PadhaikoAll subjects

Intro switched communication network, connection oriented and…

Data Communication (ENCT 253)

Open in the Padhaiko app

5.6 Intro Switched Communication Network, Connection Oriented and Connectionless

When a communication system scales beyond a handful of nearby devices, connecting every device directly to every other device via dedicated point-to-point wires (unswitched full-mesh network) becomes impractical. For stations, a full mesh requires:

Moreover, most links sit idle most of the time, and adding a single new user requires running new cables. Switching solves this scalability bottleneck by replacing pairwise point-to-point wires with a shared cloud of interconnected switching nodes.

Definition

Switched Communication Network, Stations, and Nodes

UNSWITCHED MESH (N = 6 -> 15 links):        SWITCHED NETWORK (Stations A, B, PCs, Server):
     User 1 ------- User 2                               [Mainframe]
     / |  \       /  |  \                                     |
User 6-+---\-----/---+---+-User 3            [PC]---(Node 1)-(Node 2)-(Node 3)---[Server]
     \ |    \   /    |  /                              |        |        |
     User 5 ------- User 4                             |     (Node 5)    |
                                                       |    /   |    \   |
                                           Station A-(Node 4)-(Node 7)-(Node 6)-Station B

Taxonomy of Switching Methods

Classical switched networks are classified into three major paradigms according to how links and node buffers are allocated:

                                 +---> Circuit Switching (Dedicated Physical Path)
                                 |
                                 |                        +---> Virtual-Circuit Approach
              Switching ---------+---> Packet Switching --+
                                 |                        +---> Datagram Approach
                                 |
                                 +---> Message Switching (Store-and-Forward Whole Message)

Among these three, circuit switching and packet switching form the foundation of modern telecommunications and the Internet, while message switching survives primarily in application-layer store-and-forward services (such as electronic mail and delay-tolerant messaging).

Connection-Oriented vs. Connectionless Services

Network communication services—both at the switching layer and across the OSI/TCP-IP protocol stack—operate in one of two fundamental modes:

Definition

Connection-Oriented and Connectionless Communication

  1. Connection-Oriented Service: Modeled after the telephone system. Before any user data is exchanged, the source and destination establish a logical or physical association (connection) across the network. Communication proceeds strictly through three phases:
    • Connection Establishment (Setup): End-to-end path or state tables are negotiated and initialized.
    • Data Transfer: Data units travel in order along the established path using a short connection identifier.
    • Connection Release (Teardown): Allocated switch state and channel resources are released.
  2. Connectionless Service: Modeled after the postal system. No prior setup or teardown handshake occurs. Each data unit (datagram or message) is completely self-contained, carries the full destination address, and is routed independently at every intermediate node.
AspectConnection-Oriented ServiceConnectionless Service
Setup / Teardown PhasesMandatory connection setup before data transfer and teardown after completion.None; data is transmitted immediately whenever ready.
Addressing in Data UnitsFull address needed only during setup; data packets carry a short Virtual Circuit Identifier (VCI) (or none in circuit switching).Every single packet must carry the complete source and destination address.
Routing DecisionMade once during the setup phase; all subsequent packets follow the same path.Made independently for every packet at every intermediate router/switch.
Packet Arrival OrderGuaranteed in-order delivery along the fixed route.Packets may travel via different paths and arrive out of order.
Reliability & OverheadHigh reliability with optional flow/error control; setup delay adds overhead for short transfers.Best-effort delivery; fast for short, bursty transactions; robust to single-node failure via rerouting.
Representative ExamplesCircuit Switching (PSTN), Virtual-Circuit Packet Switching, TCP.Datagram Packet Switching (IPv4/IPv6), Message Switching, UDP.

5.7 Switching Devices: Types, Importance and Application

Definition

Switch and Its Role in Data Communication

A switch is a multi-port hardware (and software) networking device that connects multiple transmission links and selectively forwards incoming signals, frames, or packets from any input port to one or more specified output ports by establishing temporary internal paths.

