Module 1 ACN

Q1. Building a Network

A computer network is a system where many devices such as computers, laptops, servers,
switches, and routers are connected together so that they can share data and resources.

Steps to Build a Network:

1. Understand the requirements – First, the purpose of the network is studied.
Example: A school network may need Internet access, file sharing, and printers.
2. Choose the topology – Topology means the structure or design of the
network. Common types are:
• Star (all devices connect to one switch).
• Bus (all devices share one line).
• Mesh (devices are fully connected).
• Ring (devices form a circular connection).
3. Select the transmission medium – The physical path of communication is
selected:
• Wired (twisted pair cable, coaxial cable, optical fiber).
• Wireless (Wi-Fi, Bluetooth, radio waves).
4. Use network devices – Devices like routers (connect networks), switches
(connect computers), and firewalls (security) are required.
5. Apply communication protocols – Protocols are rules for communication.
Example: TCP/IP is the most widely used protocol.
6. Provide security – Security features like passwords, firewalls, and
encryption are needed to prevent unauthorized access.

Diagram:
PC <--- Switch <--- Router <--- Internet

Conclusion:

Building a network involves planning, choosing devices, transmission medium, and protocols.
A properly designed network allows smooth, secure, and reliable communication among users.

Q2. Requirements of a Network

A good network must satisfy some important requirements. These requirements are like goals of
network design.

Main Requirements:
1. Interconnectivity:
• The network must allow multiple devices (computers, printers, servers) to connect
and communicate with each other.
2. Scalability:
• The network should be designed to accommodate growth, allowing for the
addition of more devices without significant changes to the existing infrastructure.
3. Reliability:
• The network must provide consistent performance and availability, ensuring that
devices can communicate without frequent interruptions or failures.
4. Performance:
• The network should support the required data transfer rates and response times
for applications being used.
5. Security:
• Measures must be in place to protect data integrity and confidentiality, preventing
unauthorized access and ensuring secure communication.
6. Cost-Effectiveness:
• The design and implementation of the network should be within budget
constraints while meeting the necessary performance and reliability standards.
7. Ease of Management:
• The network should be easy to manage and maintain, with tools and protocols in
place for monitoring performance and troubleshooting Issues.
8. Compatibility:
• The network should support a variety of devices and protocols, ensuring
interoperability among different hardware and software components.
9. Flexibility:
• The network design should allow for changes and upgrades as technology evolves
 or as organizational needs change.

Q3. Perspectives of a Network

Got it Aifa Let me expand Q3 – Perspectives of a Network into a detailed, easy-to-

understand, exam-style answer (so you can directly write it for 10 marks).

Q3. Perspectives of a Network

When we design or study a network, we must understand that different people look at a
network from different viewpoints. Each type of person has their own needs, and the network
must be designed to satisfy all of them. These viewpoints are called perspectives.

1. End-User Perspective
• End users are the people who actually use the network services like
browsing websites, sending emails, video calling, or downloading files.
• For them, the network is like a “black box” — they do not care how it works
inside, they only want it to work properly.
• Their main needs are:
• Reliability – The connection should not break often. Example: video call
should not drop again and again.
• Ease of use – Plug and play devices (Wi-Fi should connect easily
without complex setup).
• Performance – Fast access to websites, smooth video streaming, no
unnecessary delay.
• Security – Protection from hackers or viruses when using online
services.

Example: A student using Wi-Fi in a college library wants good speed,

reliability, and no complicated settings.

2. Application Developer Perspective

• Application developers create apps like WhatsApp, Zoom, or YouTube.
• For them, the network should provide clear rules and services so they can
design apps without worrying about the low-level details.
• Their main needs are:
• Standard APIs and protocols – Example: sockets in TCP/IP to send/
 • Standard APIs and protocols – Example: sockets in TCP/IP to send/
receive data.
• Transport guarantees – Some apps need reliable delivery (TCP), others
need fast but not guaranteed delivery (UDP).
• Support for multiple services – Apps may require video, audio, and text
transmission.

Example: The developer of Zoom needs the network to support real-time

audio/video streaming without much delay.

