1.

Structured Network Design (3-Layer Model)

A structured design has three main layers:
Layer Role Key Devices Example
Switches,
Access Connects end-user devices Office computers in
wireless
Layer (computers, printers, etc.) a LAN
access points
Connects multiple access Layer 3 Connects different
Distribution
layers and applies policies switches, departments in a
Layer
(e.g., VLANs, security) routers building
High-speed backbone
High-speed Connects data
connecting distribution
Core Layer routers, core centers & large-
layers across buildings or
switches scale networks
campuses
✔ Why This Matters?
 Each layer has a specific role, improving network performance and
troubleshooting.
 Prevents bottlenecks by organizing traffic flow efficiently.

2. What Do Routers Do?

Forwarding (Packet Switching)
 A router receives a packet on an incoming link and forwards it to an
outgoing link.
 It looks at the destination IP address and sends the packet to the right
path.
✔ Example:
 Your laptop sends a request to www.google.com. The router forwards
your request toward Google's servers using the best available path.

Routing (Finding the Best Path)

 Routers calculate the best path to send data across networks using
routing tables.
 They use routing algorithms to determine the most efficient path.
✔ Example:
 A router decides whether to send a packet via a fiber-optic connection
or a wireless link based on speed and reliability.

Routing Algorithms
 Used by routers to find optimal paths between sender & receiver.
 Common algorithms:
o Distance Vector (e.g., RIP) → Based on hop count (shortest path).
o Link-State (e.g., OSPF) → Considers network conditions
(bandwidth, congestion).
o Hybrid (e.g., EIGRP) → Mix of both for efficiency.
✔ Example:
  If a fiber-optic link fails, the router reroutes traffic via an alternative
path using OSPF.

What Do Switches Do?

Switch = High-Speed Local Networking
 A switch is a Layer 2 (Data Link layer) device.
 It forwards frames (Ethernet packets) between connected devices
using MAC addresses.
 Unlike routers, switches do NOT decide long-distance paths—they only
work within local networks (LANs).
✔ Example:
 In an office network, a switch connects all employee computers so they
can share files and printers.

Key Differences Between Routers & Switches

Feature Router Switch
Works at Layer 3 (Network Works at Layer 2 (Data Link
Layer
layer) layer)
Routes packets between Switches frames within a
Function
networks network
Uses IP or
Uses IP addresses Uses MAC addresses
MAC?
Connects a home to the Connects computers inside an
Example
Internet office
✔ Why This Matters?
 Routers connect different networks (e.g., home to the Internet).
 Switches connect devices inside a network (e.g., office computers).

Understanding Network Addresses: IP, MAC, and Hostnames

In computer networking, devices need unique addresses to communicate
efficiently. Three main types of addresses are used:
Type Purpose Example Scope
Identifies a device on a network Global (Internet
IP
and allows routing between 192.168.1.10 or private
Address
networks networks)
MAC Identifies a network interface 00:1A:2B:3C:4D:
Local (LAN)
Address (NIC) for local communication 5E
Host Human-readable name server.example.c Mapped to IP via
Name assigned to a device om DNS

1. IP Address (Logical Address)

 An IP address is a unique identifier for a device on a network that
follows the Internet Protocol (IP).
 Used for routing traffic across networks (WAN, LAN, Internet).
 Example: 192.0.2.1 (IPv4) or 2001:db8::ff00:42:8329 (IPv6).
✔ Why It Matters?
 Devices need an IP address to send & receive data over the Internet.
 Routers use IP addresses to forward packets to the correct destination.

2. MAC Address (Physical Address)

 A MAC address is a unique identifier assigned to a network interface
card (NIC).
 Used for local communication within a LAN.
 Format: 00:1A:2B:3C:4D:5E (Hexadecimal).
✔ Why It Matters?
 Switches use MAC addresses to forward frames within a local network.
 A device's MAC address never changes, unlike an IP address which may
be dynamic.
✔ Example:
 A laptop’s Wi-Fi adapter and Ethernet port each have a unique MAC
address.

