Transcript Network Layer - Home Pages of People@DU

Network Layer
The context
Store and Forward Packet
Switching

A packet is stored in entirety, checksum
recomputed at every hop and forwarded.
Services provided to the transport
layer



Services should be independent of the
router technology.
Transport layer be shielded from the
topology of the routers.
The network addresses made available to
the TL should use a uniform numbering
plan, even across LANs and WANs.
Two schools of thought

Whether NL provides a CL service to the
TL (Internet Community), or

Whether It provides a CO service to the TL
(telephone companies)
CL VS CO


CL : Since subnet is inherently unreliable,
host should do the error , flow control,
packet ordering etc themselves.
CO : subnet must provide reliable service;
QoS is imporatnt
Implementation of CL

Packets are called datagrams and the
subnet is called datagram subnet.
Implementation of Connectionless
Service
Routing within a diagram subnet.
Implementation of Connection-Oriented
Service
Routing within a virtual-circuit subnet.
Comparison of Virtual-Circuit and
Datagram Subnets
5-4
Comparison of Virtual-Circuit and
Datagram Subnets
5-4
Routing Algorithms : desirable
properties






Correctness
Simplicity
Robustness
Stability : converge to equilibrium
Fairness
Optimality : Minimize mean packet delay,
maximize throughput – conflicting, since queuing
near full capacity implies long delays

Minimize the number of hops – improves delays as
well as throughput
Equilibrium
Fairness Vs Optimality
Conflict between fairness and optimality.
Routing Algorithms : Adaptive or
Non-adaptive

Non-adaptive : Static

Adaptive : Dynamic
Routing Algorithms
•
•
•
•
•
•
•
•
•
•
The Optimality Principle
Shortest Path Routing
Flooding
Distance Vector Routing
Link State Routing
Hierarchical Routing
Broadcast Routing
Multicast Routing
Routing for Mobile Hosts
Routing in Ad Hoc Networks
The Optimality Principle
(a) A subnet. (b) A sink tree for
router B.
Shortest Path Routing: Static
The first 5 steps used in computing the
shortest path from A to D.
The arrows indicate the working node.
Flooding
Problems : Jamming/Congestion
Solutions :
1.
Hop count in the header
2.
Sequence number for every source : discard
duplicate packets.
Adv: Though not practical for routine routing but
useful when a system starts afresh.
Distance Vector Routing :
Dynamic
(a) A subnet. (b) Input from A, I, H, K,
and the new
routing table for J.
Distance Vector Routing (2)
The count-to-infinity problem.
Link State Routing
Each router must do the following:
 Discover its neighbors, learn their network
address.
 Measure the delay or cost to each of its
neighbors.
 Construct a packet telling all it has just
learned.
 Send this packet to all other routers.
 Compute the shortest path to every other
router.
Discovering the neighbours

Send a HELLO packet

The other router sends a reply telling who
he is
Problem
(a) Nine routers and a LAN. (b) A
graph model of (a).
Measuring the delay to
neighbours

Send an ECHO packet

The receiver sends it back immediately
with a time stamp
Building Link State Packets
(a) A subnet. (b) The link state packets for
this subnet.
Distributing the Link State
Packets


Packets are flooded,
Seq no.s to check jamming
Distributing the Link State
Packets
The packet buffer for router B in the
previous slide (Fig. 5-13).
Hierarchical Routing
Hierarchical routing.
Broadcasting in point to point
subnet




Individual packet for each destination
Flooding
Multidestination routing – each packet has to
contain either a list of destinations or a bit map
indicating all the destinations.
Spanning tree – best in terms of channel
utilization – minimum number of copies are
generated …. But knowledge of some spanning
tree at every router is required ..which is
sometimes available (as in LSR) but sometimes
not(as in DVR)
Reverse path forwarding

Approximates the behaviour of Spanning
tree algo .. When routers have no
knowledge of spanning trees.
Reverse path forwarding
Reverse path forwarding. (a) A subnet. (b) a Sink tree. (c) The
tree built by reverse path forwarding.
Reverse path forwarding





No knowledge of spanning trees required
No bit map etc required
Simple and easy to implement
No additional CPU time required
Not the best … but in practice reasonably
efficient.
Multicast Routing in the subnet

MR – sending a message to members
within a group
Multicast Routing
(a) A network. (b) A spanning tree for the leftmost router.
(c) A multicast tree for group 1. (d) A multicast tree for group 2.
Congestion Control
What is congestion?

