Professional Documents
Culture Documents
3.1 Switching:
A Switched network consists of a series of interlinked node called Switches. Switches are
network and / or software devices capable of creating temporary connections between 2 or more
devices linked to the switch but not to each other. There are three methods in Switching namely
Circuit Switching, Packet Switching and Message Switching.
are called trunks.Any phone line can connect one user to another user through the phone
system; the user has a line assigned randomly. This is called as dial-up or Switched network.
Telephone networks are connection oriented because they require the setting up of connection
before the actual transfer of information can take place. An end-to-end path setup beginning of a
session , dedicated to the application , then released at the end of session is called as Circuit
Switching.
Direct
Switch 1
path
Switch 2
Switch
Fig: 3.2 Routing in Circuit tandem
Switching
Routes between end Office A to B:
Route 1: A 1 B
Route 2: A 3 B
Switch tandem
Route 3: A B
Switch 1
Propagation delay
Processing delay
Call accept
User data
Link
Nodes
Link
Link
4
Acknowledgement signal
Message 2
Store &
Forwar
d
Message 1
Store &
Forward
Message 1
Message 2
Store &
Forwar
Store &
Forwar
PKT 1
PKT 2
PKT 3
PKT1
PKT 2
PKT1
PKT 2
PKT 3
PKT 3
Ackowledgement Packet
Nodes
1
2
Fig: 3.6 Timing Diagram
Demerits:
1. Transmission delay will be increased.
2. Each node requires large capacity for storing.
3.1.2.1 Properties:
Fixed number of inputs & outputs.
The point to point link is maintained between the switches.
Adding new host will not reduce the performance.
High scalable.
It uses the identifier to find the host. It is classified into 3 types:
Datagram approach
Virtual circuit approach
Source routing
Assumptions:
Each host has unique identifier (globally unique).
Two ways to find port
Number the ports
By name of port
Datagram approach:
Connectionless approach.
Each packet is routed independently through the network. Header is attached with each packet. It
provides all information to route the packet to the destination.
If the transmission line is busy then the packet is placed in the queue until the line becomes free.
Characteristics:
Send anywhere & anytime.
No way to know the capacity of the network.
Independent forwarding.
No problem in case of failure.
switch
consults
forwarding
table
before
starts
the
forwarding.
Port
C
E
A
B
D
3
2
3
1
3
Host C is connected with switch2 via switch1 at port 3(in switch 2). Likewise, for Host E,A,B,&
D are connected with switch2 via switch1.
Virtual circuit approach
A preplanned route is established before any packets are sent.Because of the route is fixed for the
duration of the logical connection, it is referred as VIRTUAL CIRCUIT.No routing decisions
are required.Each packet contains a virtual circuit identifier, instead of destination address.
Data Transfer
Connection setup
A connection is established in each of the switches between the source & the destination.
Data Transfer
The connection state for a single switch connection consists of an entry in VC table in each
switch through which the connection passes.
E.g.: VC Entry on a single switch contains 4 parts i.e.