Importance of Switching Devices

  1. Elimination of Wiring Complexity: Reduces cable count from mesh links to station-to-switch links.
  2. Collision Domain Isolation and Full-Duplex Bandwidth: Unlike passive hubs or shared coaxial buses where only one station can transmit at a time, a switch buffers frames and allows multiple station pairs to communicate simultaneously at full link speed.
  3. Dynamic Resource Sharing: Multiplexes many bursty stations onto high-capacity inter-switch trunks.
  4. Intelligent Path Selection and Fault Tolerance: Routes traffic around congested or failed links in mesh-connected WANs.

Types of Switching Devices and Applications

Switching hardware is categorized by both internal switching fabric mechanism (physical layer) and OSI protocol layer:

Classification of Switching Devices, Mechanisms, and Applications
Switch TypeOSI LayerSwitching MechanismPrimary Applications
Space-Division Switch (Crossbar / Multistage)Layer 1 (Physical)Connects inputs to outputs via a metallic or semiconductor grid of crosspoints. A single-stage crossbar needs crosspoints; a 3-stage Clos multistage switch reduces crosspoints dramatically.Traditional analog/digital telephone exchanges (PSTN), optical cross-connects (OXC).
Time-Division Switch (Time-Slot Interchange, TSI)Layer 1 (Physical)Uses high-speed RAM data store and a control store to reorder TDM time slots between incoming and outgoing synchronous TDM frames ( input slot output slot).Digital central office telephone switches (ESS), Private Branch Exchanges (PBX).
Data-Link Switch (Layer-2 Ethernet Switch / Bridge)Layer 2 (Data Link)Reads the MAC address in each frame header and consults an internal MAC address table (CAM table) to forward the frame only to the target port (store-and-forward or cut-through).Enterprise Local Area Networks (LANs), campus networks, data-center server racks.
Packet Switch / Router (Layer-3 Switch)Layer 3 (Network)Inspects the logical IP address in each packet header, performs longest-prefix matching against a routing table, decrements TTL, and forwards the packet to the next-hop interface.Internet backbone cores, Wide Area Networks (WANs), inter-VLAN enterprise routing.

5.8 Circuit Switching Network: Circuit Switching Concepts, Message Switching

Circuit Switching Network and Its Three Phases

A circuit-switched network consists of a set of physical switches connected by trunk links (where each trunk link is divided into channels via FDM or TDM). It was originally designed for real-time voice telephony in the Public Switched Telephone Network (PSTN) and Private Branch Exchanges (PBX), though it can also carry digital data.

Definition

Circuit Switching and Three-Phase Operation

In circuit switching, a dedicated physical communication path (a sequence of reserved FDM frequency bands or TDM time slots connected through intermediate switches) is established between the source and destination stations for the entire duration of a call. Every circuit-switched communication passes through three strict phases:

  1. Setup Phase (Circuit Establishment): Before any user signal can be transmitted, an end-to-end circuit must be reserved. In the switched network of Section 5.6, suppose Station A requests a connection to Station B via Node 4:
    • Station A sends a connection request to Node 4; the link is dedicated.
    • Based on routing tables, link availability, and cost, Node 4 selects Node 5 and allocates a free channel on trunk .
    • Node 5 similarly selects Node 6, allocating a channel on trunk , and Node 6 connects to Station B.
    • A test is performed to check whether Station B is busy or ready; if ready, a call-accept signal returns to A, completing the dedicated physical route .
  2. Data Transfer Phase: Once the dedicated path is established, analog or digital signals flow continuously from A to B at a guaranteed fixed data rate. Because the channel is pre-reserved and no per-packet routing decisions or buffering occur at intermediate nodes, there is zero queuing delay—only electromagnetic propagation delay across the links.
  3. Teardown (Terminate) Phase: After data transfer concludes, one station initiates termination. Disconnect signals propagate through Nodes and to deallocate the reserved channels and make those trunk resources available for other users.
Circuit-Switched Public Telephone Network (PSTN) Architecture:
 [Analog Phone]           [End Office]      [Long-Distance]     [Long-Distance]     [End Office]
       |  Subscriber Loop   |     |            [Office]            [Office]           |     |
       +------------------->|  o--+---------------+                   |               |     |
                            |     | Connecting    +------>o   o-------+               |     |
                            |     | Trunk                 |   |  Intercity Trunk      |     |
                            +-----+                       |   +---------------------->+--o  |
                                                          +------------------+  +-----------+
                                                                             |  |       \
                                                                             +--+    [Digital PBX]

