computer networks midterm

What is the goal of Content Distribution Networks (CDNs)?

Serve large volumes of content (e.g., video, software) to geographically distributed users with low latency and high availability.

Describe the 'Bring-home' architecture for CDNs.

Deploy fewer, larger server clusters at Internet Exchange Points (IXPs).

What are the benefits of using CDNs?

Reduced backbone traffic, faster downloads, load balancing, and fault tolerance.

How does DASH (Dynamic Adaptive Streaming over HTTP) use "Chunking"?

Video is divided into equal-duration segments, each encoded at multiple bit-rates.

What is the purpose of a Manifest File in DASH?

It lists all segment URLs, bit-rate options, durations, codecs, and resolutions for a video.

Describe the client logic in DASH.

Measure recent download throughput, consult the manifest, request the highest sustainable bit-rate, and adaptively switch bit-rates to prevent buffer underrun/overflow.

What is HTTP (HyperText Transfer Protocol)?

Stateless request/response protocol over TCP (port 80 for HTTP, 443 for HTTPS).

Explain the Non-persistent connection model in HTTP.

One request/response per TCP connection.

Explain the Persistent (keep-alive) connection model in HTTP.

Multiple requests/responses over one connection (default in HTTP/1.1).

What is the structural classification of DNS (Domain Name System)?

Hierarchical Distributed Database.

What is a Recursive DNS query?

A DNS query type where the resolver returns the final answer to the client.

What is an Iterative DNS query?

A DNS query type where the resolver returns a referral to the next-level server.

Name some common DNS Resource Record (RR) types.

A/AAAA (IPv4/IPv6 addresses), NS (Name server), CNAME (Canonical name), MX (Mail exchange).

What is the primary purpose of SMTP (Simple Mail Transfer Protocol)?

Push e-mail between mail servers.

What are the key characteristics of SMTP?

ASCII-based commands/responses with status codes and uses persistent TCP connections by default.

Describe Centralized P2P architecture.

A directory server tracks peers and resources (e.g., original Napster).

Describe Decentralized (Pure) P2P architecture.

Peers form an overlay network and route queries via flooding or Distributed Hash Tables (DHTs).

How does Hybrid P2P (e.g., BitTorrent) work?

A tracker coordinates peers in a "swarm," files are split into pieces, and peers share piece-availability bitmaps with a rarest-first strategy for downloading least-common pieces first.

What are RTP and RTCP used for in real-time streaming?

RTP (over UDP) adds sequence numbers and timestamps for playout timing and jitter buffering, while RTCP provides periodic quality-of-service feedback.

What is the function of RTSP in real-time streaming?

It controls playback (play, pause, seek) on streaming servers.

What is the primary service of the Transport Layer?

End-to-end communication between processes on hosts.

Explain Multiplexing and Demultiplexing in the Transport Layer.

Multiplexing: Sender collects data from multiple sockets, adds header (port numbers). Demultiplexing: Receiver uses header to deliver to correct socket.

Compare TCP and UDP Transport Protocols.

TCP is reliable, in-order, connection-oriented. UDP is unreliable, unordered, connectionless, with minimal overhead.

What are the key fields in a UDP header?

Source Port (16 bits), Destination Port (16 bits), Length (header + data, 16 bits), Checksum (16 bits).

What are the characteristics of UDP?

No handshake or teardown, best-effort delivery (segments may be lost, duplicated, or reordered), used by DNS, streaming audio/video, VoIP, SNMP.

What is the purpose of Checksums in the Transport Layer?

To detect bit errors in headers and payloads (UDP, TCP); it does not correct errors.

Describe the Checksum algorithm (One’s-Complement Sum) including verification at the receiver.

1. Break data into 16-bit words, pad with 0 if needed.
2. Sum using one’s-complement addition (wrap any carry back into LSB).
3. Take one’s-complement of the final sum to get the checksum field.
4. Receiver sums all words including checksum; result should be all 1s (0xFFFF) if no error.