1. Virtual Circuit Identifier (VCI).
2. An incoming interface.
3. An outgoing interface. In which packets for this VC leave the switch.
4. A different VCI will be used for outgoing packets.
VC Entry Table:
Incoming Interface
Incoming VCI
Outgoing Interface
Outgoing VCI
INCOMING
OUTGOING
OUTGOING
VCI
INFORMATION
VCI
11
INCOMING
OUTGOING
VCI
INFORMATION
11
OUTGOING
VCI
7
3.1.3 SOURCE
INCOMING
OUTGOING
VCI
INFORMATION
OUTGOING
VCI
4
ROUTING :
All the information about n/w topology that is required to switch a packet across the n/w
is provided by the source host
IMPLEMENTATION :
Assign a number to each output of each switch and to place the number in the header of the
packet. The switching is simple. For each packet that arrives on an i/p,the switch would read the
PORT NUMBER in the header and transmit the packet on that o/p
D C B
Ptr D C
B A
Ptr
3.2 IP : ( Internetworking
D C
Protocol)
Datagram delivery :
Every datagram carries enough information to let the n/w forward the packet to its
correct destination
It doesnot make any attempt to recover from the failure .so it is called as a unreliable
service
Sometimes the packet can get deliverd out of order & sometimes the same packet can
delivered more than once
Packet Format
0
4
Version
8
IHLen
16
Tos
Identification
TTL
19
Length
Flags
Protocol
31
Offset
Checksum
Source Address
Destination Address
Pad(Variable)
Data
Start of header
Ident=x
(DF)
0
Offset=0
Rest of header
1400 bytes of data
Start of header
Ident=x
(DF) Offset =0
1
Rest of header
512 bytes of data
Start of header
Start of header
Ident=x
(DF) Offset=64
1
Rest of header
512 data bytes
Ident=x
(DF) Offset=128
0
Rest of header
376 data bytes
Fig
Addressing scheme :
Way of identifying all the hosts so we need global addressing scheme
one in which no 2 hosts have the same address
Ethernet addresses are globally unique,but that alone does not sufficient for an addressing
scheme in a large internetwork
It contains 2 parts
n/w part : identifies the n/w to which host is attached
host part : identifies each host uniquely on that particular n/w
IP addresses types :
There are 5 categories they are
Class
Class
Class
Class
Class
A
B
C
D (for multicast)
E (Currently not in use )
Class A:
0
Network
Host
Class B :
1 0
Network
Host
Class C:
1 1 0
Network
Host
Class A:
First bit is 0
of addresses in 4 billion possible IP Addresses
n/w part -> 7 bits,host part -> 24 bits
can accommodate up to 2^24- 2 host (16 million)
Class B :
3.2.4 ROUTING:
It is the process of building up the tables that allow the correct output for a packet to be
determined.
Main Points for Forwarding of IP Datagrams:
Every IP datagram contains IP address of the destination host.
Network part of an IP address identifies a single physical network.
All hosts and routers that share the same network part of their address are connected to
the same physical network.
Every physical network is part of the Internet has at least one router.
Steps:
A datagram is sent from a source host to a destination host, possible passing through several
routers along the way.
It compares the network part of the destination address with the network part of the address of
each of its network interfaces.
If a match occurs then that means that the destination lies on the same physical network as the
interface and the packet can be directly delivered over that network.
If the node is not connected to the same physical network as the destination node, then it needs to
send the datagram to a router.
Each node will have a choice of several routers and so it needs to pick the best one, or at least
one that has a reasonable chance of getting the datagram closer to its destination.
The router finds the correct next hop by consulting its forwarding table.
Datagram Forwarding Algorithm:
-For a host with only one interface and only a default router in its forwarding table.
-The forwarding table has a list of <NetworkNum, NextHop> pairs.
Simplified Algorithm:
else
deliver packet to default router
-The goal is to enable each host on a network to build up a table of mappings between IP
addresses and link-level addresses.
-Since these mappings may change over time, the entries are timed out periodically and removed.
It happens on the order of every 15 minutes.
Steps:
-If a host wants to send an IP datagram to a host(or router) that it knows to be on the same
network, it first checks for mapping in the cache.
-If no mapping is found, it needs to invoke ARP over the network.
-It broadcasts an ARP query onto the network. Each host receives the query and checks to see if
it matches its IP address.
-If it does match, the host sends a response message that contains its link-layer address back to
the origination of the query.
-The originator adds the information contained in the response to its ARP table.
-When a host broadcasts a query message, each host on the network can learn the senders linklevel and IP addresses.
-It places that information in its ARP table.
-If the host already has an entry for that host in its table, it refreshes this entry. If that host is the
target of the query, it adds information about the sender to its table, even if it did not already
have an entry for the host.
-If a host is not the target and does not already have an entry for the source in its ARP table, then
it does not add an entry for the source.
HLen=48
PLen=32
Operation
Two Messages:
Request specifies both the sender &target machine.
Reply send IP address As response directly to the machine making the request
3.5.2 DHCP:
-Automated configuration protocol.
-DHCP server is responsible for providing configuration information to hosts.
There is at least one DHCP server for an administrative domain.
-The configuration information for each host could be stored in the DHCP server
and automatically retrieved by each host when it is booted or connected to the network.