Message Switching

Message switching was the precursor to packet switching (used in mid-20th-century telegraph/teletype networks, AUTODIN, and modern store-and-forward electronic mail).

Definition

Message Switching (Store-and-Forward)

In message switching, no dedicated physical path is established in advance. Instead, the sender appends a destination address header to the entire message and transmits it as a single indivisible unit:

  1. Each intermediate switching node receives the entire message bit-by-bit, stores it completely in secondary disk or RAM buffer memory, and checks it for transmission errors.
  2. When the appropriate outgoing trunk becomes free, the switch forwards the complete message to the next node (Store-and-Forward).
Message Switching (Entire Message Stored Before Forwarding):
  [Sender] ===[Full Msg M]===> (Switch 1) ===[Full Msg M]===> (Switch 2) ===[Full Msg M]===> [Receiver]
                               [Disk/RAM]                     [Disk/RAM]
                              (Store Whole                   (Store Whole
                               Msg, Queue,                    Msg, Queue,
                                Forward)                       Forward)

5.9 Packet Switching: Virtual Switching, Datagram Switching

To overcome the channel wastage of circuit switching and the long store-and-forward delay of message switching, packet switching breaks a long message into small blocks called packets (typically hundreds to a few thousand bytes). Each packet carries user data plus a header containing source/destination address information and control bits.

At each intermediate node, packets are received, stored briefly in high-speed RAM, and forwarded to the next node (packet store-and-forward). Because packets are small, nodes operate as a pipeline: while Node 1 is receiving Packet 2 from the sender, Node 1 is already transmitting Packet 1 to Node 2.

Advantages of Packet Switching over Circuit Switching

  1. Superior Line Efficiency: No link capacity is pre-reserved; a single node-to-node link is dynamically shared over time by packets from many concurrent sessions.
  2. Data-Rate Conversion: Two stations operating at different bit rates can communicate seamlessly because each station connects to its local switching node at its own native bit rate while buffers absorb the rate difference.
  3. Non-Blocking Under Heavy Traffic: In circuit switching, new calls are blocked (busy tone) when all trunk channels are occupied. In packet switching, packets are still accepted into queues—delivery latency increases gracefully instead of rejecting connections.
  4. Priority Differentiation (QoS): Switches can inspect packet priority fields and transmit high-priority real-time packets ahead of lower-priority bulk packets waiting in the queue.

Role of IP Addresses in Packet Routing

In Internet Protocol (IP) packet-switched networks, every host network interface is assigned a globally or locally unique logical IP address (a IPv4 address written in dotted-decimal notation such as 192.168.10.25, or a IPv6 address). An IP address is hierarchically partitioned into:

When a packet arrives at an intermediate packet switch (router), the router extracts the destination IP address from the packet header, performs a longest-prefix match against its routing table (which maps destination network prefixes to outgoing interfaces and next-hop router addresses), and forwards the packet along the optimal path toward the destination network.

