Today we discussed Datagram packet-switching.
Packet switching networks provide for the interconnection
of source to destinations on a dynamic basis.
Resources are typically allocated onto an information flow only when needed and then
therefore are shared among many users resulting in high efficiency and low cost.
There are two fundamental approaches to
transferring information over a packet switching network.
The connection oriented mode involves setting up
a virtual circuit connection across the network before information can be transferred.
The connection in this mode does not involve setting up connection.
Instead, packets are routed independently from
node to node until the packet arrive at its destination.
Both approaches involve the use of packet switches to direct packets across the network.
This lesson focuses on packet switching that has
its origin in messages switching for telegraphs.
A message switch typically operates in store and forward fashion,
whereby a message has to be completely received
by the switch before it can be forwarded to the next switch.
Here we will recall that the overall delay has two basic components,
transmission delay and propagation delay.
Because the transmission lines are shared,
the message may also have to wait until previously queued messages are transmitted,
which is called queueing delay.
The figure shows the minimal delay that incurs when a message is
transmitted over a pause that involves two intermediate switches.
We assume there is propagation time of tau by each link.
In total there are 3 taus as the transmission takes total three hops.
Let T be the transmission time of the message in each link.
By store and forward,
there are total 3 Ts.
It follows, the minimal end to end a message delay is 3 taus plus 3 Ts.
Note that this delay doesn't consider any queueing delay and any error checking delay.
Long messages switching not only suffer long delay but also is more vulnerable to errors.
Suppose we want to transmit a large message of 1 megabits over two hops.
So bit error rate at each hop is 10 powers negative 6.
The question is how many bits needed to
be transmitted to deliver the messages successfully.
By messages switching it sends 1 megabit message at once.
The probability of transmission success at each hop is about 33 percent.
So on average it takes about 3 transmissions per hop and
the total number of bits transmitted over two hops is approximately 6 megabits.
The second approach partitioned the 1 megabit message into 10 smaller packets,
each packet at 100 kilobits.
The probability of each packet arriving correctly per hop is about 90 percent.
So on average it takes about 1.1 transmission per hop to succeed.
Therefore, by the second approach the total number of
bits transmitted over two hops is only 2.2 megabits.
In packet switching a long message is broken into smaller packets,
source and destination addresses in the packet header.
It's connectionless, that packets are routed
independently but a packet may arrive at the destination out of order.
Pipelining all packets across network can reduce delay and increase throughput.
Lower delay is suitable for interactive traffic.
This figure shows the minimal delay that is incurred by
transmitting a message that is broken into three packets.
Here we assume that the three packets follow the same path.
For fair comparison with the messages switching we also assume no queuing delay,
no error checking in packet switching.
As in messages switching,
there is propagation delay of tau by each link.
In total, there are 3 taus because of the 3 hops,
and each packet is a one third the size of the message.
The transmission time is one third of T.
The three packets are transmitted successfully from
the source to the first packet switch and at
the first switch the packet 1 is stored and then forwarded to switch 2,
while packet 2 is being transmitted from the source to switch 2.
This packet pipelining reduce total delay of message delivery to 3 taus plus 3/5
T. One can further derive the total delay for k-packet message over L hops.
While packet switching cannot reduce the total propagation delay,
it can reduce the packet transmission delay significantly by packet pipelining.
One may also say that the smaller the packet size is,
the smaller the overall delay is because of finer granularity or packet pipeline.
However, in practice there is a packet header overhead and
all the overheads that put a certain constraint on the minimal packet size.