-It is stored in a table that is indexed by some form of unique client identifier,
typically the hardware address.
-In this model, DHCP server saves and maintains a pool of available addresses that
it hands out to hosts on demand.
-It reduces the amount of configuration an administrator must do, since now it is
only necessary to allocate a range of IP addresses to each network.
Goal : minimize the amount of manual configuration required for a host to function
Problem : server discovered.
To Contact A DHCP Server:
A newly booted or attached host sends a DHCP DISCOVER message to a special IP
address(255.255.255.255) that is an IP broadcast address
IT will be received by all host & routs on that n/w
DHCP server should be one of the nodes from the n/w
Steps:
1.A newly attached host broadcast a DHCP DISCOVER msg to a special IP address
2.The server would reply to the host that generted the discovery msg.In same n/w,it has a
potentially large no.of server that need to be correctly &consistently configured
3.DHCP uses the concept of a relay agent.There is atleast one relay agent on each n/w,&
it is configured with just one piece of info(IP address of the DHCP server).
4. When a relay agent receives a DHCP DISCOVER msg,it unicasts it to the DHCP
server & awaits the respones ,which it will then send back to the requesting client
the msg is actually sent using a protocol called User DatagramProtocol(UDP) that runs
over IP
broadcast
DHCP
Relay
others
n/ws
server
DHCP
16
HLen
24
31
Hops
X id
Seconds
Flags
Ciadds
Yiadds
Siadds
Giadds
Cbadds(16 bytes)
Sname(64 bytes)
File(128 bytes)
Options
1. Operation: Specifies whether the message is a request or a response.
2. Htype: Specifies the n/w hardware type.
3. HLen: Specifies the length of a h/w address.
4. HOPS: Specifies how many servers forwarded the request.
5.Xid: Transaction Identifies which provides a value that a client can use to find if an incoming
response matches its request.
Since it makes the binding between physical host &IP address much more dynamic.This may
make the n/w manager's job more difficult.
ICMP
message
IP header
Frame header
IP data
Frame data
Trailer
16
31
Type
Code
Checksum
Description
Echo reply
Destination unreachable
Router
can't
destination
Source quench
find
the
Redirect
Echo request
11
Time exceeded
12
Parameter pblm
13
14
17
18
Router advertisement
10
Router solicitation
Type:3
Code: 0 to 15
Checksum
Unused(All os)
IP datagram + header
Fig:3.17 Destination unreachable format
1.code: It has 0 to 15 different message
eg)
Not unreachable
Host unreachable
Protocol unreachable
Port unreachable
2.Checksum: 16 bit one's complement of the one's complement sum of the ICMP msg starting
with the ICMP type
3.Unused: 32 bit are not used & ignored
Code:0
Checksum
Unused(All os)
IP datagram + header
Fig:3.18 Source quench format
1.code: It is always 0.
All the other fields are same.
Code:0/1
Checksum
Unused(All os)
IP datagram + header
Fig:3.19 Time exceeded format
1.Code:
0
1
Code:0/1
Checksum
Unused(All os)
IP datagram + header
Fig 3.20 Parameter problem format
1.Code:
It is 0 when the pointer field indicates the error
It is set to 1, when there is no error
2.Pointer:
Identifies the original datagram's header where the error was detected.
3.6.3.5 Redirection:
If a router A receives a datagram to forward from a host in the same n/w. The router
A checks its routing path is via another router (Routre B) on the same n/w.
For redirecting message to the host.
It can only be sent by routers.
Type:5
Code:0 to 3
Checksum
3.6.4 QUERY
ICMP query message are of four types
1.Echo request &reply
Type:8(or)0
Code:0
Checksum
Identifier
Sequence number
Optional data
Fig 3.22 Echo Request & Reply format
1.Type:
0
1
Echo request
Echo reply
Code :0
Checksum
Original Timestamp:
Time in units of 32 milliseconds when the datagram was transmitted from the source
Receive Timestamp:
Time, when the datagram was received at the destination
Transmit Timestamp:
Time, when the Timestamp reply datagram reply datagram was transmitted from the
destination
Sending Time = Receive Timestamp - Original Timestamp
Receiving Time = Returned Time -Transmit Time
Round -Trip Time = Sending time + Received Time
C)Address Mask Request & Reply:
Used by a host to find what its address mask is on a network.