3. Host Name (Human-Readable Identifier)

 A host name is a user-friendly name assigned to a device.
 Used to identify devices within a network (e.g., webserver1.local or
example.com).
 Needs to be resolved to an IP address before communication.
✔ Why It Matters?
 Instead of remembering 192.168.1.100, you can use myserver.local.
 DNS (Domain Name System) translates hostnames to IP addresses.
✔ Example:
 When you type www.google.com, DNS resolves it to an actual IP
address like 142.250.190.78.

4. How They Work Together

✔ A user enters www.example.com in their browser →
✔ The DNS server resolves www.example.com to an IP address
(93.184.216.34) →
✔ The router forwards packets based on IP addresses →
✔ The switch forwards frames using MAC addresses inside the local network
→
✔ The website loads successfully! 🎉

Technical Goals in Network Design

A well-designed network must meet various technical goals to ensure
reliability, performance, scalability, security, and cost-effectiveness. Below is
a breakdown of these goals with examples.

Availability & Reliability

Definition:
 Availability refers to how often a network is operational and accessible
to users.
 Reliability measures the ability of devices and network components to
function without failure.
Measurement:
 Expressed as percent uptime over a given period (year, month, week,
etc.).
 Example: If a network is up for 165 hours in a 7-day week, its
availability is: (165/168)×100=98.21% uptime(165 / 168) \times 100 =
98.21\% \text{ uptime}(165/168)×100=98.21% uptime
Key Factors:
 Reliable devices (high-quality switches, routers, servers).
 Redundancy (backup links, failover systems).
MTBF & MTTR:
 MTBF (Mean Time Between Failures): Expected time before the next
failure.
 MTTR (Mean Time To Repair): How quickly a failure is fixed.
Example Calculation:
 MTBF Goal: 4000 hours
 MTTR Goal: 1 hour
 Mean Availability Goal: 40004000+1=99.98%\frac{4000}{4000 + 1}
= 99.98\%4000+14000=99.98%
 This means the network is designed to be available 99.98% of the
time.

Network Performance
Throughput & Bandwidth
Throughput: The actual amount of error-free data transmitted per second.
Bandwidth: The maximum capacity of a network link.
Example:
 A 1 Gbps Ethernet link might only achieve 800 Mbps due to network
congestion or errors.

Delay (Latency)
Definition: The time it takes for a data packet to travel from sender to
receiver.
Types of Delay:
1. Processing Delay (time to process data at a node).
2. Transmission Delay (time to send bits over the link).
3. Propagation Delay (time for bits to travel across the medium).
Example:
 Sending a 1500-byte packet over a 100 Mbps network:
1500×8100,000,000=120 microseconds\frac{1500 \times 8}
{100,000,000} = 120 \text{ microseconds}100,000,0001500×8
=120 microseconds
Response Time
Definition: The time between a request and a response in a network service.
Example: A web server responding in 200 ms to a user request.

Scalability
Definition: A network’s ability to grow and adapt without degrading
performance.
Example:
 A company expands from 100 users to 500 users, and the network is
designed to handle this growth.
 Cloud-based solutions allow dynamic resource allocation based on
demand.

Security
Three Key Principles:
1. Confidentiality – Prevent unauthorized access to data.
2. Integrity – Ensure data is not altered during transmission.
3. Availability – Ensure network services are always accessible.
✔ Security Strategies:
 Firewalls to block unauthorized traffic.
 Encryption (SSL, TLS, VPNs) for secure data transmission.
 Intrusion Detection Systems (IDS) to detect suspicious activity.

Total Cost of Ownership (TCO)

Definition: The total cost of operating, maintaining, and upgrading a network.
Includes:
 Hardware costs (routers, switches, servers).
 Software licensing and updates.
 Support staff salaries.
 Training for IT personnel.
 Upgrades and replacements.
✔ Example Calculation:
 A company spends $10,000 per year on hardware maintenance,
$5,000 on software licenses, and $15,000 on IT staff.
 Total TCO per year: $30,000.