Describe the IP Service Model in the Network Layer.

Unreliable, Best-Effort Datagram Delivery: No guarantees on delay, loss, throughput, or order.

Name key fields in an IPv4 Header.

Version, IHL, DSCP/ECN, Total Length, Identification, Flags (DF/MF), Fragment Offset, TTL, Protocol (e.g., TCP=6, UDP=17), Header Checksum, Source IP, Destination IP.

What is Forwarding in the Network Layer?

Looking up the destination IP in a forwarding table using longest-prefix match.

Explain Subnetting and CIDR.

Techniques using variable-length prefixes (e.g., 192.0.2.0/24) to allocate IP address blocks flexibly.

Describe Distance-Vector Routing (e.g., RIP).

Neighbors exchange full distance vectors periodically. Updates use Bellman–Ford formula: Dx(y) = \min{v \in \text{neighbors}} { c(x,v) + D_v(y)}. Susceptible to slow convergence and count-to-infinity.

Describe Link-State Routing (e.g., OSPF).

Routers flood link-state advertisements (neighbor lists & link costs) to entire area. Each router runs Dijkstra’s algorithm to compute shortest paths. Offers faster convergence but is computationally intensive per update.

What is the purpose of ICMP (Internet Control Message Protocol)?

Error reporting (e.g., Destination Unreachable, Time Exceeded) and diagnostic tools (Echo Request/Reply for ping).

What is the function of DHCP (Dynamic Host Configuration Protocol)?

Clients obtain IP address, subnet mask, gateway, and DNS servers from a DHCP server via UDP.

Explain the difference between Static NAT and Dynamic/Overload NAT (PAT).

Static NAT is a one-to-one mapping between private and public IP addresses. Dynamic/Overload (PAT) allows many internal hosts to share one public IP with distinct port mappings.

What is Framing in the Link Layer?

Encapsulating network-layer datagrams into frames.

Explain the Byte-stuffing delimiter technique.

Using a special flag byte; if this flag appears in the data, it is escaped with an additional byte.

Explain the Bit-stuffing delimiter technique.

Using a special bit pattern (e.g., 01111110); if five consecutive 1s appear in data, a 0 is inserted after them.

What is a Parity Bit and its limitation in error detection?

Detects any single-bit error (odd number of bit flips). Fails on an even number of bit flips.

Explain CRC (Cyclic Redundancy Check) and its error detection capabilities.

Treats frame as a polynomial and divides it by a generator polynomial. The remainder is appended as the Frame Check Sequence (FCS). Detects all single/double-bit errors and burst errors up to the generator length.

Compare Pure Aloha and Slotted Aloha.

Aloha: Random transmissions; Pure Aloha has higher collision probability. Slotted Aloha reduces the collision window by half by requiring transmissions to start at fixed time intervals.

Differentiate between CSMA/CD and CSMA/CA.

CSMA/CD (Ethernet): Detects collisions and aborts transmission, sending a jam signal. CSMA/CA (Wi-Fi): Uses collision avoidance with random backoff delays and acknowledgments to prevent collisions.

Describe the Ethernet Frame Format.

Preamble | Dest MAC | Src MAC | EtherType/Length | Payload (46–1500 bytes) | CRC.

What are MAC Addresses?

48-bit addresses, globally unique, burned into Network Interface Cards (NICs).

Compare an Ethernet Switch vs. a Hub.

Hubs create a shared medium where collisions affect all ports. Switches learn MAC-to-port mappings, isolate collision domains per port, and support full-duplex communication.

What is the purpose of ARP (Address Resolution Protocol)?

To map an IPv4 address to a MAC address within a Local Area Network (LAN).

Describe the operation of ARP.

1. Sender broadcasts an ARP Request: "Who has IP X.X.X.X? Tell me."
2. The host with the matching IP Unicasts an ARP Reply with its MAC address.
3. The sender caches the IP-to-MAC mapping.