Type:17/18
Code :0
Identifier
Checksum
Sequence No.
Address mask
Fig: 3.23 Address Mask Request & Reply
Type: 17 --> request, 18 --> reply.
Address mask: Contains 32-bit subnet mask for the network (enabled only in the response)
d)Router Solicitation & Advertisement:
Router Solicitation --> To send notification to all routers, which broadcast on current
network.
Router advertisement --> broadcast their routing information
Type :10
Identifier
Code :0
Checksum
Sequence No.
Type :19
Code :0
No. Of addresses
checksum
lifetime
Router address 1
Address preference 1
Router address 2
Address preference 2
.
.
.
Fig:3.25 Router Advertisement message format
FIFO:
Idea: first come, first served queueing.
It is simple.
The first packet that arrives at a router is the first packet to be transmitted.
If a packet arrives & the queue is full, then the router discards that packet,as shown in the
following figure.
It is done without regard to which flow the packet belongs to which flow the packet belongs to as
how much important the packet is. It is called as tail drop.
Drop tail/Drop Policy:
Since the packets that arrive at the tail end of the FIFO are dropped.So it is called as tail
drop/drop tail/drop policy
Two Techniques:
1.Scheduling discipline --> It determines the order in which packets are transmitted.
2.Drop Policy --> It determines which packets get dropped.
Arriving packets
Free buffers
Next to trans
Queued packets
Fig :3.26 FIFO Queueing
Arriving packets
Drop
Next to transmit
Fig :3.27 Tail drop at a FIFO Queue
It pushes all responsibility for congestion control & resource allocation out to the edges of the
network.
Assume that there is no help from the routers, then TCP takes responsibility for detecting &
responding to congestion.
3.7.1.2 Priority queueing:
A Simple variation on basic FIFO queueing.The idea is to mark each packet with a priority.The
mark could be carried.
It implements multiple FIFO queues, one for each priority class
Three priority classes
Low
Medium
High
The router always transmits packet out of the highest-priority queue if that queue is nonempty
before moving on to the next priority queue.
Problem: The idea is a small departure will win during the transmission.so the high-priority
queue can starve out all the other queues.
Solution: The network could change more to deliver high priority packets than low-priority
queue can starve out all the other queues.
Here, the routing updates that are necessary to stabilize the routing tables after a topology
change.
For that packets, there is a special queue, which can be identified by the differentiated services
code point.
Flow 2
Flow 3
Problem in FQ:
1.The router is not telling the traffic sources anything about the state of the router
2.no way for limiting how quickly a given source sends packets.
It segregates the traffic so that ill-behaved traffic sources do not interfere with those that are
faithfully implementing the end-to-end algorithm.
Problem:
The packets being processed at a router are not necessarily the same length.so we have to
consider the length of the packets
Eg:Router has 2 flows
Flow1 -->1000 bytes packets -->allocate 2/3 of bandwidth
Flow2 -->500 bytes packets of bandwidth 1/3 of bandwidth
3.8 ROUTING
Forwarding table
Routing Table
6
F
1
B
E
1
D
Fig: 3.28 Network represented as a graph
Problem of Routing:
Find the lowest cost path between any two nodes, where the cost of a path equals the sum of
the costs of all the edges that make up the path.
Two Types of Routing Approach:
3.8.2.1.Static approach: load all shortest path into some nonvolatile storage.
Problem:
Does not deal with node/link failure.
Does not consider the addition of new nodes (or) links.
Imples that edge costs cannot change.
3.8.2.2 Dynamic
Approach:
In this approach, the problems of static approach will be resolved. There are 2
main classes of routing protocols, such as,
Distance vector(DVR)
Link state (OSPF)
3.8.2.2.1 DISTANCE VECTOR ROUTING: (DVR)
Each node constructs a one-dimensional array (a vector) containing the distance (costs)
to all others nodes & distributes that vector to its immediate neighbours.