Too many packets in a part of the subnet,
too many to handle.
Some of the Causes of congestion

Lot of packets arriving at 3-4 lines of a router, all
needing the same output line. As a result, a
queue will build up on the outgoing line. When
the Q is full, packets will start dropping.

Increasing the size of the buffer may help initially
but after a certain limit may have adverse effect-- By the time a packet moves up in the Q, it
times out and retransmitted increasing the load
on the network.
Causes of congestion





Slow processors: If the computational powers of
a router are weak, it will take time in bookkeeping and processing, again resulting in
building up of Qs.
Low bandwidth
In fact, Slow processors and High bandwidth, or
Fast processors but Low Bandwidth
a good combination of fast processors and high
bandwidth is required to improve the situation.
Solutions to Congestion

Classified into two categories

Open Loop solutions: Static Solutions, take
preventive measures by good design, but no
corrections are done once the system is up, that is do
not take the current state of the system into account.
Adv : simplicity

Closed Loop Solutions: Based on feedback:



a router detects congestion,
pass the information (feedback) to nodes where action can
be taken, say the sender, and
adjust the system parameters to fix the problem.
Open Loop Solutions

Policy-Decisions at various levels
Congestion Prevention Policies
Policies that affect congestion.
5-26
Open Loop Solutions contd..




Flow Control --- window size
Acknowledgement policy --- piggyback or
not
Out-of-order policy --- Go back N/Selective
repeat
Retransmission Policy --- Time out etc
Closed Loop Solutions

Detecting/Predicting congestion: Let ‘u’ be any parameter being monitored
say, output line utilization or queue length or buffer utilization; Let ‘a’ be a
constant between 0 and 1

U_new is predicted as folllows:
u_new = a * u_old + (1-a)*u_new
A value of ‘u’ going above a certain threshold is considered as a situation for
congestion

Congestion Control in Virtual-Circuit Subnets

Congestion Control in Datagram Subnets
Congestion Control in Virtual-Circuit
Subnets


Admission Control : No more new VCs
Allow new VCs but route them around the
problem area.
Congestion Control in Virtual-Circuit
Subnets
(a)
A congested subnet. (b) A redrawn
subnet, eliminates congestion and a
virtual circuit from A to B.
Congestion Control in Datagram
Subnets

Warning Bit is set in the forwarded packet and
copied en-route by the destination in the
acknowledgement packet.

Choke Packets are sent back to the sender as a
feedback.


The corrective measures (slow down the
transmission) are taken only at the source, Fig a.
The corrective measures (slow down the forwarding
of the packets coming from the source) are taken at
every Hop en-route the Choke Packet, Fig b.
Hop-by-Hop
Choke
Packets
(a) A choke packet that
affects only the
source.
(b) A choke packet that
affects each hop it
passes through.
Load Shedding in Datagram Subnet

Throw the packet when nothing else works


Select the packet to be dropped randomly
Select intelligently : For example
In a file transfer, older is better so drop the later
ones. Dropping older would cause a gap at the
destination and more packets would have to be
retransmitted.
 In audio/video file: loosing few bits is not important,
so dropping older is better, no retransmission may
be asked by the destination.
 Or, sender may specify the priority

Internetworking
Connecting Networks of different
type by routers



So far we have assumed that our subnet is
connected to LANs of same type. So the only
role of routers was to route the packets.
As the h/w and n/w gets cheaper, the place
where decisions are made move downwards in
the hierarchy in an organization. For eg ..in a
univ .. Each department decides on its own what
type of LAN do they want. Hence Maths deptt
may have an Ethernet LAN whereas CS may
have a wireless LAN.
Now the univ. must be able to provide a subnet
to connect these two LANs so that a host on one
Connecting LANs of different
types by bridge

Sort of assumes that their NL are same..