Application Layer & Network Design Considerations

The Application Layer is responsible for how applications communicate over
a network, defining protocols, data formats, and communication rules.
However, while software developers implement application logic, the
network designer selects the appropriate architecture for deploying these
applications efficiently.

Key Responsibilities of the Application Layer

✔ Defines how network applications function
✔ Specifies communication protocols & data formats
✔ Ensures interoperability between different systems
✔ Relies on application architectures chosen by network designers
📌 Example Protocols in the Application Layer:
 HTTP/HTTPS (Web browsing)
 SMTP, IMAP, POP3 (Email communication)
 FTP/SFTP (File transfers)
 DNS (Translating domain names into IP addresses)
 VoIP Protocols (SIP, RTP) (Voice over IP calls)

Application Architectures Chosen by Network Designers

The network designer selects an architecture based on the application’s
requirements and the three key technical goals (Availability, Response Time,
Scalability).

How the 3 Key Technical Goals Influence Architecture Choice

🔹 Availability – Cloud-based services or redundant server setups ensure
minimal downtime.
🔹 Performance – Client-server or edge computing improves response times.
🔹 Scalability – Cloud and P2P architectures scale efficiently with demand.
✔ Example:
A company expecting rapid growth in users should choose a cloud-based
architecture because it can scale without requiring major hardware
upgrades.

Types of Network Applications

Network applications can be classified into two main categories:
Transaction Processing Systems (TPS)
✔ Handles business transactions in real-time
✔ Requires a database server & DBMS
✔ Follows a structured processing cycle
📌 TPS Processing Cycle:
1. Data Entry: Captures transaction details.
2. Transaction Processing: Validates and processes the data.
3. Database Update: Stores transaction results in a DB.
4. Output Generation: Produces reports and confirmations.
📌 TPS Examples:
 Airline booking systems (e.g., Amadeus, Sabre)
 Order management systems (e.g., Shopify, SAP)
 Accounting systems (e.g., QuickBooks, Oracle ERP)
 University enrollment systems (e.g., YU Enrolment System)
 Payment processing systems (e.g., PayPal, Stripe)

Non-TPS Networked Applications

✔ Do not involve structured transaction processing
✔ Do not require a database server or DBMS
📌 Examples:
 Email services (e.g., Outlook, Gmail)
 Web browsing (e.g., Chrome, Firefox)
 File storage & sharing (e.g., Google Drive, OneDrive)
 Video conferencing (e.g., Zoom, Microsoft Teams)
 Instant messaging (e.g., Slack, WhatsApp)
Key Difference: TPS vs. Non-TPS Applications
Feature TPS Applications Non-TPS Applications
Database Required? ✅ Yes ❌ No
Processes Business
✅ Yes ❌ No
Transactions?
Airline booking, Email, web browsing, video
Example Applications
accounting systems conferencing

Application Architectures for Transaction Processing Systems (TPS)

TPS applications perform five key functions:
✔ Data Storage: Stores and retrieves data.
✔ Data Access Logic: Queries data from storage.
✔ Application Logic: Defines business rules & processing.
✔ Presentation Logic: Handles the user interface.
✔ Services Logic (if needed): Allows automated access to other apps.
The application architecture determines how these five functions are
distributed among clients and servers. This is different from network
architecture, which focuses on hardware and connectivity.

1️.Host-Based Architecture
✔ All five functions are handled by a central mainframe (host computer).
✔ The client (terminal) is only used for displaying results.
Example:
 Banking systems in the 1970s where branch terminals connected to a
mainframe for transaction processing.
Pros:
 Simple management & security since everything is on one machine.
Cons:
 Scalability issues – the host is a bottleneck when user demand
increases.

2️.Client-Based Architecture (Bad Practice)

✔ The client computer handles everything except data storage.
✔ The database server only stores data.
Example:
 Early Microsoft Access applications where clients handled queries and
logic.
Pros:
 Reduces load on the server.
Cons:
 Bad practice because high network traffic occurs when multiple clients
query data.

