They can be unique within a broader scope.
• For all unicast addresses, except those that start with the binary
value 000, Interface IDs are required to be 64 bits long and to be
constructed in Modified EUI-64 format.
• Modified EUI-64 format-based interface IDs may have universal
scope (stemmed from IEEE 802 48-bit MAC or IEEE EUI-64 IDs, u
bit=1) or local scope (u bit=0). Figure below shows the company ID
part.
0 0 0 1 1 2
|0 7 8 5 6 3|
+----+----+----+----+----+----+
|cccc|ccug|cccc|cccc|cccc|cccc|
+----+----+----+----+----+----+

88
IPv6 – Addresses
Modified EUI-64 format Interface Identifier (ID)
• Example: creation of a Modified EUI-64 format Interface Identifier
from an IEEE 48-bit MAC identifier.
MAC Addr: 00:07:A5:60:0E:F0 (6 Bytes)
02:07:A5: FF:FE: 60:0E:F0Modified EUI-64 ID:

89
IPv6 – Addresses
Configuration based on Interface Identifier (ID)
•MAC Addr: 00:07:A5:60:0E:F0 (6 Bytes)
•Modified EUI-64 ID: 02:07:A5:FF:FE:60:0E:F0
•Link Local Unicast Addr: FE80::0207:A5FF:FE60:0EF0

90
IPv6 – Addresses
Global Unicast Address
• General Format:
| n bits | m bits | 128-n-m bits |
+------------------------+-----------+----------------------------+
| global routing prefix | subnet ID | interface ID |
+------------------------+-----------+----------------------------+
• Global Routing Prefix is hierarchically-structured:
• Regional Internet Registries (RIPE, LACNIC, ARIN, APNIC, AfriNIC): /23
prefixes?
• Providers, research networks, etc.: usually /32 prefixes.
• Site with subnetworks: /48
• Subnetworks: /64
• Global Unicast Address: 2002:C3D4:6EED:1:0207:A5FF:FE60:0EF0 /64

92
IPv6 – Addresses
Zones
• A Zone (or scoped zone) is a connected region of topology of a given
scope.
• Determined by the context and not encoded in the address.
• Different types:
• An interface on a node comprises a single zone of interface-local scope.
• Each link and interfaces attached to it comprise a single zone of link-
local scope.
• A single zone of global scope comprises all links and interfaces in the
Internet.
• The boundaries of other zones are defined by network administrators.
• At each node, a zone is represented by an index.

93
IPv6 – Addresses
How zones affect us
•Text representation:
FE80::0207:A5FF:FE60:0EF0%2
•Programmer: the zone index is set in the
sin6_scope_id field of the sockaddr_in6 structure
(Socket API).
•Network Administrator:
• When a router transmits routing information through one
interface it must not transmit information about zones the
interface does not belong to.
• Probably different zones within an organization will be
assigned different Unique Local IPv6 Unicast Addresses.

94
IPv6 Addresses must be assigned to the
Network Interfaces
C
R3
E
F
G
R1 R2
R4

95
IPv6 – Address Assignment
• It is necessary to assign addresses to a node’s
network interfaces so that packets destined to
this node can reach it.
• IPv6 allows Address Autoconfiguration.

96
IPv6 – Address Autoconfiguration
• The procedure a host follows to create its interface
addresses without manual assistance.
• Autoconfiguration is performed on a per-interface basis
on multicast-capable links in multicast-capable
interfaces. Begins when the interface becomes
enabled.
• It uses Neighbor Discovery Protocol messages.
• Address Autoconfiguration can be:
– Stateless: without assistance of stateful servers.
– Stateful: with assistance of a stateful server; for example
DHCPv6.
• Both Stateless and Stateful are complementary.