If that is not the case, then the scheme
studied earlier will not suffice.
Connecting Networks
A collection of interconnected networks.
How Networks Differ
5-43 networks can differ.
Some of the many ways
How Networks Can Be
Connected
(a) Two Ethernets connected by a
switch.
(b) Two Ethernets connected by
routers.
Concatenated Virtual Circuits
Internetworking using concatenated virtual
circuits.
Connectionless Internetworking
A connectionless internet.
Tunneling
Tunneling a packet from Paris to London.
Tunneling (2)
Tunneling a car from France to England.
Autonomous System

Each network in an Internet is
independent and hence it is called an
Autonomous System.
Internetwork Routing
(a) An internetwork. (b) A graph of the
internetwork.
Two level routing algorithm in IN

Once a graph of multi-protocol
routers(gateways) is constructed , routing
algorithms such as DVR and LSR can be
applied. This leads to 2-level routing in
internetworks


Interior Gateway Protocol (within a network).
Exterior Gateway Protocol (across the
networks)
Routing in Internet contd..










A host H1 on LAN 1 wants to send a packet to a host H2 on LAN 2.
A packet is prepared by NL of H1 with Network address of H2 but,
Encapsulated in a frame by the DLLwith MAC address of the multi-protocol
router connected to LAN 1
Packet arrives at MPR1 on LAN1
NL at MPR1 uses the Network address to decide which MPR2 to forward
the packet to.
If the Network Protocol used by the Network through which MPR1 sends a
packet to MPR2 is same as that used by LAN1, the packet is sent directly
with no change in the packet.
Else, the packet is encapsulated in the payload field of the packet of the
Network Protocol used by the connecting Network and tunneled.
Of course, assuming that the connecting network uses MAC addressing,
DLL of MPR1 puts the MAC address of MPR2.
The process is repeated at MPR2
Until the packet reaches the destination network.
Fragmentation
(a) Transparent fragmentation. (b)
Nontransparent fragmentation.
Fragmentation (2)
Fragmentation when the elementary data size is 1
byte.
(a) Original packet, containing 10 data bytes.
(b) Fragments after passing through a network with
maximum packet size of 8 payload bytes plus
The Network Layer in the Internet
•
•
•
•
•
•
•
The IP Protocol
IP Addresses
Internet Control Protocols
OSPF – The Interior Gateway Routing
Protocol
BGP – The Exterior Gateway Routing
Protocol
Internet Multicasting
Mobile IP
Design Principles for Internet








Make sure it works.
Keep it simple.
Make clear choices.
Exploit modularity.
Expect heterogeneity.
Avoid static options and parameters.
Look for a good design; it need not be
perfect.
Be strict when sending and tolerant when
receiving.
Collection of Subnetworks
The Internet is an interconnected collection of many
networks.
The IP Protocol
The IPv4 (Internet Protocol) header.
IP Protocol contd..








Version
IHL
Type of Service
Total Length, Identification, DF, MF, Fragment
Offset
TTL
Protocol
Header Checksum
Options
Version Number

To let several versions to work
simultaneously…actually two during a
transition period which takes years.
IP Protocol contd..
IHL : IP Header Length

Header length is variable :


specified as number of 32 bit words
20 bytes (5 32 bit words) to 60 bytes (15 32
bit words)
IP Protocol contd..
Type of Service

Various combinations of reliability and speed
can be specified here. For Example, digitized
voice prefers fast over error-free transmission
and file transfer prefers error-free over fast
transmission. Routers on the way use this
information to choose a path. If the shortest path
is error-prone, it may use an alternate path to
transfer a packet for file-transfer whereas for
voice packet it will choose the shortest path
even if it is error-prone.
Type of Service contd..

First 3 bits : precedence or priority bits

Next 3 are flags : D (delay), T (Throughput) and
R (Reliability)

Allow the routers to make a choice between high
throughput and high delay link like satellite and
low throughput, low delay link like leased line.
IP Protocol contd..
Other fields




Total Length: Header + Data: 16 bits …..65,535
bytes
Identification No.: To know which datagram the
fragment belongs to
DF, MF: Don’t Fragment and More Fragment
Fragment Offset:



specified in number of elementary fragment unit i.e. 8
bytes i.e multiple of 8 bytes.
13 bits … 2^13=8192
Total Datagram : 8192 * 8.
IP Protocol contd..
Time to Live

Specified in seconds and decremented on
every Hop and even when in the queue.