3. Network Designer Perspective

• Network designers are engineers who plan and build the network
architecture.
• They focus on technical performance and efficiency of the network.
• Their main needs are:
• Scalability – The network should grow easily from 10 users to millions.
• Efficiency – The network should use minimum bandwidth and hardware
resources.
• Fault tolerance – If one part fails, the rest of the network should still
work.
• Flexibility – Should support new technologies in future (like 5G, IoT).

Example: Designers of Google’s data center networks ensure huge

scalability (millions of users) and high reliability.

4. Network Provider Perspective

• Network providers are organizations or Internet Service Providers (ISPs) that
own and manage the network infrastructure.
• They focus on cost, management, and security.
• Their main needs are:
• Manageability – Easy to monitor performance, detect faults, and update
software.
• Cost-effectiveness – Build and maintain networks with minimum cost,
but maximum service.
• Security – Prevent unauthorized access, attacks, or misuse of
resources.

Example: Airtel or Jio must manage millions of users while keeping costs
low and ensuring security.
Conclusion

A network is not only about cables and routers; it is about meeting the expectations of all
stakeholders.
stakeholders.
• End-users want reliability and speed.
• Developers want clear APIs and delivery guarantees.
• Designers want scalability and efficiency.
• Providers want manageability, security, and low cost.
⸻
Q4. Scalable Connectivity

Scalable connectivity means that a network should be able to grow (more users, devices, sites,
or traffic) without becoming slow, overloaded, or failing. A scalable network handles small
groups of devices as well as millions of devices with the same reliability.

Key Points Explained

1. Hierarchical Design (Core → Distribution → Access)
• Networks are built in layers:
• Core layer → backbone (fast switching and routing).
• Distribution layer → policies, VLANs, routing.
• Access layer → users and IoT devices.
• Advantage: Easy to expand. If one part grows, you only upgrade that
part, not the whole network.

2. Addressing that Scales

• CIDR (Classless Inter-Domain Routing) and subnetting help divide IP
addresses efficiently.
• IPv6 gives a huge address space for billions of devices (important for
IoT).
• This ensures that when more devices join, we don’t “run out” of
addresses.

3. Name Resolution that Scales (DNS)

• Humans use names ([Link]) instead of IPs ([Link]).
• DNS (Domain Name System) is hierarchical (root → .com →
[Link]).
• Even with millions of websites, DNS works fast because it distributes
the load across many servers.

4. Routing that Scales

• Inside the network (LAN/WAN), we use Interior Routing Protocols like
OSPF or IS-IS.
• Between networks (Internet), we use Exterior Routing Protocols like
BGP.
• Route summarization/aggregation reduces routing table size, so
 • Route summarization/aggregation reduces routing table size, so
routers don’t get overloaded.

5. Controlling Broadcast and Multicast Traffic

• Too many broadcast messages (like ARP requests) slow down the
network.
• Use VLANs to separate traffic.
• Use Spanning Tree Protocol (STP) to avoid loops.
• Use IGMP snooping and PIM (Protocol Independent Multicast) to
manage multicast traffic.

6. Stateless Services
• Some services (like web servers) don’t “remember” past requests
(stateless).
• Stateless services are easy to replicate behind load balancers, which
improves scalability.
• Example: Multiple identical web servers handle different users at the
same time.

7. Automation
• Large networks cannot be configured manually.
• Use configuration templates, intent-based networking, and zero-
touch provisioning.
• This makes scaling faster and reduces human errors.
Mini diagram
Example: Adding a new office = add an access block + summarize its routes
upstream; no global rework.

Conclusion: Plan hierarchy, addressing, and routing from day one to grow
painlessly.

Q5. Cost-Effective Resource Sharing

1. Concept: Cost-Effective Resource Sharing

Computer networks are designed to allow many users to share expensive resources—such as
transmission links, routers, servers, and storage—rather than providing a separate physical
resource for each user.

This sharing reduces cost while still providing acceptable performance.

• Example: Instead of giving every user a dedicated 10 Mbps line, the network
allows many users to share a high-speed 100 Mbps link, allocating bandwidth as
 needed.

The key mechanism that enables this is Multiplexing.

2. Multiplexing

Definition:
Multiplexing is the process of combining multiple independent data streams into one shared
physical channel. At the receiving end, a demultiplexer separates them back into their original
streams.

Goal: Efficient utilization of available bandwidth and cost-effective sharing of resources.