Datagram Packet Switching

Definition

Datagram Packet-Switched Network

In datagram packet switching (used at the OSI Network Layer, e.g., IPv4/IPv6):

Datagram Packet Switching (Packets 1, 2, 3, 4 Take Different Routes and Arrive Out of Order):
                                      (Node B)
                                    1/        \1
  [Sender] ===[4|3|2|1]===> (Node A)----3------>(Node D) ===[3|1]===> (Node E)
                            4,2|               3/    \                 | [3]
                               v      4       /       \                v [1]
                            (Node C)------>(Node F)    \               |
                                 \            \4        \              |
                                  \-----2----->(Node G) ===[2|4]===> (Node H) ===> [Receiver]
                                                                                Arrives: [2|4|3|1]

Virtual-Circuit Packet Switching

Definition

Virtual-Circuit (VC) Packet-Switched Network

In virtual-circuit packet switching (commonly implemented at the Data Link Layer or MPLS/WAN layer):

Virtual-Circuit Packet Switching (All Packets Follow the Same Preplanned Route and Arrive in Order):
                                      (Node B)          (Node E)
                                     /        \        /        \ [4|3|2|1]
  [Sender] ===[4|3|2|1]===> (Node A). . . . . .(Node D). . . . . .(Node H) ===[4|3|2|1]===> [Receiver]
                                    \  [4|3]  /        \        /
                                     \  [2|1]/          \      /
                                      (Node C)          (Node G)
                       Legend:  . . . = Preplanned Virtual Circuit Path (A -> C -> D -> E -> H)

Advantages of Datagram vs. Virtual-Circuit Approach

Effect of Packet Size on Transmission Time (Pipelining Analysis)

Breaking a message into multiple packets reduces end-to-end transmission time through store-and-forward pipelining, up to a point where per-packet header overhead begins to dominate.

Consider a virtual-circuit path (: link , link , and link ). Suppose a message of (bytes) is to be transmitted, and every packet requires a fixed control header:

Pipelining Across 3 Hops (X -> a -> b -> Y):
  (a) 1-Packet Message (43 octets/pkt):   (b) 2-Packet Message (23 octets/pkt):   (c) 5-Packet Message (11 octets/pkt):
  Time    X->a   a->b   b->Y              Time    X->a   a->b   b->Y              Time    X->a   a->b   b->Y
   |     [Pkt1]                            |     [Pkt1]                            |     [P1]
   |     (43)                              |     [Pkt2] [Pkt1]                     |     [P2]   [P1]
   |            [Pkt1]                     |            [Pkt2] [Pkt1]              |     [P3]   [P2]   [P1]
   |            (43)                       |                   [Pkt2]              |     [P4]   [P3]   [P2]
   |                   [Pkt1]              v  Total = 4 x 23 = 92 octet-times      |     [P5]   [P4]   [P3]
   v                   (43)                                                        |            [P5]   [P4]
  Total = 3 x 43 = 129 octet-times                                                 v                   [P5]
                                                                                  Total = 7 x 11 = 77 octet-times

In general, if a message of is divided into equal packets across , each packet has payload and total size . The first packet takes octet-times to reach the destination, and the remaining packets arrive one after another every octet-times:

Effect of Packet Count P on Total Transmission Time (M = 40 Octets, h = 3 Octets, K = 3 Hops)
Number of Packets ()Data per Packet ()Header ()Packet Size ()Pipelined Slots ()Total Time

Notice that splitting the message from packet to packets reduces transmission time from to thanks to pipelining, but splitting further into tiny packets increases total time to because the fixed header is repeated times ( of header overhead plus extra per-packet node processing and queuing delays).

Delay Comparison: Circuit, Virtual-Circuit, and Datagram Switching

Three distinct delay components govern the performance of any switched network:

