GTU 3150710 Computer Networks (CN) Summer 2026 Solved Paper & Solutions

Client-server architecture is a network architecture where central machines called servers provide services, and host computers called clients request those services.

Servers run continuously with a fixed IP address. Clients initiate requests and do not communicate directly with other clients.

Applications on the Internet include:

• Web browsing: Web browsers query web servers for pages using HTTP.
• Electronic mail: Mail clients send and retrieve messages through mail servers using SMTP and IMAP.
• File transfer: Users download and upload files through FTP servers.
• Domain Name System (DNS): Resolvers query DNS servers to map domain names to IP addresses.

The Simple Mail Transfer Protocol (SMTP) is an application-layer protocol used to transfer email messages reliably across the Internet over a TCP connection on port 25.

SMTP is a push protocol: the sending mail server pushes the message to the receiving mail server. The transfer follows three main phases:

1. Connection setup: The client opens a TCP connection to the server. The server greets the client with a 220 code, and the client introduces itself using HELO or EHLO.
2. Mail transfer: The client identifies the sender with MAIL FROM and the recipient with RCPT TO. It then issues the DATA command, transmits the message body line by line, and signals the end of the text with a single period (.) on an empty line.
3. Connection closure: The client sends the QUIT command, and the server closes the connection.

The lines starting with C: are sent by the client and the lines starting with S: by the server.

The Open Systems Interconnection (OSI) model is a theoretical seven-layer reference framework developed by the International Organisation for Standardisation (ISO) to guide network protocol design. Each layer provides services to the layer above it and receives services from the layer below it.

Seven layers of the OSI model

1. Physical layer: Transmits raw unstructured bit streams across physical transmission media. It defines electrical signals, data rates, pin layouts, and connector types.
2. Data link layer: Packages raw bits into data frames for node-to-node delivery. It handles physical MAC addressing, media access control, and link-level error detection.
3. Network layer: Routes packets across diverse networks from source to destination host. It performs logical IP addressing and path determination using routing algorithms.
4. Transport layer: Delivers complete messages between application processes. It provides segmentation, end-to-end flow control, connection management, and error recovery.
5. Session layer: Establishes, maintains, and terminates communication sessions between software processes. It manages dialog control and session checkpoints.
6. Presentation layer: Formats data for the application layer. It handles character code translation (such as ASCII to EBCDIC), data compression, and encryption.
7. Application layer: Serves as the user interface to network services. It supports applications with protocols such as HTTP, DNS, and SMTP.

Comparison with the TCP/IP model

The TCP/IP model is a practical four-layer architecture developed for the Internet.

A network topology is the physical or logical layout of nodes and links in a computer network.

In a star topology, each node connects to a central device, such as a switch or hub, using a point-to-point link.

Key characteristics:

• Centralised traffic: All communication between nodes passes through the central device.
• Fault isolation: Failure of a single link or node affects only that node; other nodes continue operating.
• Single point of failure: If the central switch fails, the entire network stops working.
• Easy maintenance: Simple to add, remove, and troubleshoot nodes.

TCP measures the Round Trip Time (RTT) to calculate its Retransmission Timeout (RTO) so it knows when a segment is lost. The estimation uses four steps:

1. SampleRTT: The measured duration between transmitting a segment and receiving its acknowledgement. By Karn's algorithm, TCP never calculates SampleRTT for retransmitted segments to prevent timing ambiguity.
2. EstimatedRTT: An exponential weighted moving average of past SampleRTT values that smoothes out short-term network fluctuations:

RFC 6298 recommends setting $\alpha = 0.125$.

3. DevRTT: An estimate of how much SampleRTT deviates from EstimatedRTT:

RFC 6298 recommends setting $\beta = 0.25$.

4. Timeout Interval (RTO): TCP sets the timer by adding a safety margin to EstimatedRTT:

When sample variance is high, DevRTT expands the timeout margin to prevent premature retransmissions. If a timer expires, TCP doubles the timeout interval.

The Hypertext Transfer Protocol (HTTP) is an application-layer request-response protocol used across the Web. HTTP messages are formatted as plain human-readable ASCII text.

HTTP request message format

An HTTP request message has four main parts:

1. Request line: The first line of the message. It consists of three fields separated by spaces:

• Method: The action to perform, such as GET (fetch a resource), POST (submit data), PUT (upload a file), or DELETE (remove a resource).
• URL: The path specifying the requested resource on the server (for example, /index.html).
• HTTP version: The protocol version being used, such as HTTP/1.1.