Assumption:
Each node knows the cost of the link to each of its directly connected neighbours.
The costs may be provided when the router is configured by a n/w manager.
A link is down, it is assigned an infinite cast.
Infinite Cost:
Eg) consider the following graph, The cost of each link is set to 1.
B
A
E
F
G
Fig 3.29 Example Graph
We can represent each nodes knowledge about the distances to all other nodes as a table
STEP 1:
TABLE : Initial distances stored at each node
Each node knows only the information in one row of the table
Node is directly connected
cost is 1
cost is
Initially A believes that it can reach B in one hop and that D is unreachable
STEP 2:
Every node sends a message to its directly connected neighbours containing its
personal list of distances
Destination
Cost
NextHop
Like the above table all nodes have its personal list of distances
Send the routing information as < Destination , Cost , Next Hop>
Eg) F tells node A that it can reach node G at a cost of 1. (ie like < G , 1 , G >
ie) A D , A G , cost is from the initial distances ( GLOBAL VIEW )
Now A found that it can reach G via F at the cost of 2 ( A F (cost 1 ) + F G ( cost 1 )
so totally 2 cost )
Likewise A found that it can reach D via C at the cost of 2 ( A C ( cost 1 ) + C D ( cost 1 )
so total cost is 2)
Similarly, it will find all distances ( costs ) for each node.
TABLE : Final routing table at node A
Destination
Cost
NextHop
If any topology changes it takes only a few exchanges of information between neighbours
before each node has a complete routing table
Convergence:
The process of getting consistent routing information to all the nodes is called
convergence
This algorithm enable all nodes to achieve a consistent view of the network in the absence
of any centralized authority
TABLE: Final distances stored at each node ( GLOBAL VIEW )
Now A concludes that it can reach E in 4 hops. Now C concludes that it can reach E
in 5 hops. Likewise it increases the distance but none of the nodes actually knows that E is
unreachable this is called as Count to infinity
Solution:
Use some relatively small number as an approximation of infinity.
Techniques To Stabilize Routing
SPLIT HORIZON:
When a node sends a routing update to its neighbours it doesnot send those routes it
learned from each neighbour back to that neighbour.
Eg) B has the route < E , 2 , A > it is learned from A. so it wants forward to A.
SPLIT HORIZON WITH POISON REVERSE:
Instead of omitting the forwarding to the learned node, here assign a negative
information in the route.
Problem In Both The Techniques :
They only work for routing loops that involve 2 nodes.
B
A
5
2
D
C
3
Routes C advertises to route A the fact that it can reach networks 2 and 3 at cost of 0,
networks 5 and 6 at cost of 1, and networks 4 at cost of 2.
The packet is sent as,
< Address , mask , distance>
RIP send their advertisements every 30 secs.
0
Command
16
Version
Family of net 1
31
Must be zero
Route Tags
Address Prefix of net 1
Mask of net 1
Distance to net 1
First X checks to see if it has already a copy of an LSP. If not stores the LSP.
If it already has a copy it compares the sequence nos.
Is larger than the stored old LSP(It replace the
old LSP by new LSP)
Sequence no of LSP
If the received LSP was the newer one, X then sends a copy of that LSP to all of its neighbours
except the neighbour from which the LSP was just received.
(a)
B
(b)
(c)
(d)
Fig 3.31 Reliable Flooding
C(n)=l(s,n)
While(Nm)
M=mu{w}such that c(w) is the minimum for all w in (N-M)
For each n in (N-M)
C(n)=MIN(c(n),c(w)+c(w,n))
It is called the forward search algorithm. each switch maintains two lists .
Tentative.
Confirmed.
Each of these lists contains a set of entries of the form <destination, cast, next hop>
Algorithm:
Initialize the confirmed lists with an entry for myself; this entry has a cost of zero.
Added node in the previous step, call it node next & select its LSP.
For each Neighbour of Next, calculate the cast to reach this Neighbour as the sum of the
sum of the cast from myself to Next & from Next to Neighbour.
a)Neighbour is on neither Confirmed nor Tentative list then add <Neighbour, cast ,next
hop> to the Tentative list.
b)Neighbour is currently on Tentative list &cast is less than the currently listed cost for
Neighbour then replace the current entry with <Neighbour, cast, next hop>
If the Tentative list is empty, stop otherwise pick the cast, move it to the Confirmed list &
return to step 2
Eg:
step
confirmed
Tentative
1.