  1. Propagation Delay (): Time taken by an electromagnetic signal to travel along a physical link from one node to the next.
  2. Transmission Time (): Time required for a transmitter to push a block of onto a link of bit rate (e.g., on a line takes ).
  3. Node Delay (): Time spent inside each intermediate switch processing headers and waiting in output queues.
Delay Timing Diagrams Across Nodes 1 -> 2 -> 3 -> 4:
(a) Circuit Switching:            (b) Virtual-Circuit Packet:       (c) Datagram Packet Switching:
 N1    N2    N3    N4              N1    N2    N3    N4              N1    N2    N3    N4
 |\Call|     |     |               |\Call|     |     |               |\Pkt1|     |     |
 | \Req|     |     |               | \Req|     |     |               |\Pkt2|\Pkt1|     |
 |     |\    |     | (Proc Delay)  |     |\    |     |               |\Pkt3|\Pkt2|\Pkt1|
 |     | \   |\    |               |     | \   |\    |               |     |\Pkt3|\Pkt2|\Pkt1
 |     |  Accept/  |               |     |  Accept/  |               |     |     |\Pkt3|\Pkt2
 |    /Call  |     |               |    /      |     |               |     |     |     |\Pkt3
 |\User Data |     |               |\Pkt1|     |     |               |     |     |     |
 | \Continuous Block\              |\Pkt2|\Pkt1|     |              (No Call Setup Phase!)
 |  \   No Node Delay\             |\Pkt3|\Pkt2|\Pkt1|
Comprehensive Comparison: Circuit Switching, Datagram Packet Switching, and Virtual-Circuit Packet Switching
CriterionCircuit SwitchingDatagram Packet SwitchingVirtual-Circuit Packet Switching
1. Call SetupCircuit setup is required before data transfer.Circuit setup is not needed.Virtual-circuit setup is required before data transfer.
2. Transmission PathDedicated physical transmission path.No dedicated path (independent per packet).No dedicated physical capacity, but fixed logical path.
3. Primary OSI LayerGenerally implemented at the Physical Layer.Implemented at the Network Layer (e.g., IP).Implemented at the Data Link Layer (or MPLS).
4. Route DeterminationPath is established once for the entire conversation.Route is established independently for each packet.Route is established once for the entire conversation.
5. Dominant DelaysMajor delay is call setup delay; negligible node delay during transfer.Delay involved is packet transmission and per-node routing/queuing delay.Delays include both call setup delay and per-packet node queuing/transmission delay.
6. Storage at SwitchesMessages are not stored (continuous bit/signal stream).Packets are queued/stored in RAM until forwarded.Packets are queued/stored in RAM until forwarded.
7. Overhead BitsNo overhead bits added after call setup is complete.Full address and control overhead bits added in every packet.Short VCI and control overhead bits added in every packet.

Worked Out Examples

Example 5.1

5.1

Suppose a virtual circuit is established from station to station through two intermediate switching nodes and (forming the path ). A message of (bytes) is to be sent, and each packet carries of control header information. Calculate and compare the total transmission time (in octet-times) when the message is transmitted as:

Explain why dividing a message into smaller packets cannot reduce transmission time indefinitely.

Case 1 — Message sent as : The single packet carries all of data plus of header, giving packet size . Node must receive all from before forwarding to , and must receive all before forwarding to :

Case 2 — Message broken into : Each packet carries of data plus of header, so . As soon as node finishes receiving the packet (), node forwards the packet to while simultaneously sends the packet to :

Cases 3 and 4 — Message broken into and :

  • For : Each packet has , and requires pipelined intervals:
  • For : Each packet has , and requires pipelined intervals:

Explaining why excessively small packets increase total delay: Each packet carries a fixed header. When increases from to , total header bytes sent by each node double from to (nearly matching the message itself!), and each switch also incurs additional per-packet header processing and queuing delay.

Answer:
  1. : .
  2. : .
  3. : .
  4. : (cumulative header overhead and per-packet node processing cause delay to rise when packets become too small).
Example 5.2

5.2

Station A needs to send a file to Station B across a path of (). Each link has a transmission rate of and a propagation delay of per hop.

  1. Calculate the total end-to-end time under circuit switching if the call setup time is .
  2. Calculate the total end-to-end time under datagram packet switching if the file is split into , each carrying plus a , and each intermediate switch introduces a processing/queuing delay of .

1. End-to-end delay under Circuit Switching: In circuit switching, once the setup phase completes, the flow continuously across the links with zero intermediate buffering delay:

2. End-to-end delay under Datagram Packet Switching: Each of the packets has total length , so the transmission time of one packet on one link is:

Across hops with intermediate switches (each adding ):

Answer:
  1. Circuit switching end-to-end time .
  2. Datagram packet switching end-to-end time .
Example 5.3

5.3

A corporate campus has computers that need to communicate with one another.