3. Multiplexing Techniques

(i) Time Division Multiplexing (TDM)

• The available link time is divided into fixed slots.
• Each user (or data stream) gets an assigned time slot in a round-robin
sequence.
• Transmission occurs whether or not the user has data to send.

Example:
• Telecommunication networks → digital voice channels (like old PCM
telephone systems).
• Four users share a 1 Mbps line, each allocated 250 kbps via time slots.

Limitation:
• Inefficient if a user is silent → slot goes unused (wasted bandwidth).

(ii) Frequency Division Multiplexing (FDM)

• The total available bandwidth is divided into non-overlapping frequency
bands.
• Each user transmits data continuously over a separate frequency channel.
• Requires band-pass filters to separate signals.
 Example:
• Radio & TV broadcasting → each channel has its own frequency band.
• DSL broadband → multiple users share telephone lines via different
frequency bands.

Limitation:
• Frequency bands are fixed and reserved, even when idle.
• Less flexible, can lead to under-utilization.

(iii) Statistical Multiplexing (Packet Switching)

• Unlike TDM/FDM, resources are allocated dynamically.
• Data is sent in the form of packets, only when a user has data to send.
• Packets from different users are stored in a queue and transmitted as soon
as the link is free.

Advantages:
1. Efficient use of bandwidth → no slot/frequency wastage.
2. Supports variable packet sizes (flexibility).
3. Dynamic allocation → bandwidth given to active users only.
4. Ideal for bursty traffic (e.g., web browsing, emails, video streaming).

Core Idea of the Internet:

Statistical multiplexing = packet switching, which is the fundamental basis of IP
networks.

Limitation:
• May cause queuing delay or packet loss during congestion.
Mini Diagram
-----------

Q6. Support for Common Services

Modern networks are not just about connecting machines; they must support the needs of
applications. Applications (like web browsing, video streaming, or client–server systems) rely on
channels that provide reliable communication despite errors and failures.

Channel Types

1. Request/Reply Channels
• Used by client-server applications.
• A client sends a request message, and the server replies with a response
message.
• Requires low delay, correctness, and reliability, but usually not continuous
transmission.

Examples:
• Web browsing (HTTP request/response).
• DNS lookup (query/response).

2. Message Stream Channels

• Designed for continuous media applications such as audio or video
streaming.
• Focus is on maintaining order and timing rather than strict reliability.
• A few packet losses can be tolerated, but delay and jitter (variation in delay)
 must be minimized.

Examples:
• Voice over IP (VoIP).
• Online video conferencing.

Reliability Classes

Networks must handle different kinds of reliability issues. Errors occur at different levels, so we
classify them into three categories:

1. Bit-Level Errors
• Occur when individual bits are flipped due to noise or interference in
transmission.
• Can be single-bit errors or burst errors (several consecutive bits
corrupted).
• Detected and corrected using error detection codes (parity, CRC,
checksums).

Example: Wireless communication often suffers from bit errors.

2. Packet-Level Errors
• Even if bits are transmitted correctly, entire packets can be lost, duplicated,
or delivered out of order.
• Causes: buffer overflows in routers, congestion, or incorrect packet
forwarding.
• Managed by protocols like TCP (which retransmits lost packets, removes
duplicates, reorders packets).

Example: Internet packet loss during congestion.

3. Node/Link Failures
• Failure of a network node (router, switch, server) or physical link (fiber,
copper, wireless channel).
• Causes: equipment failure, power outages, cable cuts.
• Reliability provided by redundancy (backup links, multiple paths) and
routing protocols that adapt (like OSPF, BGP).

Example: If a fiber-optic cable between two ISPs is cut, routing protocols

automatically find another path
⸻

Q7. Manageability
Concept

A network is not useful unless it can be easily managed. As networks grow in size and
complexity, administrators need tools and mechanisms for configuration, monitoring,
troubleshooting, optimizing performance, enforcing security, and scaling administration.
Manageability ensures that the network remains reliable, efficient, secure, and scalable
throughout its lifetime.

Management Requirements

1. Network Configuration and Monitoring

• Involves setting up devices (routers, switches, firewalls, servers) and
ensuring they run with the correct parameters.
• Continuous monitoring is needed to detect abnormal traffic, bottlenecks, or
failures.
• Tools: SNMP (Simple Network Management Protocol), modern SDN
controllers, dashboards.