2. Header lines: Lines containing key-value pairs (Header-Name: value) that provide metadata about the client and request. Examples include Host (domain name), User-Agent (browser details), and Accept-Language (preferred language).
3. Blank line: A mandatory empty line containing carriage return and line feed (CRLF). It marks the end of the header section.
4. Entity body: Optional payload containing data sent to the server. It is present in requests like POST and PUT (holding form data or files) and empty in GET requests.

HTTP response message format

An HTTP response message also consists of four main parts:

1. Status line: The first line of the response. It contains three fields:

• HTTP version: Protocol version used by the server (HTTP/1.1).
• Status code: A three-digit integer indicating the result of the request (for example, 200 for success, 404 for not found, 500 for server error).
• Reason phrase: A short human-readable explanation of the status code (such as OK or Not Found).

2. Header lines: Metadata regarding the response and server. Examples include Date (timestamp), Server (server software), Content-Length (size of entity body in bytes), and Content-Type (media type, such as text/html).
3. Blank line: An empty line (CRLF) indicating the end of headers.
4. Entity body: The requested content itself, such as an HTML document, image, or JSON string.

Hierarchical routing is a routing strategy that divides a large network into administrative regions called Autonomous Systems (AS) or areas to keep routing tables small.

In this approach:

• Intra-AS routing: Routers within an autonomous system execute an internal routing protocol (such as RIP or OSPF). Each router only maintains routes to destinations inside its own region.
• Inter-AS routing: Designated gateway routers at the boundary of each region communicate with gateway routers of other regions using an exterior protocol (such as BGP).

Hierarchical routing prevents routing tables from growing excessively, saves router memory, and limits the spread of route updates.

A computer network is divided into the network edge, where applications run, and the network core, which transports data between systems.

Error detection techniques enable a receiver to detect bit errors introduced by channel noise during frame transmission.

Parity checks

A parity check appends one or more redundant parity bits to a data block to make the count of 1s even (even parity) or odd (odd parity).

1. Single-bit even parity: The sender appends a bit so the total number of 1s in the transmitted block is even.

• Example: Data is $1011001$ (four 1s). The sender appends parity bit $0$, transmitting $10110010$.
• If a single bit flips during transmission to $10110110$ (five 1s), the receiver detects an error.
• Limitation: It detects any odd number of bit errors, but fails completely if an even number of bits flip.

2. Two-dimensional parity: Bits are arranged in a matrix of rows and columns. Parity bits are calculated for each row and each column. This scheme detects and corrects any single-bit error by identifying the intersecting row and column.

Cyclic Redundancy Check (CRC)

CRC is an error-detection technique based on polynomial codes and modulo-2 binary arithmetic (XOR operations without carry or borrow).

Sender and receiver agree in advance on a generator pattern $G$ of length $r + 1$ whose highest-order bit is 1.

Algorithm:

1. Append $r$ zero bits to the $k$-bit data string $D$.
2. Divide the extended string $D \cdot 2^r$ by $G$ using modulo-2 division.
3. The resulting $r$-bit remainder $R$ is appended to $D$ to form the transmitted frame.
4. The receiver divides the received frame by $G$. A non-zero remainder indicates transmission error; a zero remainder indicates no error.

Example: let data $D = 1010$ and generator $G = 1011$, so $r = 3$. Append 3 zeros to get $1010000$ and divide by $G$ using XOR:

The remainder is $R = 011$, so the transmitted frame is $1010011$.

Receiver check: divide the received frame $1010011$ by $G$.

IP fragmentation is the process where a router splits an IPv4 datagram into smaller packets called fragments so they can pass through a link with a smaller Maximum Transmission Unit (MTU).

Three IPv4 header fields manage this process:

• Identification: Uniquely labels all fragments originating from the same datagram.
• Flags: Contains the More Fragments (MF) bit, set to 1 for every fragment except the last one.
• Fragment offset: Specifies the relative offset of the fragment payload in 8-byte units.

The destination host reassembles all fragments before delivering the packet to higher layers.

Slotted ALOHA is a random access protocol in which time is divided into discrete equal-length slots, each equal to the transmission time of one frame. A station can transmit data only at the beginning of a time slot.

By eliminating partial collisions, Slotted ALOHA cuts the vulnerable window in half and doubles channel efficiency compared to Pure ALOHA.

TCP provides reliable communication by managing transmission rates through two distinct mechanisms: flow control (protecting the receiving host) and congestion control (protecting the intermediate network).

TCP flow control

Flow control prevents a high-speed sender from overflowing the receive buffer of a slower receiver.