3️.Client-Server Architecture (Best Practice)

✔ Tasks are shared between client and server:
 Client: Handles presentation logic
 Server: Handles data storage & access logic
 Either client or server: Handles application logic
Example:
 Online banking systems (e.g., RBC, TD)
 Enterprise ERP systems (e.g., SAP, Oracle)
Pros:
 Efficient data processing, scalable.
Cons:
 Requires proper load balancing.

4️.Cloud Computing Architecture (A variation of Client-Server)

✔ The server-side processing is moved to the cloud.
✔ Clients only need an internet connection.
Example:
 Amazon Web Services (AWS) for hosting applications.
 Google Docs & Microsoft 365 for cloud-based document editing.
Pros:
 Highly scalable, reduces local infrastructure costs.
Cons:
 Dependent on internet connection & cloud provider reliability.

5️.Peer-to-Peer (P2P) Architecture (A variation of Client-Based)

✔ Clients share data directly without a central server.
✔ Used when users need to share resources directly.
Example:
 BitTorrent for file sharing.
 Cryptocurrency networks like Bitcoin.
Pros:
 No need for central servers, decentralized & scalable.
Cons:
 Difficult to manage security & performance.

Comparison of TPS Application Architectures

Example Use
Architecture Pros Cons
Case
Host-Based Centralized control, Not scalable, Early banking
secure expensive systems
 Example Use
Architecture Pros Cons
Case
mainframes
Client-Based (Bad Microsoft
Reduces server load High network traffic
Practice) Access DB
Load balancing
Client-Server Efficient, scalable Online banking
needed
Scalable, low local Google Docs,
Cloud Computing Internet dependency
cost AWS apps
No central control, BitTorrent,
Peer-to-Peer Security issues
decentralized Bitcoin

Two-Tier, Three-Tier, and N-Tier Architectures in Client-Server Systems

The tiered architectures define how application logic is partitioned in a Client-
Server (C/S) model. Each tier is responsible for specific tasks.

Two-Tier Architecture (2-Tier)

✔ One client, one server
✔ The application logic is split between the client and server.
✔ Used when fewer than 50 users need to access the system.
Example:
 Small business database applications (e.g., MS Access with a SQL
Server backend).
Pros:
 Simple design, easy to manage.
Cons:
 Not scalable – performance decreases as users increase.

Three-Tier Architecture (3-Tier)

✔ Three layers:
 Client (Presentation Layer): Handles the user interface.
 Application Server (Logic Layer): Processes business rules & service
logic.
 Database Server (Data Layer): Manages data storage & access.
Example:
 E-commerce websites like Amazon where:
o The browser (client) displays products.
o The application server handles search & shopping cart logic.
o The database server stores product details & orders.
Pros:
 More scalable than 2-tier
Cons:
 More complex, requires load balancing.

3️N-Tier Architecture (More than 3 Tiers)

✔ Adds extra layers like:
 Web Server Layer (to handle HTTP requests).
 Load Balancer (to distribute traffic).
 Caching Servers (to speed up data retrieval).
Example:
 Netflix or YouTube (which use multiple layers to handle millions of
requests).
Pros:
 Highly scalable, supports millions of users.
Cons:
 Expensive & complex to manage.

Comparison of Architectures
Architectur
Pros Cons Example Use Case
e
Simple, good for small Small business
Two-Tier Not scalable
systems DBs
Scalable, better Requires more E-commerce
Three-Tier
performance resources websites
Highly scalable, better
N-Tier Expensive, complex Netflix, YouTube
security

Thin vs. Thick (Fat) Clients

The difference between thin and thick clients lies in how much of the
application logic runs on the client device.

Thin Client
✔ Minimal application logic on the client.
✔ Most processing happens on the server.
✔ The client is mainly for display and user input.
Example:
 Google Docs – You only need a browser, all processing happens on
Google’s servers.
 Virtual Desktop Infrastructure (VDI) – Citrix, Remote Desktop.
Pros:
 Easier maintenance (only update the server).
 Cheaper devices (no powerful CPU or storage needed).
Cons:
 Relies on a constant network connection.
 Performance depends on server speed.