97
Address Autoconfiguration Steps – Link-Local Address
PC3
switch
PC2
PC1
1. Interface is enabled.
2. “Tentative” link-local address
is formed.
3. Duplicate Address Detection
is performed: A Neighbor
Solicitation is sent to
Solicited-Node Multicast
Address of the Target Address
4. If the node ascertains the link-
local address is unique, it is
assigned to the interface.
5. The node has only link
connectivity.
FE80::0207:A5FF:FE60:0EF0
FE80::0207:A5FF:FE60:0EF0
Neighbor Solicitation to
FF02:0:0:0:0:1:FF60:0EF0
2
4
3

98
Stateless Address Autoconfiguration - Next Steps
PC3
switch
PC2
PC1
1. Routers periodically send out
Router Advertisements with
Prefix Information Options.
2. If a prefix published in the
Prefix Information Option has
the appropriate flag set, it can
be used to form a Global
Address or a Local Unicast
Address.
3. The node has Site or Internet-
wide connectivity. 2002:C3D4:6EED:1:0207:A5FF:FE60:0EF0
Router Advertisements.
Prefix:
2002:C3D4:6EED:1
ROUTER
1
2

99
How is an IPv6 packet physically relayed
from a node’s NIC to the next hop’s NIC?
C
R3
E
F
G
R1 R2
R4
A
NIC: Network Interface Card

100
Neighbor Discovery Protocol for IPv6
• This protocol is used in IPv6 to:
– Determine neighbor’s link-layer address, i.e. address
resolution (NO ARP).
– Determine neighbor bidirectional reachability: Neighbor
Unreachability Detection.
– Discover neighboring routers.
– Learning link specific parameters: network prefixes, MTU, etc.
– Next-Hop determination and Redirect.
– Duplicate Address Detection in Address Autoconfiguration.
– Proxy support.
• Messages used in ND Protocol are part of ICMPv6.

101
IPv6 – ND Protocol Address Resolution
PC3
switch
PC2
PC1
1. PC3 is to send a packet to
2002:C3D4:6EED:1:0207:A5FF:
FE61:2AB8 but first it needs to
know the corresponding HW
address.
2. PC3 sends a Neighbor
Solicitation asking for the owner
of the destination address: it is
sent to the solicited-node
multicast address for the target
addr.
3. PC1 reads the Neighbor
Solicitation and responds with a
Neighbor Advertisement telling
its link-layer address.
4. PC3 can physically send the
packet to the HW (link-layer)
address of PC1.
2002:C3D4:6EED:1:0207:A5FF:FE60:0EF0
2002:C3D4:6EED:1:0207:A5FF:FE61:2AB8

102
IPv6 – Neighbor Discovery Protocol ICMPv6
Messages
• ND Protocol defines the following ICMPv6
messages:
– Neighbor Solicitation (Duplicate Address Detection,
Address Resolution, Neighbor Unreachability Detection).
– Neighbor Advertisement (Duplicate Address Detection,
Address Resolution , Neighbor Unreachability Detection,
Proxy Support).
– Router Solicitation (Address Autoconfiguration, Network
Parameters).
– Router Advertisement (Address Autoconfiguration,
Network Parameters).
– Redirect (Next-Hop correction).

103
Now a Packet sent from G arrives to A
C
R3
E
F
G
R1 R2
R4
A
The packet is physically sent from G’NIC to A’s NIC
Neighbor
Solicitation

104
Now a Packet sent from G arrives to A
C
R3
E
F
G
R1 R2
R4
A
The packet is physically sent from G’NIC to A’s NIC
Neighbor
Advertisement

105
Now a Packet sent from G arrives to E
C
R3
E
F
G
R1 R2
R4
Router
Advertisement
2002:C3D4:6EED:1:0207:A5FF:FE60:0EF0
20B1:CAD1:55ED:1:0207:A5FF:FE61:2AB8
For the sake of simplicity, in this presentation it is assumed intermediate routers know
where to send packets.

106
Now a Packet sent from G arrives to E
C
R3
E
F
G
R1 R2
R4
2002:C3D4:6EED:1:0207:A5FF:FE60:0EF0
20B1:CAD1:55ED:1:0207:A5FF:FE61:2AB8
For the sake of simplicity, in this presentation it is assumed intermediate routers know
where to send packets.

107
IPv6 Operation Summary
• The topics presented encompass the basics about
IPv6:
• Packet Structure: Base Header and Extension Headers.
• IPv6 Addresses.
• Stateless Address Autoconfiguration.
• Neighbor Discovery Protocol Basics.
• The topics specified above are enough to send a
packet from a source node to a destination node.

