UDP Port Numbers: Advantages & Disadvantages

Advantages:

• Simplified server implementation: Preassigned ports eliminate the need for servers to dynamically allocate ports for each incoming connection, reducing processing overhead.
• Well-known services: Certain ports are associated with specific services (e.g., port 80 for HTTP), allowing clients to easily connect without manual configuration.
• Firewall management: Firewalls can easily filter traffic based on destination port, simplifying security rules.

Disadvantages:

• Port exhaustion: A limited number of ports can be overwhelmed during peak usage, leading to connection failures.
• Security risks: Well-known port numbers are often targeted by attackers, making services more vulnerable.
• Limited flexibility: Dynamic port allocation might be needed for certain applications where static ports are impractical.

Protocol Ports vs. Process Identifiers

Advantages of Protocol Ports:

• Standardization: Ports are standardized across different operating systems, facilitating communication between diverse machines.
• Multiplexing: A single server can listen on multiple ports, handling different services simultaneously.
• Client anonymity: Clients don’t need to reveal their internal process identifiers, enhancing privacy.

Disadvantages of Process Identifiers:

• Non-standard: Process identifiers differ across operating systems, hindering interoperability.
• Security concerns: Exposing process identifiers can create security vulnerabilities.
• Limited scope: They only identify processes within a single machine, not across networks.

Reliable Datagram Protocol with Timeouts and Acknowledgements

Concept:

1. Sender assigns sequence numbers to datagrams.
2. Sender transmits datagrams with timeouts.
3. Receiver acknowledges received datagrams with corresponding sequence numbers.
4. Sender retransmits unacknowledged datagrams after timeout.

Overhead:

• Sequence numbers and acknowledgements add to packet size.
• Retransmissions can increase traffic and delay.

Delay:

• Timeouts introduce waiting time before retransmission.
• Round-trip time for acknowledgements contributes to delay.

TCP Stream Transmission with Finite Sequence Numbers

Mechanism:

• Initial sequence number (ISN) chosen randomly for each connection.
• Sequence number increments with each byte sent.
• Wrap-around occurs after reaching the maximum value, but subsequent ISNs differ significantly due to randomness.
• Receiver acknowledges data based on expected sequence number and window size.

Lost TCP Acknowledgements and Retransmissions

Reasons for not retransmitting:

• Delayed acknowledgement: Acknowledgement might arrive later due to network delays.
• Fast retransmission: TCP can trigger retransmission based on timeouts or duplicate ACKs without waiting for lost ACK.
• Congestion avoidance: Retransmitting immediately during potential congestion might worsen network conditions.

Arguments for and Against Automatic Idle Connection Closure

For:

• Reduces resource usage on servers, especially for long-lived connections with no activity.
• Prevents denial-of-service attacks by keeping connection slots available.
• Encourages clients to explicitly close connections when finished, improving network efficiency.

Against:

• Can inconvenience users if connections close prematurely due to brief pauses in activity.
• Requires clients to handle reconnection gracefully, adding complexity.
• Might cause unnecessary connection establishment overhead if activity resumes shortly after closure.

TCP System Crash and Confusion with Initial Sequence Number (ISN)

Scenario:

1. Client establishes connection with server (ISN1 used).
2. System crash and restart with ISN1 still in use.
3. Client reconnects and uses a new ISN2.
4. Server receives data with ISN2 but expects ISN1 due to crash, leading to confusion.

Solutions:

• Use timestamps in addition to ISNs to distinguish old and new connections.
• Maintain connection state information on the server to identify valid connections even after restarts.