• Receive window (rwnd): The receiver advertises its available buffer space in the 16-bit rwnd field of every returning TCP acknowledgement segment.
• Window management: The sender limits in-flight unacknowledged bytes so they never exceed rwnd:

• Zero-window probing: If the receiver buffer fills up, it announces $\text{rwnd} = 0$, telling the sender to pause. To prevent deadlock when buffer space reopens, the sender periodically transmits a 1-byte probe segment. The receiver responds with its updated rwnd.

TCP congestion control

Congestion control prevents sender traffic from overwhelming routers and communication links across the Internet.

The sender maintains an internal congestion window (cwnd). The effective transmission window is $\min(\text{cwnd}, \text{rwnd})$. TCP adjusts cwnd through three main algorithms:

1. Slow start: Begins with $\text{cwnd} = 1\text{ MSS}$. The sender doubles cwnd every round-trip time (adding 1 MSS per received ACK) until it reaches the slow start threshold (ssthresh).
2. Congestion avoidance: Once $\text{cwnd} \ge \text{ssthresh}$, growth switches from exponential to linear. The sender increases cwnd by 1 MSS per RTT (additive increase) to probe for bandwidth cautiously.
3. Loss recovery:

• Triple duplicate ACKs: Indicates mild congestion where packets are still flowing. TCP sets $\text{ssthresh} = \text{cwnd} / 2$, reduces cwnd by half (multiplicative decrease), and performs fast retransmit without waiting for a timer.
• Timeout: Indicates severe congestion. TCP sets $\text{ssthresh} = \text{cwnd} / 2$, collapses $\text{cwnd} = 1\text{ MSS}$, and restarts slow start.

True.

Justification:

The Routing Information Protocol (RIP) is an interior gateway protocol (IGP) designed specifically for routing within a single Autonomous System (AS).

• Design scope: RIP is based on the distance-vector routing algorithm and uses hop count as its metric. It is suited only for small and medium internal networks because its maximum hop count is limited to 15.
• Inter-AS routing difference: Routing between different autonomous systems requires exterior gateway protocols like the Border Gateway Protocol (BGP). BGP handles policy routing, autonomous administration, and global scaling, which RIP cannot support.

Therefore, RIP operates strictly as an intra-AS routing protocol.

The User Datagram Protocol (UDP) is a lightweight, connectionless transport protocol that provides best-effort packet delivery without connection setup delays, flow control, or retransmission overhead.

Major applications of UDP include:

1. Real-time multimedia and streaming: Applications such as Voice over IP (VoIP), video conferencing (Google Meet, Zoom), and live broadcast streaming rely on UDP. They tolerate occasional lost packets but cannot tolerate delay jitters caused by TCP retransmissions.
2. Online multiplayer gaming: Fast-action games transmit real-time player position updates over UDP. Dropped position updates are superseded by newer packets, making TCP retransmissions unnecessary.
3. Domain Name System (DNS): DNS queries use UDP on port 53. Since a query and its response fit in single datagrams, UDP avoids the three-way handshake delay of TCP, delivering faster lookups.
4. Network management and discovery: Protocols like Simple Network Management Protocol (SNMP) and DHCP use UDP to collect device metrics and assign IP addresses even under congested network conditions.
5. Simple file and time protocols: Trivial File Transfer Protocol (TFTP) and Network Time Protocol (NTP) use UDP for low-overhead communication.

IP addressing and packet forwarding are the foundational functions of the network layer, enabling end-to-end data communication across the Internet.

IP addressing

An IP address is a unique logical numerical identifier assigned to every network interface connected to the Internet.

• IPv4: Uses 32-bit addresses written in dotted-decimal notation (such as 192.168.1.1), providing approximately $4.3$ billion distinct addresses.
• Classless Inter-Domain Routing (CIDR): Replaced older class-based schemes. An address is divided into a network prefix and a host identifier written in the format $a.b.c.d/x$, where $x$ denotes the number of bits in the prefix (subnet mask).
• IPv6: Uses 128-bit addresses written in hexadecimal notation (such as 2001:0db8::1) to eliminate IPv4 address exhaustion and support modern networking requirements.

Packet forwarding

Forwarding is the data-plane operation where a router moves an incoming packet from its input link interface to the appropriate output link interface.

1. Forwarding table lookup: When an IP datagram arrives, the router inspects the destination IP address in the header and searches its forwarding table.
2. Longest prefix matching: If the destination address matches multiple table entries, the router selects the entry with the longest prefix match (the most specific network route).
3. Header processing: The router decrements the datagram's Time to Live (TTL) field to prevent infinite loops, updates the header checksum, and drops the packet if TTL reaches zero.
4. Switching fabric transit: The datagram travels across the internal switching fabric to the designated output port buffer.
5. Link transmission: The output port encapsulates the datagram into a new link-layer frame and transmits it across the outgoing physical link.