In practice, Number of Hops is used.
IP Protocol contd..
Protocol


Mentions the number of Transport Layer
Protocol to which the packet must be
handed over. For eg. TCP/UDP or any
other.
These TP are assigned numbers (called
ports) are unique across the globe.
IP Protocol contd.
Header Checksum

Computed at every hop


to take care of error that might creep in due to
bad bits in the router memory.
TTL field changes at every hop
IP Protocol contd..
Some of the IP Options
.
5-54
IP Addresses : 32 bit number

Assigned by central naming authority
ICANN: Internet Corporation for Assigned
Names and Numbers
IP Addresses
IP address formats.
Dotted decimal Notation



202.14.13.1
192.133.13.5
Each 8 bit block is written as its decimal
eqvt.
IP Addresses (2)
Special IP addresses.
Routing Tables

Routing tables at each router has two
types of entries:



(network,0)
(this network, host)
Note the first type of entry. By keeping only
the network number and not all the IP
addresses belonging to a distant network, the
size of the routing table is greatly reduced.
Problems in Class based
Addressing
What to do when the network grows beyond the current
maximum.
Solution : To allow a network to be split into several
parts (called subnet) for internal use but appear to be
a single network to the outside world.
Note : The word “subnet” has been used again now to
define a different context. The difference will be clear
from the context.
Subnets
A campus network consisting of LANs for
various departments.
Subnets (2)
A class B network subnetted into 64 subnets.
Routing Tables

Routing tables at each router now has
three types of entries:



(network,0)
(this network, subnet, 0)
(this network, subnet, host)
CDR – Classless InterDomain
Routing
A set of IP address assignments.
5-59
IP Addresses are scarce



Most of the people are opting for
broadband Internet Connection i.e. a
permanent IP address
One solution : IPV6 : 128 bit address, but
it will take years to come.
A quick solution is needed : NAT
NAT – Network Address
Translation
Placement and operation of a NAT box.
Three reserved ranges of IP
addresses for Internal Use




10.0.0.0 -- 10.255.255.255/8
172.16.0.0 – 172.31.255.255/12
192.168.0.0 – 192.168.255.255/16
For example : Delhi University Intranet

IP addresses are of the form : 10.25.2.23
Gateway : 10.25.1.4

DNS : 10.2.1.13, 10.2.1.16

Working of NAT



Before a packet from internal host exits the
company (connected to ISP through say a
leased line) as shown in the figure or
a packet from a home/ business user
connected through broadband to ISP exits
ISP
local IP address (10.x.y.z etc) is mapped to
the company's/ ISP' s true IP address and
sent out.
NAT contd..





NAT must remember the internal
addresses. How?
It uses Header of the Transport Layer.
Source Port field is replaced by a pointer to
an entry in a table (maintained by NAT box)
containing the local address.
How does it remember the Source Port
then?
The table entry contains the source port
Other Network Layer Protocols
used in Internet: Internet Control
Protocols




ICMP : Internet Control Message Protocol :
used by the routers to monitor the Internet
for unexpected events, and also to test the
Internet from time to time.
ARP : Address Resolution Protocol: maps
an IP address to a unique DLL address
RARP : Reverse ARP: DLL address to IP
BOOTP, DHCP and others
Internet Control Message
Protocol
The principal ICMP message types.
5-61
ARP


Although every machine on the Internet
has one or more IP addresses, they are
not sufficient for sending packets as the
DLL h/w doesn’t understand the IP
addresses.
How are IP addresses mapped to DLL
addresses?
ARP– The Address Resolution
Protocol
Three interconnected /24 networks: two
Ethernets and an FDDI ring.
ARP: How does it work




Suppose H1 wants to send a packet to H2.
It sends a broadcast packet (broadcast address
in the DLL address for destination) on its LAN
asking: “ who owns the IP address
192.31.65.5”?
Everyone on LAN1 gets it but only H2 replies
with its DLL address.
H1 now prepares the data packet meant for H2
and sends it.
ARP: How does it work





Now Suppose H1 wants to send a packet to H4.
It sends a broadcast packet (broadcast address in the
DLL address for destination) on its LAN asking: “ who
owns the IP address 192.31.63.8”?
Everyone on LAN1 gets it but this time the router replies
with its DLL address.
H1 now prepares the data packet meant for H4 with the
DLL address of the router and sends it to the router.
The process is repeated on the FDDI ring and so on.
DHCP: Dynamic Host Configuration
Protocol