Example: Setting IP addresses, routing tables, VLANs, and monitoring link

utilization.
2. Fault Detection and Recovery
• Networks face hardware failures, software bugs, and human errors.
• Fault management involves:
• Detecting faults quickly (alarms, alerts).
• Isolating faulty components.
• Recovering via rerouting or replacing equipment.
• Techniques: redundant links, failover routing protocols (e.g., OSPF, BGP).

Example: If a fiber cable is cut, traffic is rerouted via an alternate path

automatically.
3. Performance Optimization
• Network performance (throughput, delay, jitter, packet loss) must be
optimized to meet application requirements.
• Involves:
• Load balancing across multiple paths/servers.
• Congestion control to prevent overload.
• QoS (Quality of Service) to prioritize critical traffic.
 Example: Giving higher priority to video conferencing traffic compared to
bulk file downloads.
4. Security Policy Enforcement
• Ensures the network is protected from unauthorized access, attacks, and
misuse.
• Enforces security rules like:
• Firewalls and access control lists (ACLs).
• Encryption (IPSec, TLS).
• Authentication/authorization systems.
• Policies must be centrally managed but enforced locally.

Example: Preventing unauthorized users from accessing confidential data.

5. Scalable Administration
• As networks expand to thousands or millions of devices, management must
remain feasible.
• Requires:
• Automation (scripts, SDN controllers).
• Hierarchical management (divide network into regions).
• Centralized policies with distributed enforcement.

Example: An ISP managing millions of broadband subscribers efficiently.

⸻
Q8. Protocol Layering

Benefits of Layering
• Modularity → Each layer has specific functions.
• Abstraction → Higher layers hide lower-layer details.
• Standardization → Well-defined interfaces between layers.
Types of Protocol Layering
→ OSI 7-layer model
→TCP/IP 5-layer model.

✦ OSI Model (10 Marks)

Definition

The OSI (Open Systems Interconnection) reference model is a 7-layer conceptual framework
developed by ISO. It standardizes how data is transmitted over a network, ensuring
interoperability between different hardware and software systems.
Layers of OSI Model
1. Physical Layer (Layer 1)
• Transmits raw bits over physical medium (cables, fiber, radio).
• Defines voltage levels, connectors, frequencies.
2. Data Link Layer (Layer 2)
• Provides error detection/correction, framing, and flow control.
• Examples: Ethernet, Wi-Fi, PPP.
3. Network Layer (Layer 3)
• Handles logical addressing and routing of packets.
• Example protocols: IP, ICMP.
4. Transport Layer (Layer 4)
• Provides end-to-end reliability, segmentation, and error recovery.
• Protocols: TCP (reliable), UDP (fast but unreliable).
5. Session Layer (Layer 5)
• Manages dialogs between applications (setup, maintain, terminate).
6. Presentation Layer (Layer 6)
• Translates data formats, encryption, and compression.
• Example: SSL/TLS for encryption.
7. Application Layer (Layer 7)
• Provides services to users: web browsing, email, file transfer.
• Examples: HTTP, FTP, SMTP, DNS.

Advantages
• Provides modularity → Each layer has unique function.
• Enables interoperability between different vendors.
• Facilitates troubleshooting since problems can be isolated per layer.
✦ TCP/IP Model (10 Marks)

Definition

The TCP/IP (Transmission Control Protocol / Internet Protocol) model is a 5-layer practical
model used in the Internet. It was developed by DARPA and forms the basis of today’s global
networking.

Layers of TCP/IP Model

1. Physical Layer (Layer 1)
• Responsible for transmission of raw data bits via physical media.
2. Data Link Layer (Layer 2)
• Error detection/correction, MAC addressing.
• Examples: Ethernet, Wi-Fi.
3. Network Layer (Layer 3)
• Logical addressing, routing, packet forwarding.
• Protocols: IP, ICMP, ARP.
4. Transport Layer (Layer 4)
• End-to-end communication, segmentation, error control.
 • Protocols: TCP, UDP.
5. Application Layer (Layer 5)
• Provides user services like web, email, file transfer.
• Protocols: HTTP, FTP, SMTP, DNS.