(d,0,-)
2.
(D,0,-)
(B,11,B)(C,2,C)
3.
(D,0,-)(C,2,C)
(B,11,B)
4.
(D,0,-)(C,2,C)
(B,5,C)(A,12,C)
5.
(D,0,-)(C,2,C)
(A,12,C)
(A,10,C)
(B,5,C)
6.
Comments
(D,0,-)(C,2,C)
(B,5,C)
It is based on the distributed map concept ie)all nodes have a copy of the network map, which
is regularly updated.
Each node contains a routing directory database. It has information about the routes
interfaces like status of adjacent routes.
It computes the shortest of adjacent routes.
Version
Type
Router ID
Area ID
Packet length
Checksum
Authentication type
Authentication
Data
3.9 Subnetting:Motivation:To find some way to use the networks more efficiently.
Problem:Assigning many networks number has another drawback that becomes apparent when we
think about routing.
Big forwarding tables add costs to routers,they are potentially slower to search than smaller
tables for a given technology,so they degrade router performance.
Solution:Subnetting provides a first step to reduce total no of network number that are assigned.
Idea:-take a single IP networks number & allocate the IP addresses with that network number to
several physical networks.
Steps:-
1)the subnets should be close to each other.ie)having only one network number between
them.ie)a router will be able to select one route to reach any of the subnets.
From outside the campus ,all we need to know to reach any inside the campus is where the
campus connects to the internet.
3.9.1 Subnet mask:A single network number can be shared among mulitiple networks involves configuting all the
nodes on each subnet with a subnet mask. It enables a subnet number all hosts on the same
physical network will have the same subnet number ,which means that hosts may be different
physical networks but share a single network number.
Benefits of subnetting:1)Reduced network traffic.
2)Optimized network performance.
3) Simplified network management.
4)Facilities spanning large geographical distances.
The network administrator creates a 32 bit subnet mask comprised of one &zeros.
1 ->positions representing network or subnet addresses.
0->positions representing the host address.
Class
Format
Net.Node.Node.Node
255.0.0.0
Net.Net.Node.Node
255.255.0.0
Net.Net.Net.node
255.255.255.0
3.9.2 Masking:
It is a process that extracts the address of the physical network from an
IP address is called Masking.
There are two types of masking namely
1. Boundary level masking (either 0
or 255)
2. Non boundary level masking
(<255&>0)
Rules for Boundary Level:
Mask no 255, IP address is repeated
Mask no 0, 0 is repeated
Rules for Non-Boundary Level:
Mask no 255 ,IP address is repeated
Mask no 0, 0 is repeated.
Mask no (>0 & <255), do bitwise AND between mask number and IP address
Eg)
IP address----> 140.11.36.22 (Boundary level)
Mask ----------> 255.255.255.0
-------------------Subnet address
140.11.36.0
---------------------
Mask -------------->255.255.224.0
------------------Subnet address
141.181.0.0
-------------------
224 ---->11100000
14 ----->00001110 (Bitwise AND)
--------------0------>00000000
-----------------
LAN
1
2
10.0.0.1
198.60.42.12
3
4
5
6
ISP router
PC
Leased line
Server
NAT box
The router ANDs the packets destination address with Subnet Mask for each entry. If the result
matches the subnet number of the entry , then this is the right entry to use and it forwards the
packets to the next hop router indicated.
The CIDR aggregates the routes. It tries to balance the desire to minimize the number of routes
that a router needs to know against the need to band out addresses efficiently.
It require a new type of notation to represent network numbers or prefixes as they are known
because the prefixes can be of any length.
Add /X after the address where X is the prefix length in bits.
Eg) 20 bit prefix for all the networks 194.4.16 through 192.4.31 is represented as 194.4.16/20
Advertise
128.112.128/2
ISP
128.112.128/2
44
128.112.129/2
4
128.112.130/2
4
128.112.135/2
4
Fig 3.33 Route Aggregation with CIDR
Solution:
Keep the polices private.