  1. How many physical point-to-point cables are needed in an unswitched full-mesh network?
  2. How many station-to-switch cables are needed if all computers connect to a single central switch, and what is the percentage reduction in cabling?

Computing point-to-point links in an unswitched full mesh: For stations, every station connects to the other stations:

Computing links with a central switch and percentage cable reduction: With a central switch, each of the stations requires only link to its switch port, so . The percentage reduction in cabling is:

Answer:
  1. Unswitched full mesh requires .
  2. Switched star topology requires , a reduction in cabling.

Exercises

Exercises — Switched Networks, Circuit, Message, and Packet Switching

Question 1

Differentiate between circuit switching and packet switching with suitable diagrams. Explain the three operational phases of a circuit-switched network. [2072 Chaitra, 2074 Ashwin]

Three operational phases of circuit switching: A circuit-switched network (such as the PSTN) establishes a dedicated physical channel (via FDM or TDM) across intermediate switches before information transfer:

  1. Setup Phase: The caller sends a connection request through intermediate switches; each switch allocates a dedicated channel to the next hop until the receiver acknowledges with a call-accept signal.
  2. Data Transfer Phase: Signals flow continuously along the reserved end-to-end circuit at a guaranteed fixed bit rate with no intermediate buffering or queuing delay.
  3. Teardown Phase: When the conversation ends, a disconnect signal propagates along the path to release the reserved channels.

Key differences between circuit switching and packet switching:

  • Resource Allocation: Circuit switching reserves dedicated physical bandwidth for the entire call (wasting capacity during silence); packet switching allocates link bandwidth dynamically on demand using store-and-forward packet queues.
  • Path & Ordering: Circuit switching sends a continuous bit stream over a single physical path in exact order; packet switching splits data into addressed packets that may follow virtual circuits or independent datagram paths.
  • Delay Profile: Circuit switching incurs an initial call setup delay followed by constant propagation delay; packet switching incurs variable per-node queuing and transmission delays.
  • Congestion Behavior: Circuit switching blocks new calls when trunks are full; packet switching accepts packets into buffers and experiences increased queuing delay.
Answer:

Circuit switching operates in three phases (Setup, Data Transfer, and Teardown) over a dedicated physical channel at the physical layer, guaranteeing fixed data rate and zero queuing delay at the cost of idle capacity wastage. Packet switching segments messages into addressed packets forwarded on demand via store-and-forward buffers, maximizing trunk efficiency for bursty computer data.

Question 2

What is Data Switching? Clarify the differences between datagram packet switching and virtual-circuit packet switching. [2073 Shrawan, 2075 Ashwin]

Defining Data Switching: Data switching is the mechanism by which intermediate switching nodes in a communication network dynamically route or forward data units (circuits, messages, or packets) from an incoming link to an appropriate outgoing link so that many stations can communicate across a shared network topology without full-mesh point-to-point wiring.

Comparing Datagram Packet Switching and Virtual-Circuit Packet Switching:

  1. Connection Setup: Datagram switching is connectionless (no setup or teardown phase); virtual-circuit switching is connection-oriented (establishes a preplanned route during a setup phase before data transfer).
  2. Addressing: Every datagram carries the full global destination address; virtual-circuit packets carry only a short local Virtual Circuit Identifier (VCI).
  3. Routing and Path: In datagram switching, every router makes an independent routing decision for each packet, so packets of the same message may follow different paths; in virtual-circuit switching, routing is decided once at setup and all packets follow the exact same path.
  4. Packet Ordering: Datagrams may arrive out of order at the destination; virtual-circuit packets always arrive in order.
  5. Fault Resilience: Datagram networks can automatically route around a failed node mid-transfer; in virtual-circuit networks, a node failure terminates all virtual circuits passing through that node.
Answer:

Data switching forwards data across shared intermediate nodes. Datagram switching is connectionless, routes each packet independently using full destination addresses, and can deliver packets out of order. Virtual-circuit switching is connection-oriented, establishes a preplanned route during setup, forwards all packets in sequence using a short VCI label, and shares link capacity via node buffers.

