Computer Networks

Answer

(iii) Both (a) and (b) — i.e. Data Link Layer and Transport Layer.

Explanation

Flow control means controlling the rate at which a sender transmits data so that a receiver is not overwhelmed.

Flow control can be performed at more than one layer, but the mechanism and scope differ:

1. Data Link Layer: Provides hop-to-hop flow control between directly connected devices. Protocols at this layer can regulate the amount of data sent over a particular link.

2. Transport Layer: Provides end-to-end flow control between communicating processes. For example, TCP uses a receiver window (rwnd) so that a sender does not transmit more data than the receiving host can buffer.

Therefore, both the Data Link and Transport layers can perform flow-control functions, but at different scopes.

Exam point

Data Link = hop-to-hop flow control

Transport = end-to-end flow control

Final answer: (iii) Both (a) and (b).

Answer

Hamming code is an error-detecting and error-correcting block code that adds parity bits to data so that certain bit errors can be detected and corrected at the receiver.

For a data word containing $m$ data bits, the number of parity bits $r$ is chosen so that:
$2^r \ge m+r+1$.

The parity bits are placed at positions that are powers of 2: $1,2,4,8,\ldots$. Their parity checks allow the receiver to determine the position of a single-bit error.

Exam-ready answer

Hamming code is a forward error-correction technique that adds redundant parity bits to a data word. It can generally correct a single-bit error and detect certain multiple-bit errors depending on the variant.

For 2 marks: Definition + single-bit error-correction property is sufficient.

Answer

(ii) is NOT true.

Explanation

The statement reverses the addresses used by bridges and routers.

• A transparent bridge operates at the Data Link layer and learns a forwarding table by examining the source MAC addresses of incoming frames.
• A router operates at the Network layer and makes forwarding decisions using logical IP addresses.
• Both can selectively forward traffic, and a router can connect different networks such as a LAN and WAN.

Therefore, (ii) A bridge uses IP addresses while a router uses MAC addresses is false.

Final answer: (ii).

Answer

(iv) Channel coding

Explanation

The Data Link layer is responsible for functions such as framing, error control, and flow control. Channel coding, in the sense of encoding bits for reliable physical transmission over a medium, is primarily associated with the Physical layer.

• Framing: Data Link layer groups the bit stream into frames.
• Error control: Data Link layer can detect/correct errors or arrange retransmission.
• Flow control: Data Link protocols can prevent a fast sender from overwhelming a receiver on a link.
• Channel coding: Primarily a Physical-layer transmission function.

Final answer: (iv) Channel coding.

Answer

Piggybacking is a technique in bidirectional data communication where an acknowledgement (ACK) is attached to an outgoing data frame instead of sending a separate ACK frame.

Example

If A sends a data frame to B and B also has data to send to A, B can include the acknowledgement for A's frame in B's outgoing data frame.

Advantage

Piggybacking reduces the number of separate control frames and therefore improves bandwidth efficiency.

Limitation

The receiver may need to wait briefly for outgoing data before sending the ACK, so a maximum waiting/delay limit is normally used.

Exam-ready: Piggybacking means combining an ACK with an outgoing data frame to reduce overhead.

Answer

A network switch is a Data Link-layer device that connects devices within a LAN and forwards Ethernet frames based on MAC addresses.

Working

A switch learns the source MAC address of incoming frames and associates it with the receiving port in its MAC/forwarding table. When a frame arrives, it checks the destination MAC address:

• If the destination is known, it forwards the frame only through the appropriate port.
• If unknown, it floods the frame within the relevant LAN/VLAN.

Advantages

Switches reduce unnecessary traffic and allow multiple devices to communicate simultaneously using separate switch ports.

Exam-ready: A switch is a multiport Data Link-layer device that forwards frames using MAC addresses.

Answer

The jam signal in CSMA/CD is sent after a collision is detected so that all stations on the shared Ethernet medium become aware that a collision has occurred.

Why it is needed

A station detecting a collision transmits the jam signal, ensuring the collision is long enough to be noticed by other transmitting stations. The affected stations then stop the current transmission and apply the collision backoff procedure before trying again.

Exam-ready: The jam signal informs other stations of a collision and ensures the collision is propagated throughout the shared medium.

Solution

The minimum number of distinct sequence numbers must be large enough to distinguish all frames that may be outstanding at the same time. For a sliding-window ARQ scheme with transmitter window size $N$ and receiver window size $M$, the required number of distinct sequence numbers is determined by the larger window size.

