Circuit, packet and virtual-circuit switching

Why it exists

Two computers on opposite sides of a country are not joined by one wire. Their data crosses many links and nodes shared with many other users. How the nodes forward the data decides the delay, how many users can share the links, and what happens when a node fails. There are three classical answers.

Prerequisites

You need the idea of layers (layering).

Precise definitions

Circuit switching. Before data flows, a connection set-up reserves a fixed share of every link on the path (a time slot or frequency band, a circuit). The reserved capacity belongs to this connection for the whole call even when nothing is sent. Data then travel without waiting at nodes. Telephone networks work this way.

Packet switching. The message is cut into packets, each carrying its own header; packets share the links and are queued at nodes. Each node receives a whole packet before it starts to forward it (store and forward). Capacity is shared on demand, which is called statistical multiplexing.

Packet switching comes in two forms:

• Datagram network: no set-up. Every packet carries the full destination address and each router chooses the outgoing link for every packet independently from a forwarding table; packets of one message may take different routes and arrive out of order (this is how IP works).
• Virtual-circuit network: a path is chosen and state is set up before data (the connection setup); every packet carries only a short virtual-circuit identifier (VCI) which each switch rewrites on every link; all packets of the call follow the same path in order; the capacity is not reserved exclusively (unless the network also reserves resources). ATM and frame relay worked this way.

The table is quoted from the textbooks.

How it works

Delay of a message over several links

Take $N$ links in a row, each with rate $R$ and propagation delay $d$. A message of $M$ bits is cut into $P$ packets of $L = M/P$ bits (no headers, no processing and no queueing delay other than waiting for the previous packet on the same link).

• One packet ($P = 1$): every node has to receive the whole packet before forwarding it, so the transmission delay is paid on every link: $N \cdot L/R + N d$.
• $P$ packets: the first packet needs $N L/R$ to reach the end; each further packet follows one packet-time behind it, so the last packet arrives after

• Circuit switching: after the set-up time $t_s$ the message is sent as a stream over the reserved path without store and forward: $t_s + M/R + N d$.

A $1$-megabit message over three $1$-Mbps links takes $3$ s as one packet, but only $1.2$ s as $10$ packets, $1.02$ s as $100$ packets and $1.002$ s as $1000$ packets: splitting lets the links work in parallel (a pipeline). In the limit the time tends to $M/R$ (the time on one link) plus the delay of one small packet per extra link. With a set-up time of $0.1$ s circuit switching needs $1.1$ s, so packet switching with enough packets can be faster even for one message. The saving is real only while the header added to each packet is small; with headers of $h$ bits per packet each term grows to $(L + h)/R$ and there is an optimum number of packets.

Sharing a link: statistical multiplexing

A $10$ Mbps link serves users who each send at $1$ Mbps when active. Circuit switching can serve $10$ users (one circuit each). Packet switching admits more users because users are not active all the time:

With $35$ users, each active $10$ percent of the time independently, more than $10$ are active at once with probability only $0.000424$: the link is overloaded rarely, and then packets queue briefly. Packet switching therefore carries $3.5$ times as many users as circuit switching at almost the same quality. Packet switching is better for bursty traffic; circuit switching for steady streams that need guaranteed rate and no delay variation.

Worked examples

Cheat box

Traps and GATE patterns

Connections

The delay terms are defined on the next page (performance metrics). The Internet is a datagram network at the IP layer (IPv4 addressing, routing) on top of which TCP builds a reliable stream; the switches of a LAN forward frames in a similar store and forward way.

PYQs

Consider a file of size 4 million bytes being
transferred between two hosts connected via a path
consisting of three consecutive links of bandwidth
2 Mbps, 500 kbps, and 1 Mbps, respectively.

All processing delays and propagation delays are
negligible. Assume there is no other background
traffic over the path and no additional overhead to
transfer the file.

Which one of the following is the total time

(in seconds) to transfer the file?

Note: 1M = 106, 1k = 103

Two hosts are connected via a packet switch with 107
bits per second links. Each link has a propagation
delay of 20 microseconds. The switch begins
forwarding a packet 35 microseconds after it receives
the same. If 10000 bits of data are to be transmitted
between the two hosts using a packet size of 5000
bits, the time elapsed between the transmission of the
first bit of data and the reception of the last bit of the
data in microseconds is:

COMPUTER NETWORK

Consider the store and forward packet switched
network given below. Assume that the bandwidth of
each link is 10^6 bytes/sec. A user on host A sends a
file of size 10^3 bytes to host B through routers R1
and R2 in three different ways. In the first case, a
single packet containing the complete file is
transmitted from A to B. In the second case, the file
is transmitted from A to B. In the third case, the file
is split into 20 equal parts and these packets are sent
from A to B. Each packet contains 100 bytes of
header information along with the user data. Consider
only transmission time and ignore processing,
queuing and propagation delays. Also assume that
there are no errors during transmission. Let T1, T2,
and T3 be the times taken to transmit the file in the
first, second, and third case respectively. Which one
of the following is CORRECT?

Consider a source computer (S) transmitting a file of
size 106 bits to a destination computer (D) over a
network of two routers (R1 and R2) and three links
(L1, L2, and L3).

COMPUTER NETWORK

L1 connects S to R1; L2 connects R1 to R2; and L3
connects R2 to D. Let each link be of length 100
km.

Assume signals travel over each link at a speed of
108meters per second.

Assume that the link bandwidth on each link is 1
Mbps. Let the file be broken down into 1000
packets, each of size 1000 bits.

Find the total sum of transmission and propagation
delays in transmitting the file from S to D?

Consider a network path P → Q → R between
nodes P and R via router Q. Node P sends a file of
size 106 bytes to R via this path, splitting the file
into chunks one after the other, each of size 103
bytes. Node P sends these chunks without any wait
time between the successive chunk transmissions.
Assume that the size of the extra headers added to
these chunks is negligible, and that the chunk size
is less than the MTU. Each of the links P → Q and
Q → R has a bandwidth of 106 bits/sec, and
negligible
propagation
latency.
Router
Q
immediately transmits every packet it receives
from P to R, with negligible processing and
queuing delays.

Assume P starts transmitting the chunks at time t =
0. Which one of the following options gives the
time (in seconds, rounded off to 3 decimal places)
at which R receives all the chunks of the file?

In a packet switching network, packets are routed
from source to destination along a single path
having two intermediate nodes. If the message size
is 24 bytes and each packet contains a header of 3
bytes, then the optimum packet size is:

Self-check

Revision card

References