Advantages
• Practical implementation (basis of the Internet).
• Simpler than OSI (5 layers vs 7).
• Supports scalability for millions of devices.
⸻

Q9. Performance

Definition: How well the network carries traffic for users/apps.

Key metrics
• Throughput: Useful bits per second delivered.
• Latency (delay): Time from send → receive.
• Jitter: Variation in latency (important for voice/video).
• Loss: % packets dropped (hurts TCP and real-time).
• Goodput: Throughput minus protocol overhead/retransmissions.
• Utilization: Load vs capacity; avoid persistent >70–80%.

Factors
• Link rate, queue sizes, congestion control, RTT, errors, protocol overhead
(MSS/MTU), application patterns.

Conclusion: Measure, then tune: capacity, queues, and protocols to meet app
SLAs.

⸻
Q10. Internet/network architecture
Q10. Bandwidth and Latency

Bandwidth: Capacity of a link (e.g., 100 Mbps).

Latency: Time for a bit/packet to travel (ms).

Delays inside latency

• Propagation: Distance / signal speed (fiber ≈ 200,000 km/s).
• Transmission: Packet_size / link_rate.
• Queuing: Waiting in device buffers (varies with load).
• Processing: Header checks, routing lookups.

Examples
• 1,500-byte packet on 10 Mbps → ~1.2 ms transmission.
• 2,000 km fiber → ~10 ms one-way propagation.

Conclusion: Bandwidth sets how much; latency sets how fast the first bit arrives.
Both matter.

Q11. Delay × Bandwidth Product (DBP)

Definition: Amount of data “in flight” on a path.

Formula

DBP (bits) = Path bandwidth (bits/s) × Round-trip or one-way delay (s)

Why important
• TCP window must be ≥ DBP to fully utilize the pipe.
• Buffer sizing in routers/senders.
• Explains why long-fat networks (LFNs) need large windows and window
scaling.

Example
• 100 Mbps, RTT 100 ms → DBP = 10 Mb = 1.25 MB.
TCP needs ≥1.25 MB in flight to fill the pipe.

Conclusion: DBP tells you the minimum data in flight for full throughput.

Q12. Perspectives on Connecting

Definition: Ways to connect machines, from simplest to Internet-scale.

Ways
• Direct (point-to-point): Two nodes, simple but not scalable.
• LAN with switching: Many hosts share switches (MAC learning, VLANs).
• WANs: Leased lines/MPLS/SD-WAN interconnect sites.
• Internetworking: Different networks joined by IP routers → global
connectivity.
• Overlay/VPN: Logical networks (IPsec, SSL-VPN, WireGuard) over the
Internet.

Conclusion: We connect locally with switches; we connect sites/globally with

routers and overlays.

Q13. Classes of Links

Point-to-Point: One sender ↔ one receiver (fiber run, serial link).

Broadcast/Shared Medium: Many nodes share (classic Ethernet bus, Wi-Fi).

Switched: Each pair gets a virtual circuit/path (modern Ethernet with switches).

Multipoint: More than two nodes on one link (satellite uplink).

Duplex modes:
Simplex (one-way),
Half-duplex (turn-taking),
Full-duplex (both ways).

Implications
• Access control on shared media (CSMA/CA in Wi-Fi).
• Dedicated links give predictable bandwidth/low collisions.

Q14. Reliable Transmission

Goal: Deliver data correctly and in order over unreliable channels.

Building blocks
• Error detection: Parity, checksum, CRC.
• ARQ (Automatic Repeat reQuest): ACK/NACK + retransmission on loss.
• Sequencing: Sequence numbers to keep order & detect duplicates.
• Timers: Retransmit if ACK not seen in time.
• Flow control: Don’t overwhelm receiver (e.g., TCP receive window).
• Congestion control: Don’t overwhelm network (e.g., TCP Reno/CUBIC).

⸻
Q15. Stop-and-Wait ARQ
Definition:
Stop-and-Wait is a basic data link layer protocol where the sender sends one frame at a time
and then waits for an ACK (acknowledgment) from the receiver before sending the next frame.
Working
1. Sender sends one frame.
2. Receiver gets the frame and sends back ACK.
3. After receiving ACK, sender sends the next frame.
4. This repeats until all data is sent.

This ensures orderly delivery of data.