Maps DLL address to IP address
When a diskless workstation boots from a remote machine, how
does it get its IP address? Remember,




IP addresses are assigned in the s/w and,
DLL address in h/w
When a machine boots from a local OS it learns its IP address from
the settings already done (settings in TCP/IP etc) but,
When it boots from a remote machine it gets the binary image of its
OS from a remote file server. The IP address cannot be included in
this binary image for then a separate binary image will be required
to boot each host.
DHCP contd..





Such a host (say H1) asks a question: “ My DLL address is……..Does
anyone know my IP address?”
Another machine running DHCP server responds back with the IP address
of H1.
DHCP Server maintains a table of (DLL address, IP address) of the nodes it
serves.
How does H1 get down (IP address of) to the DHCP server?
If the DHCP server is on the same LAN as the host H1, there is no problem



A broadcast packet from H1 is enough.
However, if DHCP server is on a remote machine, a machine called DHCP
relay agent (who knows the IP address of the DHCP server) is required on
each LAN.
DHCP relay agent relays the packet from H1 to the DHCP server and back.
Dynamic Host Configuration
Protocol
Operation of DHCP.
Routing Protocols in Internet

OSPF: Open Shortest Path First (Interior
Gateway Protocol)

BGP: Border Gateway Protocol (Exterior
Gateway Protocol)
OSPF – Open Shortest Path First:
Interior Gateway Protocol

Routing algorithm within an AS:


Initially when As were small .. A variant of
DVR(RIP) was used.
DVR suffered from count to infinity problem
and was replaced by Link State Routing
algorithm in 1979.
OSPF – Basic form


AS is small – same as LSR.
To understand OSPF lets see how an AS
looks like : AS is a collection of routers
and networks.
OSPF – hierarchical structure –
when ASes became large


When ASes became large they were
further divided into areas
Each area is a collection of networks and
routers now.
OSPF – works in an AS

Three types of Routers



Internal routers (used for routing within an area hence
keep the Link State Database for routers within an
area and run Shortest Path Algorithm locally )
Area border routers (Inter-area routing through
backbone routers, keeps the LSDB for all the areas(2
areas in case of non-backbone area router and may
be more than 2 areas for a backbone area router)
each is connected to and runs a possibly different
SPA for each area separately.
Backbone routers : could be internal or area border
routers of Area 0
OSPF - Hierarchical Structure
The relation between ASes, backbones, and areas in OSPF.
OSPF – Hierarchical Structure
contd..
Internal Routers
These are routers that are only connected
to other routers or networks within a single
area. They maintain an LSDB for only that
area, and really have no knowledge of the
topology of other areas.
Area Border Routers
These are routers that connect to routers
or networks in more than one area. They
maintain an LSDB for each area of which
they are a part. They also participate in the
backbone.
Backbone Routers
These are routers that are part of the OSPF
backbone. By definition, this includes all area
border routers, since those routers pass routing
information between areas.
However, a backbone router may also be a router
that connects only to other backbone (or area
border) routers, and is therefore not part of any
area (other than Area 0).
Back
BGP – Border Gateway Protocol
Exterior Gateway Protocol

Issues –





No transit traffic thru certain Ases
Never put Iraq on a route starting at Pentagon
DO not use US to get from British Coumbia to Ontario
Only transit Albania if there is no alternative to the
destination
Traffic starting or ending at IBM should not transit
Microsoft
These kind of issues cannot be solved by computing
the Shortest Path trees.
BGP cntd




BGP is basically a DVR protocol.
However it maintains and advertise the
entire path.
Since choice of entire paths are available
.. decisions around the routers can be
taken on a datagram basis.
Hence, it takes care of the count to infinity
problem also.
BGP contd..
(a) A set of BGP routers.
(b)
Information sent to F.
The Main IPv6 Header
The IPv6 fixed header (required).
Extension Headers
5-69
IPv6 extension headers.
Extension Headers (2)
The hop-by-hop extension header for large datagrams
(jumbograms).
Extension Headers (3)
The extension header for routing.