Title: EE3900 Data Communications and LANs Packet Switching Slide 1
1Packet Switching
2Outline
- Packet-switching Principles
- Switching Techniques
- Comparison of Circuit Switching and Packet
Switching - Simple Routing Schemes
- Congestion Control
3Circuit Switching
- Circuit switching designed for voice or constant
rate services - Resources are dedicated to a particular call for
the call duration - In data traffic, some of the time, no data are
sent - Data rate is pre-arranged
- Both ends must operate at the same rate
4Packet Switching
- Data transmitted in small packets
- Typically 1k to 2k octets
- Longer messages are split into series of packets
- Each packet contains a data field plus some
control info - Control information
- Routing (addressing) info
- Packets are received, stored briefly (buffered)
and passed on to the next node - Store and forward
5Advantages
- Line efficiency
- Single node to node link can be shared by many
packets on a on-demand basis, improve efficiency
and reliability - Packets are queued and transmitted as fast as
possible - Data rate conversion
- Each station connects to the local node at its
own speed - Nodes buffer data and send to the station
- When the network is busy, packet accepting and
delivery rates may be reduced - Priority services can be implemented
6 Packets
7Use of Packets
8Example of Packet Routing
A Packet Switching Network
9Packet Size
10Outline
- Packet-switching Principles
- Switching Techniques
- Comparison of Circuit Switching and Packet
Switching - Simple Routing Schemes
- Congestion Control
11Switching Technique
- Station breaks long message into packets
- Packets are sent one at a time to the network
- Packets are handled in two ways
- Datagram
- Virtual circuit
12Datagram Mode
- Each packet is treated independently
- Packets can take any route
- Packets may arrive out of order
- Packets may go missing
- Up to the receiver to re-order the packets and to
recover from missing packets - IP uses datagram
13Datagram Mode
14Virtual Circuit Mode
- Preplanned route decided at connection
establishment - Call request and call accept packets establish
the connection (handshake) - Each packet contains a virtual circuit identifier
instead of destination address - No routing decisions required for each packet
- Clear request to terminate circuit
- Not a dedicated path, links are shared by
different packets
15Virtual Circuit Mode
16X.25 Use of Virtual Circuits
17Public Switched Data Networks
- An example of packet switching network
- A network for data communications over a wide
area - Also operated by telephone companies
- Based on X.25 standard from ITU-T, specifying an
interface between a host system and a packet
switching network - low speed, up to 64kb/s
18Virtual Circuits vs Datagram
- Virtual circuits
- Network provides sequencing and error control
- Packets are forwarded more quickly
- No routing decisions to make
- Less reliable
- Loss of a node will lose all circuits through
that node - Datagram
- No call setup phase
- Better for low number of data packets
- More flexible
- Routing can be used to avoid congested parts of
the network
19Outline
- Packet-switching Principles
- Switching Techniques
- Comparison of Circuit Switching and Packet
Switching - Simple Routing Schemes
- Congestion Control
20Delays
- Propagation delay time for a signal to propagate
from one node to the next - Transmission time time for a transmitter to send
out a block of data, transmission capacity
dependent - Node delay the time for a node to process and to
switch the data
21Circuit Switching vs Packet Switching
Event Timing for Circuit Switching and Packet
Switching
22Circuit Switching vs Packet Switching (Contd)
- Circuit switching connection set-up is required
before data transfer - Virtual circuit packet switchingconnection
set-up is needed before data transfer - Datagram packet switching no set-up is required
23(No Transcript)
24Outline
- Packet-switching Principles
- Switching Techniques
- Comparison of Circuit Switching and Packet
Switching - Simple Routing Schemes
- Congestion Control
25 - Introduction
- The need for routing e.g. Getting from A to F
D
B
F
A
E
C
Path 1 A-B-D-F Path 2 A-B-E-F Path 3
A-B-C-E-F Path 4 A-C-E-F Path 5
A-C-E-B-D-F Path 6 A-C-B-D-F Path 7
A-C-B-E-F
Which one?