Therefore, the answer is:

(ii) $\max(M,N)$

So the minimum number of distinct sequence numbers required is $\max(M,N)$.

Answer

(ii) Two times

Explanation

In pure ALOHA, a frame can be transmitted at any time. Suppose a frame takes time $T$ to transmit.

For another frame to avoid collision with it, another transmission must not begin during the interval from $T$ before the frame starts until $T$ after the frame starts.

Therefore the vulnerable period is:

$T_v = 2T$

So the vulnerable time is twice the frame transmission time.

This is also why the maximum theoretical throughput of pure ALOHA is approximately $18.4\%$.

Final answer: (ii) Two times.

Answer

(i) Media Access Control (MAC) sublayer

Explanation

The Data Link layer is commonly divided into two sublayers:

1. LLC (Logical Link Control): Provides a common interface to the Network layer and handles functions independent of the particular transmission medium.
2. MAC (Media Access Control): Handles functions that depend on the characteristics of the shared medium, especially access to the transmission medium.

Therefore, the sublayer whose functions depend on the kind of medium is the MAC sublayer.

Final answer: (i) Media Access Control sublayer.

Answer

(iv) GBN discards packets that are received out of order.

Explanation

In Go-Back-N (GBN) ARQ, the receiver accepts only the next packet in sequence. If a packet arrives out of order, the receiver does not buffer it as an accepted packet; it discards it and typically sends a duplicate cumulative ACK for the last correctly received in-order packet.

Why the other options are wrong

• (i) False: GBN does not require the sequence-number space to be at least twice the window size. A common condition is $W \le 2^k-1$, where $W$ is the sender window and $k$ is the number of sequence-number bits.
• (ii) False: Standard GBN uses one timer, associated with the oldest unacknowledged packet.
• (iii) False: Duplicate ACKs inform the sender about the missing packet, but standard GBN retransmission is primarily triggered by the timer expiring, after which the outstanding packets are retransmitted from the missing one onward.
• (iv) True: Out-of-order packets are discarded by the GBN receiver.

Final answer: (iv) Discards packets received out of order.

Answer

Dividing an Ethernet LAN using a bridge provides several advantages:

1. Reduces collision domain size: Each bridged segment is a separate collision domain, reducing contention.
2. Improves performance: Local traffic can remain within its segment instead of consuming bandwidth on every segment.
3. Filters frames: The bridge learns MAC addresses and forwards frames only when necessary.
4. Extends the LAN: Different physical Ethernet segments can be interconnected.
5. Reduces congestion: Separating traffic between segments can reduce the load on each segment.

Exam-ready: A bridge divides a LAN into smaller collision domains, filters unnecessary traffic using MAC addresses, and improves overall network performance.

Solution

Go-Back-N (GBN) and Selective Repeat (SR) are sliding-window ARQ protocols used for reliable data transmission. Both use acknowledgements, sequence numbers, timers, and retransmission to recover from lost or damaged frames, but they differ in how they handle errors.

Feature	Go-Back-N ARQ	Selective Repeat ARQ
Receiver window	Usually 1	Greater than 1; accepts multiple frames
Out-of-order frames	Discarded	Buffered
Retransmission	Retransmits the erroneous/lost frame and all subsequent outstanding frames	Retransmits only the specific lost/damaged frame
ACKs	Typically cumulative	Typically selective/individual
Receiver complexity	Lower	Higher
Buffer requirement	Low	Higher
Bandwidth efficiency	Lower when errors are frequent	Higher, especially on noisy links
Implementation	Simpler	More complex

Example

Suppose frames 0, 1, 2, 3 are sent and frame 2 is lost.

Go-Back-N: The receiver detects the missing frame and does not accept frame 3 as the next in-order frame. The sender eventually retransmits frames 2 and 3 (and any later outstanding frames).

Selective Repeat: The receiver can accept and buffer frame 3 while waiting for frame 2. The sender retransmits only frame 2. After frame 2 arrives, the buffered frame 3 can be delivered in order.

Conclusion

Go-Back-N is simpler and requires less receiver buffering, but it can waste bandwidth because correctly received frames may be retransmitted. Selective Repeat is more efficient on error-prone or long-delay links because only lost or damaged frames are retransmitted, but it requires more receiver memory and more complex control logic.