Types of inter domain Routing Protocols:
BGP (Border Gateway protocol)
EGP(Exterior Gateway Protocol)
EGP is designed when the internet had a treelike topology. Since AS are connected only as
parents and children and not as peers.
BGP has no assumption about how AS are interconnected. They form an arbitrary graph. It is
enough to accommodate non-tree structured internetworks. Many internet service providers exist
to provide service to consumers. Interconnects other providers and larger corporations.
Types of Traffic:
Local traffic
Transit traffic
Local traffic: Originates (or) terminates on nodes within an AS
Transit traffic: Passes through an AS.
Types of Autonomous systems:
Stub AS
Multihomed AS
Transit AS
Stub AS : has only a single connection to other AS, it will carry local traffic.
Multihomed AS: has connection to more than one other AS but that refuses to carry transit
traffic
Transit AS : has connections to more than one other AS and is designed to carry both transit and
local traffic.
Goals of Inter Domain Routing:
Necessary to find some path to be intended destination that is loop free
Policy determination of Routing ie) Finding a non-looping, policy-complaint path
Additional Factors:
Scalability
Reachability
Issue of trust
Basics of BGP:
Each AS has one or more border routers through which packets enter and leave the AS. It is
known as Gateways.A border route is simply an IP router that is charged with the task of
forwarding packets between autonomous systems(AS)
Each AS also have atleast one BGP speaker.
BGP speaker duty:
Speak with other BGP speaker in other AS
Find that border routers
BGP advertises Complete paths as an enumerated list of AS to reach a particular network. It is
called as a path-vector protocol. The advertisement enables the sorts of policy decisions and
routing loops to be readily detected.
Large
Corporation
Peering point
Backbone
service provider
Consumer ISP
Small
Corporation
Peering point
Consumer ISP
Large
Corporation
Fig 3.34 A Simple Multi-Provider Internet
Many provides arrange to inter connect with each other at a single peering point.
Jobs of BGP:
To prevent the establishment of the looping paths.
Regional
Provider A (AS
2)
128.96
Customer P (AS
4)
Customer Q (AS 5)
Loop
Backbone n/w
(AS 1)
Customer R (AS
6)
Regional
Provider B (AS
3)
Customer S (AS
7)
Fig :3.35 A network running BGP & Loop among autonomous systems
The extra link (dashed line) between As2 and AS3 ,it has a loop in it. AS1 learns that it can reach
network 128.96 through AS2, so it advertises to AS3 , who in turn advertises it back to AS2. The
advertisement passes to AS3, then As3 is passed it to AS1, As1 is passed to AS2 and so on. The
loop will occur.(problem)
Solution:
BGP carries AS numbers to be unique. We note that a stub AS does not need a unique AS
number, because it will carry load traffic.
Path vector routing table:
Network
Next Router
path
N1
R01
N2
R05
N3
R06
Types of messages:
Open
Update
Keepalive
Notification
Open to create neighbourhood relations IP and opens a TCP connection with a neighbour and
sends open message. Neighbour answers with a keep-alive message.
Update the path vector routing table by using 2 techniques
Local preference
BGP router advertise the local preference value, it will choose lighter
preference.
Multi exit discriminator
Entry point to forward and exit point for advertise. The area send the advertisements with
multi exit attribute go
Keepalive
BGP routers periodically inform each other about their presence. The maximum value is
1 sec
Notification
Send if a router wants to close or if it detects an error
BGP Packet Header:
All BGP packet share the same common header
15
23
MARKER
Marker
Reserved for authentication
Type (8 bits)
31
Length
length of the packet and the size is 16 bits
Type
Specifies the type of the message size is 8 bits ie) update , keepalive , notification, open
message.
3.13 IPV6 (IP Version 6) :IP addresses are assigned not just to computers but also to mobile phones,television and others
household appliances. Since the IP address is carried in the headers of every IP
packet,increasing a need of the size of address dictates a change in the packet header.
Generation/Version of IP :-
3.13.1 SIPP (Simple Internet Protocol Plus) :Doubling of the IP address size to 64bits.
IPV6 :Doubling of the SIPP address to 128 bits.