108
IPv6 core specifications
• According to the IPv6 Forum (IPv6 Forum Roadmap &
Vision,
http://www.6journal.org/archive/00000261/02/WWC_IPv6_Forum_Roadmap__Visio
n_2010_v6.pdf)
• RFC 1981: Path MTU Discovery.
• RFC 2460: IPv6 Protocol.
• RFC 2461: IPv6 Neighbor Discovery.
• RFC 2462: IPv6 Stateless Auto-Configuration.
• RFC 2463: Internet Control Message Protocol for IPv6
(ICMPv6).
• RFC 4291: IPv6 Address Architecture.
• RFC 4301: Security Architecture for IP (IPsec).
• IPv6 over “XYZ” Link Layer (Ethernet, ATM, PPP, etc.)

109
Why do I need IPv6?
• There is no killer application for IPv6.
• Although not compatible, IPv6 can be seen as an
evolution from the experience gained with IPv4 since
1981.
• The major great advantage is the wider address space.
This brings a lot of changes to the communications
industry.

110
Why do I need IPv6?
• The wide address space allows each device in the
world to have its own IPv6 address:
• Every device can be connected (PDA, cell phone, robot,
sensor, camera, medical equipment, etc.).
• NAT is not necessary to save IP addresses. Without NAT it is
possible to achieve E2E communications between nodes.
• This leverages the creation of new ubiquitous services without
a centralized server, using P2P architectures:
• A digital camera can send a picture to a PDA without the
need of a server in the middle.
• My MP3 Player can directly access my PC and get read
music files.
• I can turn on the air conditioner at home from my cell phone
without a server in the middle.

111
Why do I need IPv6?
• IPsec is part of the Protocol.
• It is easier to extend the Protocol by adding new Extension
Headers. For example: Mobile IPv6.
• Network parameters can be managed from the router configuration
(Network Renumbering and new configuration features).
• Support for Stateless Address Autoconfiguration, allowing “Plug-
and-Play” capability.
• Better support for QoS (based on Type of Service and flows).
• Better support for mobility: Route Optimization.

112
IPv6 new challenges and opportunities
• IPv6 also involves new challenges:
• How can a chat application using E2E communications and
without a centralized server be implemented? How participating
nodes, which join and leave the chat session, can be tracked?
• A proprietary protocol can be implemented at layer 3 by just
adding new Extension Headers.
• TCP might need some changes in order to use all IPv6
advantages.
• Applications using TCP and UDP might need some changes (for
example, the case of prefix lifetime).
• Transition from IPv4 to IPv6 is said to be easy. However, some
features must be considered. In addition, IPv4 and IPv6 must
coexist for a long time (some say 25 years).
• In E2E connectivity, it is more obvious that Security is essential.

113
IPv6 and other standards groups
• 3GPP mandates that IMS (IP Multimedia Subsystems) will
internally use IPv6 (Within the Core Network).
• 3GPP2 determined that IMS will internally use IPv4 or IPv6.
• In the IETF, 6lowpan WG (IPv6 over Low power WPAN) is
investigating the use of IPv6 over 802.15.4
• IMS Links:
• http://www.motorola.com/networkoperators/pdfs/Motorola-IP-Multimedia-Subsystem-
Brochure.pdf#search=%22%2BIMS%20%2B%22Ip%20Multimedia%20Subsystem%22%22
• http://www.ericsson.com/technology/whitepapers/ims_ip_multimedia_subsystem.pdf#search=%22%2BIMS%20%2B%2
2Ip%20Multimedia%20Subsystem%22%22

114
IPv6 Status and Future
• IPv6 Penetration estimates generated by RTI International in
United States (http://www.nist.gov/director/prog-ofc/report05-2.pdf).

115
• IPv6 Rollout in the world (IPv6 and Broadband, EC IST 6LINK,
http://www.ipv6tf.org/pdf/ISTClusterbooklet2005.pdf).

116
• Major IPv6 activities worldwide (IPv6 and Broadband, EC IST
6LINK, http://www.ipv6tf.org/pdf/ISTClusterbooklet2005.pdf).