Advantages
• Very simple to understand and implement.
• Ensures frames are received in correct order.
Disadvantages (Problems)
• Inefficient use of bandwidth → only one frame in transit, rest of the channel
is idle.
• Slow on long-distance links (e.g., satellite), as sender waits too much.
• No error recovery → assumes noiseless channel.
• Idle waiting → sender can’t send new frames until ACK is received
Figure : [Link]
⸻

Q16. Sliding Window Protocols (Go-Back-N & Selective Repeat)

• It is a data link layer protocol used to ensure reliable and sequential
delivery of data frames.
• Instead of sending one frame at a time (like Stop-and-Wait), it allows the
sender to send multiple frames at once, before needing an acknowledgment.
• Each frame has a sequence number → helps detect lost frames, duplicates,
and ensures correct order.
• The receiver uses acknowledgments (ACKs) to confirm received frames.
• Used in TCP (Transmission Control Protocol).
• Two major types:
1. Go-Back-N ARQ
2. Selective Repeat ARQ

Example analogy: Think of sending pages of a document — instead of

waiting after every page, you send a bunch of pages (window) and the
receiver confirms them together.

Go-Back-N ARQ Protocol (a type of Sliding Window)

 • A flow control method where the sender can send multiple frames (up to
N) before stopping for acknowledgment.
• Sender window size = N → Can send N frames ahead.
• Receiver window size = 1 → Can only accept the next expected frame in
sequence.
• If a frame is lost or arrives corrupted:
• Receiver rejects it (does not acknowledge).
• Sender must “go back” and retransmit that frame and all following
frames, even if some were correctly received.

Example of Go-Back-N ARQ

Suppose window size N = 4:

1. Step 1 → Sender sends frames 0, 1, 2, 3.
2. Step 2 → Sender waits for acknowledgment (ACK) of frame 0.
3. Step 3 → When ACK for frame 0 arrives, the window slides → Sender can
now send frame 4.
4. Step 4 → Receiver acknowledges frame 1 → Sender sends frame 5.
5. Step 5 → If ACK for frame 2 is lost or corrupted, sender does not slide the
window further.
6. Step 6 → Sender must go back to frame 2 and retransmit frames 2, 3, 4, 5.

Selective Repeat ARQ (10 Marks Answer)

Definition:
Selective Repeat ARQ (Automatic Repeat reQuest) is an error-control protocol that retransmits
only the specific frames that are lost or corrupted, instead of resending all frames. This makes it
more efficient than Go-Back-N ARQ.

How It Works
1. Sliding Window:
• Sender can send multiple frames in sequence without waiting for ACKs.
• Both sender and receiver maintain a window size (usually half the
sequence number range).
2. Acknowledgments:
• Each frame has an individual acknowledgment (ACK).
• If a frame is lost/corrupted, receiver sends a Negative ACK (NACK) only
for that frame.
3. Retransmission:
• Sender retransmits only the erroneous frame, not all subsequent ones.
4. Receiver Window:
 • Receiver can store out-of-order frames until the missing one arrives,
then reassemble data in correct order.

Step-by-step Example
• Sender sends frames: 0, 1, 2, 3.
• Receiver ACKs: frame 0 , frame 1 .
• Frame 2 (lost/corrupted) → Receiver sends NACK for frame 2.
• Sender resends only frame 2.
• Meanwhile, frames 3 and 4 may already be stored in buffer at receiver until
frame 2 arrives.

Advantages
• Efficient: Only retransmits erroneous frames.
• Faster recovery: No need to resend correctly received frames.
• Good for high-speed networks with few errors.

(Figure: [Link]

Q17. Concurrent Logical Channels

Definition: Multiple logical channels share one physical link so several

conversations run in parallel.

How it’s done

• Multiplexing:
• TDM: Time slots per flow (e.g., T1/E1).
• FDM/OFDMA: Different frequencies/subcarriers (DSL, LTE/5G).
• CDM: Different codes (3G).
• Virtual circuits/labels: ATM VCs, MPLS labels create many logical paths on
one trunk.
• Ports/streams: TCP ports, QUIC streams = many logical flows over one 5-
tuple.
• VLANs/VXLANs: Many L2/L3 “virtual networks” over the same wires.

Benefits
 • Higher utilization, isolation between flows/tenants, simpler management.

Conclusion: Logical channels turn one pipe into many, giving concurrency and
separation.