26Routing
Routing
- Complex, crucial aspect of packet switched
networks - Performance used for selection of route Minimum
hop, Least cost - Characteristics required
- Correctness, Simplicity, Robustness, Stability
- Fairness, Optimality, Efficiency
27Routing Attributes
- Routing information gathering
- Information source local, adjacent nodes, all
nodes - Update time never (fixed routing), regular
update (adaptive routing) - Routing path selection/calculation
- Time per packet, connection set-up
- Place distributed nodes (distributed routing),
centralized node (central routing), source node
(source routing)
28Fixed Routing
- Single permanent route for each
source-destination pair of nodes - Determine routes using a least cost algorithm
- the link cost is based on expected traffic or
capacity, but not on any dynamic variable such as
instant traffic volume - each node needs only to store the next forwarding
address for each destination
29Cost of Routes in a Packet-Switched Network
30Fixed RoutingTables
31Least-Cost Path
- What is meant by least cost?
- It can be number of hops, physical distances,
costs, delay, link capacity, current loading - Or an arbitrary function of the above parameters,
32Flooding
- A packet is sent by the source node to every one
of its neighbors - At each intermediate node, an incoming packet is
retransmitted on all outgoing links except for
the incoming link - Eventually a number of copies will arrive at
destination - Each packet is uniquely numbered so duplicates
can be discarded - Advantage no network information is required
- Disadvantage unlimited amount of traffic will be
generated unless something is done
33Flooding Example
34Solution for Flooding
- Method 1 each node remembers the identity of
those packets it has already transmitted. When
duplicate copies of the packet return, they are
discarded. - Method 2 Each packet contains a hop count field.
Each time a node passes on a packet, it
decrements the count by one. When the count
reaches zero, the packet is discarded.
35Properties of Flooding
- All possible routes between source and
destination are tried. Therefore, a packet will
always get through if there is a connection
between source and destination - Because all routes are tried, at least one copy
of the packet will arrive at the destination
using a minimum-hop route can be used to set up
VC - All nodes that are directly or indirectly
connected to the source node are visited useful
to distribute information (routing)
36Random Routing
- A node selects only one outgoing path for
transmission of an incoming packet - The outgoing link is chosen at random, round
robin or probability based excluding the link on
which the packet arrived - Like flooding, no network information is required
- Random routing carries less traffic than
flooding, but higher than optimum load - Route is typically not least cost nor minimum hop
37Adaptive Routing
- The routing decisions change as conditions on the
network change, such as node failure or link
congestion - Information about the state of the network are
required and exchanged among the nodes - The higher the amount and the more frequent
information is exchanged (higher network
overhead), the better decisions can be made - React too quickly can cause oscillation, too
slowly will not be adaptive
38Advantages of Adaptive Routing
- improve performance, as seen by the network user
- By-pass link congestions and node failures
39Drawbacks of Adaptive Routing
- The routing decision is more complex hence,
increase node processing - Adaptive strategies depend on status information,
collected at one location but used in another
therefore, the traffic loading on the network
increases
40Isolated Adaptive Routing
41Outline
- Packet-switching Principles
- Switching Techniques
- Comparison of Circuit Switching and Packet
Switching - Simple Routing Schemes
- Congestion Control
42Congestion
- As packets arrive, they are stored in the input
buffer of the corresponding link - The node sends each incoming packet to the
appropriate output buffer - If packets arrive too fast for the node to
process, or faster than the outgoing speed
packets will arrive for which no memory is
available and packets will be lost.
43Causes of Congestion
- Insufficient router memory, slow processors (at
routers), low-bandwidth links, etc. - Normal Traffic
- Congestion
Router
incoming links
outgoing links
Router
incoming links
outgoing links
Packets are dropped due to lack of buffers. Can
we solve congestion by using huge amount of
buffers?
44Effects of Congestion
- Throughput As the offered load increases,
throughput first increases, then drop and finally
approaches to zero (Why? Any solution?) - Delay the average delay grows without bound as
the load approaches the capacity of the system.
45Congestion
- Introduction
- The effect of network congestion
During congestion, the more you send, the less
the network can deliver.
46The effect of congestion.
47General Strategies for Congestion Control
- discard any incoming packets for which there is
no available buffer space - exercise flow control over its neighbors (ask
them to reduce sending rate), so that the traffic
flow remains manageable. In fact, the traffic on
the entire network needs to be managed.
48Congestion Control Algorithms
- General Principles
- Open-loop congestion control algorithms
- Prevent congestion from occurring in the first
place. - Close-loop congestion control algorithms
- Monitor state of the network
- Feedback state information to the senders
- Adjust system operation to correct the problem
- Specific algorithms will be discussed in TCP