Question 3

Compare connection-oriented and connectionless services in switched communication networks. Briefly explain how IP addresses are used by packet switches (routers) for routing.

Connection-oriented vs. connectionless services:

  • Connection-Oriented Service: Requires three phases (connection establishment, data transfer, connection release). State is stored at intermediate switches (or end hosts), data units arrive in sequence, and short connection identifiers replace full addresses during data transfer (e.g., Circuit Switching, Virtual-Circuit Switching, TCP).
  • Connectionless Service: Transmits self-contained data units immediately without handshaking. Each unit carries the full destination address and is routed independently, offering zero setup latency and automatic rerouting around failures, but without built-in ordering guarantees (e.g., Datagram Switching, IP, UDP).

Role of IP addresses in packet routing: An IP address (32-bit IPv4 or 128-bit IPv6) logically identifies both the destination network prefix (NetID) and the host interface (HostID). When a packet arrives at an IP router:

  1. The router inspects the destination IP address in the packet header.
  2. It performs a longest-prefix match against its routing table to find the matching network prefix.
  3. The routing table entry specifies the outgoing interface and next-hop router IP address to which the packet is forwarded until it reaches the destination subnet.
Answer:

Connection-oriented service sets up a preplanned logical/physical path before sending data in order, whereas connectionless service routes each self-addressed packet independently with no setup phase. Routers use the network prefix of the packet's destination IP address to look up the next-hop interface in their routing table.

Question 4

A message of is to be transmitted across a virtual-circuit path (). Each packet requires a control header. Assuming negligible propagation and node processing delays, calculate the total transmission time (in octet-times) when the message is sent as: (i) , (ii) , and (iii) .

Setting up the pipelined store-and-forward equation for : With message size , header size , and , dividing the message into equal packets gives packet size and requires pipelined packet-transmission intervals:

Evaluating for and packets:

  • (i) For : :
  • (ii) For : :
  • (iii) For : :
Answer:

(i) : ; (ii) : ; (iii) : .

Question 5

Compare the end-to-end delay of transferring a message across a path where each link operates at with propagation delay per hop:

  1. Using circuit switching with a call setup time of .
  2. Using store-and-forward message switching with zero setup time and negligible queuing delay.

1. Calculating end-to-end delay for Circuit Switching: The transmission time of the message on the dedicated circuit is , plus setup time and link propagation delays of :

2. Calculating end-to-end delay for Message Switching: In message switching, each intermediate switch must receive the entire message ( per hop) before forwarding it to the next link, so the transmission time is incurred times in series:

Answer:
  1. Circuit switching end-to-end delay .
  2. Message switching end-to-end delay (more than slower because the entire message is stored and forwarded without packet pipelining).
Question 6

What are the main types of switching devices and their applications? If an office has stations that need to communicate with one another, compare the number of physical links required in an unswitched full-mesh topology versus a single central switch topology, and find the percentage reduction in links.

Summarizing types of switching devices and applications:

  1. Space-Division Switch (Crossbar / Multistage): Connects inputs to outputs via physical crosspoints (used in PSTN and optical cross-connects).
  2. Time-Division Switch (Time-Slot Interchange, TSI): Swaps TDM time slots in RAM (used in digital telephone exchanges and PBXs).
  3. Layer-2 Ethernet Switch: Forwards frames based on 48-bit MAC addresses using a CAM table (used in enterprise LANs).
  4. Layer-3 Packet Switch (Router): Routes packets across networks based on logical IP addresses (used in WANs and the Internet core).

Comparing link counts for stations:

  • Unswitched full mesh:
  • Central switch star topology:
  • Percentage reduction in links:
Answer:

Full-mesh topology requires , whereas a single central switch requires —a reduction in physical links.

Past exam questions for Data Communication

← Access introduction to multiplexing, application of multiplexing ·…Cellular Wireless Communications and Latest Trends →