Address and Routing
IPv6 provides a 128bit address space
It can address 3.4*10^38 nodes
Use
000..............0(128 bits)
Unspecified
000..............1(128 bits)
Loopback
1111 1111
Multicast address
1111 1110 10
Link-Local unicast
Everything else
IP/IPv4
IPv6
Class A,B,C
Class D
Multicast Address
Link-local Address :Use addresses to enable a host to construct an address that will work on the
network to which it is connected without being concerned about global unigueness of the
address.
Conversion of IPv4 to IPv6 :A node may be assigned on IPv4 compatible IPv6 address By ZERO-EXTENDING
a 32bit address (IPv4) to 128bits.It is called as IPv4 mapped IPv6 address
32 bits IPv4 + 2 bytes of + Zero 128 bits of
Address all is extending
IPv6 address
x:x:x:x:x:x:x:x:x
: : FFFF : 128.96 . 33 . 81
IPv6 address
The unicast address allocation plan that determines how unicast address will be
assignes to service providers, autonomous system, network, hosts & routers
_ Assume that nontransit As as a subscriber & transit As as a provider. We may subdivide
providers into direct & indirect.
_ The former are directly connected to subscribers. It may correct other providers are not
connected directly to subscribers, are known as backbone network.
Goal :- To provide aggregation of routing information to reduce the burden on intra domain
routers.
Idea :- Use an address prefix a set of contiguous bits at the most significant end of the address
Way :- Assign an address prefix to a direct provide & then for that direct provider to assign
longer prefixes that begin with that prefix to its subscribers.
Problem :If a site decides to change providers, it will need to obtain a new address prefix &
renumber all nodes in the site.
Solution :Geographic addressing, in which a sites address is a function of its location rather than
the providers to which it attaches.
IPv6 Provider Based Unicast Address :Format :3
010
RegistryID
ProviderID
p
SubscriberID
InterfaceID
no relationshIP with
prefix Y
then Y advertise a
prefix, no relationship
it causes the half of
their site become unreachable
Packet Format :0
Version
12
16
Traffic Class
Payload Len
24
Flowlabel
Next header
31
Hoplimit
Source Address
Destination Address
Next header/data
0
Next
16
Reused
29
Offset
31
RES
Header
Indentification
Next Header :-
5. Hop Limit :Simply the TTL of IPv4 maximum number of hop limits.
6. Source Address & Destination Address :The bulk of the header is taken up with the source & destination addresses, each of
which is 16 bytes (128 bits) long.
7. Next Header :Type of each extension header.
Auto Configuration :Requirements / Needs to Configure :- Valid IP address
- A subnet mask for the link to which it attaches
- Address of a name server.
* IPv6 support Plug-and-Play operation for auto configuration. It needs Stateless Auto
configuration, which needs does not require a server.
* IPv6 uni cast addresses are hierarchical, thus we can subdivide the auto configuration
problem into 2 parts.
1. Obtain an Interface ID ie) unique on the link to which the host is attached.
2. Obtain the correct address prefix for the subnet.
Advanced Routing Capabilities :- IPv6 defines an anycast address ie) assigned to a set of interfaces & packets sent to
that address will go to the nearest of those interfaces, determined by the routing protocols.
- Provide enhanced routing support to mobile hosts.
Other Features :1.Mobility
2.Network Security
3. New service model for the internet
>= 1
Set this bit in packet
Queue Length
Averaging Interval
Current Time
Time
Previous cycle
Fig :3.36 Queue Length
Current cycle
- The router maintain a running average of their queue lengths.Each time a packet
arrives, this alignment computes the average queue length.
- The router discards one (or) more packets before the buffer becomes completely
full.
Average Queue Length
Congestion Level
Action to be taken
Minimal (or)
Non-existent
>Upper Threshold
Between 2 Threshold
Onset of Congestion
Calculate the
Probability
Flow Control
Congestion Control
- Buffering is used
congestion
- Packet is lost between one sender & - Other user packet is also lost
receiver only
- It affects network performance highly
- It affects less on network perforance