117
• The Asia-Pacific region is leading IPv6 adoption with Japan at the
head.
• In most developed regions, government agencies have great
participation in the IPv6 adoption effort.
• The US DoD has announced support for IPv6 back in June 2003.
Its networks will completely be IPv6 capable by 2008.
• In Argentina there is a recently formed IPv6 Task Force

118
Questions?

119
Link References
• IST IPv6 Portal: http://www.ist-ipv6.org
• IPv6 Forum: http://www.ipv6forum.org
• Argentina IPv6 Task Force: http://www.ar.ipv6tf.org (?)
• NTIA comments on IPv6: http://www.ntia.doc.gov/ntiahome/ntiageneral/ipv6
• North American IPv6 Task Force: http://www.nav6tf.org/
• IPv6 and Broadband: www.ist-ipv6.org/pdf/ISTClusterbooklet2005.pdf
• IPv6 Forum Roadmap & Vision:
http://www.6journal.org/archive/00000261/02/WWC_IPv6_Forum_Roadmap__Visio
n_2010_v6.pdf
• IETF: www.ietf.org
• HP IPv6 tutorial: http://h10026.www1.hp.com/netipv6/IPv6_seminar_Oct2004.pdf
• IPv4-IPv6Transition:
• http://www.6journal.org/archive/00000046/01/trans_ipv6_v014.pdf
• http://usipv6.unixprogram.com/North_American_IPv6_Summit_2004/
IPv6_Tutorial/marc_blanchet_tutorial_ipv6_transition.pdf

120
C Source Code:
IPv6 Client and IPv6 Server

121
Example: Server
Socket Creation (Family Type AF_INET6):
if ((listenSocket = socket(AF_INET6, SOCK_STREAM, 0)) == -1)
{
printf("ERROR - Socket could not be open: %dn", errno);
return 1;
}
Address and Port to bind to:
bzero(&my_addr, sizeof(my_addr));
my_addr.sin6_family = AF_INET6;
my_addr.sin6_port = htons(LISTEN_PORT);
my_addr.sin6_addr = in6addr_any;
my_addr.sin6_scope_id = 2;
Socket Structure Data Type for IPv6:
struct sockaddr_in6 my_addr;
struct sockaddr_in6 cliaddr;

122
Example: Server
Listen to the socket (Like IPv4):
res = listen(listenSocket, BACKLOG);
Bind to the interface and port (Like IPv4):
res = bind(listenSocket, (struct sockaddr *)&my_addr,
sizeof(struct sockaddr_in6));
Accept connections (Like IPv4):
newSocket = accept(listenSocket, (struct sockaddr *) &cliaddr,
(socklen_t*)&size);
Convert client address to print:
if (inet_ntop(AF_INET6, (const void *)&cliaddr.sin6_addr,
addrBuffer, INET6_ADDRSTRLEN) == NULL)
Read message (Like IPv4):
if ((bytesRead = read(newSocket, (void *)buffer, 100)) == -1)
Close sockets(Like IPv4):
close(newSocket);
close(listenSocket);

123
Example: Client
Socket Creation (Family Type AF_INET6):
if ((listenSocket = socket(AF_INET6, SOCK_STREAM, 0)) == -1)
{
printf("ERROR - Socket could not be open: %dn", errno);
return 1;
}
Address and Port to bind to:
#define DEST_ADDR "::1“
bzero(&their_addr, sizeof(their_addr));
their_addr.sin6_family = AF_INET6;
their_addr.sin6_port = htons(LISTEN_PORT);
their_addr.sin6_scope_id = 2;
if (inet_pton(AF_INET6, DEST_ADDR,(void *)&their_addr.sin6_addr)
!= 1)
Socket Structure Data Type for IPv6:
struct sockaddr_in6 their_addr;

124
Example: Client
Connect to Destination (Like IPv4):
if (connect(sock, (const struct sockaddr *)&their_addr,
(socklen_t)sizeof(their_addr)) == -1)
Address and Port to bind to (Like any fd):
if ((bytesWritten = write(sock, (const void *)buffer,
sizeof(buffer))) == -1)
Close socket(Like any fd):
close(sock);

Networking - TCP/IP stack introduction and IPv6

More Related Content

What's hot

Viewers also liked

Similar to Networking - TCP/IP stack introduction and IPv6

Recently uploaded

Networking - TCP/IP stack introduction and IPv6

Editor's Notes