Thick (Fat) Client

✔ Most application logic is on the client.
✔ Server mainly stores data, while the client does the processing.
Example:
  Microsoft Excel installed on a laptop (instead of using Google Sheets
online).
 Gaming PCs that run games locally instead of streaming from a cloud.
Pros:
 Works offline without needing a server.
 Faster performance for local processing.
Cons:
 Harder to maintain (updates required on every client).
 More expensive (requires powerful hardware).

Comparison of Thin vs. Thick Clients

Feature Thin Client Thick Client
Processing location Server Client
Hardware
Low (cheaper) High (expensive)
requirements
Network
Always needed Works offline
dependency
Easier (centralized
Maintenance Harder (manual updates)
updates)
Google Docs, Virtual Installed Microsoft Excel, PC
Example
Desktops gaming

Cloud Computing Architecture: Explanation & Examples

Cloud computing is a variation of the client-server (C/S) model, where
organizations outsource computing resources (hardware, software, storage,
etc.) to a third-party provider. Instead of maintaining on-premise servers,
businesses pay for computing services over the Internet, MAN, or WAN.

1️Three Cloud Computing Models

Who Manages
Model What It Provides Example
What?
Third-party provides Vendor manages Google Docs,
SaaS (Software
everything: apps, OS, hardware + Microsoft 365,
as a Service)
and hardware software Dropbox
Provides hardware AWS EC2,
IaaS Vendor manages
(servers, storage, Google
(Infrastructure hardware, user
networking). User installs Compute
as a Service) manages software
software. Engine
Vendor manages
Provides development AWS Lambda,
PaaS (Platform hardware + OS,
tools + hosting. User Google App
as a Service) user develops
codes apps. Engine
apps

Example: Cloud Computing in Action

Scenario: A Startup Wants to Launch an Online Store
1️SaaS Approach
 The startup uses Shopify to build the store.
 No need to install or manage servers—just sign up and start selling!
2️IaaS Approach
 The startup rents virtual servers from AWS EC2.
 They install Linux, a database, and an e-commerce platform on the
rented servers.
3️PaaS Approach
 The startup uses Google App Engine to build a custom store.
 The cloud provider manages OS, runtime, and hosting, while the
startup only writes code.

Advantages & Disadvantages of Cloud Computing

Pros:
 Scalability – Add more storage/computing power as needed.
 Cost-Effective – No need to buy expensive servers.
 Anywhere Access – Accessible from any device with internet.
 Security & Maintenance – Managed by cloud provider.
Cons:
 Internet Dependency – No internet = no access.
 Security Risks – Data stored in third-party servers.
 Less Control – Limited control over cloud infrastructure.

Peer-to-Peer (P2P) Architecture: Explanation & Example

Definition:
Peer-to-Peer (P2P) architecture does not follow the traditional client-server
model. Instead, every device (computer) acts as both a client and a server,
meaning they can request and provide resources (files, processing power,
etc.).
How It Works:
 Each device (peer) has equal responsibility in the network.
 A central server is often used only to help peers locate each other.
 Once a connection is made, peers share data directly without needing
a central server.

Best Practice Architecture Selection
1️Host-Based Architecture
Best for:
 Large organizations with high transaction processing (TPS) needs.
 Companies that process huge volumes of structured data (e.g.,
banking systems, airline reservations).
Limitations:
 Scalability issues – requires expensive mainframe upgrades.
 Single point of failure – if the central host fails, the entire system is
down.
2️Client-Server (C/S) Architecture
Best for:
 Most modern applications (whether TPS or not).
 Businesses that require flexibility, scalability, and cost-effectiveness.
 Can use thin or thick clients depending on the needs.
Limitations:
 Requires server maintenance and updates.
 More complex than a host-based model.

3️Cloud Computing Architecture

Best for:
 Businesses that want to outsource some or all of their IT infrastructure.
 Companies that need rapid scalability and lower initial costs.
Limitations:
 Security & privacy concerns – Data is stored on third-party servers.
 Ongoing costs – Instead of one-time investments, cloud services
require continuous payments.