Network Layer Protocol Practice Questions

TTL (Time To Live) in IPv4: originally intended as seconds but in practice implemented as a hop count. Each router decrements TTL by 1 before forwarding. If TTL reaches 0: the router discards the packet and sends an ICMP Time Exceeded message (Type 11, Code 0) back to the source IP. This prevents packets from circulating forever in routing loops. Common initial TTL values: Windows = 128, Linux = 64, Cisco routers = 255. Traceroute exploits TTL: sends packets with TTL=1,2,3,... - each router that discards a packet (TTL=0) sends an ICMP Time Exceeded, revealing the router's IP address.

IPv4 uses 32-bit addressing.

Routing algorithm classification: Static (non-adaptive) routing: routes are preconfigured by the network administrator and remain fixed regardless of network conditions. Simple but cannot adapt to failures or congestion. Suitable for small, stable networks. Dynamic (adaptive) routing: routes are automatically recalculated based on current network topology and traffic conditions. Routers exchange routing information to build and maintain routing tables. Subtypes: distance vector (RIP - uses Bellman-Ford), link-state (OSPF - uses Dijkstra's), path vector (BGP). Dynamic routing adapts to failures but has higher complexity and overhead.

/26 means 64 addresses per subnet. The network 192.168.10.64 spans 64 addresses: 192.168.10.64 to 192.168.10.127. The broadcast address is the last address: 192.168.10.127.

The four /24 networks: 200.23.16.0, 200.23.17.0, 200.23.18.0, 200.23.19.0. Convert the third octet to binary: 16=00010000, 17=00010001, 18=00010010, 19=00010011. Common prefix in bits 1-22: the first 22 bits are identical (00010000, 00010001, 00010010, 00010011 share 000100xx in octet 3). Supernet = 200.23.16.0/22, which covers 200.23.16.0 to 200.23.19.255 (2^10 = 1024 addresses). All four /24 networks fall within this /22 supernet. A /22 replaces 4 routing table entries with 1.

IPv4 header fields: A: TRUE - Version = 4 for IPv4, 6 for IPv6 (in IPv6 header). B: TRUE - IHL (4 bits) specifies header length in 32-bit (4-byte) words. Minimum IHL = 5 (20 bytes); maximum = 15 (60 bytes). C: TRUE - Protocol field (8 bits) identifies the next-level protocol: 1=ICMP, 6=TCP, 17=UDP, 89=OSPF, 50=ESP (IPsec). D: FALSE - the IPv4 checksum covers ONLY the IP header, not the data payload. Transport layer protocols (TCP, UDP) have their own checksums covering the entire payload. Covering only the header reduces checksum computation overhead and allows routers to quickly verify header integrity.

VLSM allocates different subnet sizes to match actual needs. For Department A needing 500 hosts: need 2^h - 2 ≥ 500 → 2^h ≥ 502 → h ≥ 9 (2^9 = 512). Mask = /23 (32-9=23 network bits), giving 2^9 - 2 = 510 usable hosts ✓. For Department B needing 250 hosts: h ≥ 8 (2^8=256), mask /24, usable = 254 ✓. For Department C needing 100 hosts: h ≥ 7 (2^7=128), mask /25, usable = 126 ✓. VLSM allocates: A gets /23 (512 addresses), B gets /24 (256 addresses), C gets /25 (128 addresses) - efficiently using the address space compared to fixed subnets.

BGP (Border Gateway Protocol) is the standard inter-domain routing protocol used between Autonomous Systems (AS) on the Internet. OSPF, RIP, and EIGRP are intra-domain (interior gateway) protocols.

Bellman-Ford update rule: new cost via N = link cost to N + N's cost to D = 3 + 5 = 8. Compare with existing best: current cost to D = 7 (via M). Since 8 > 7, the route via N is NOT better - the router keeps its existing route via M with cost 7. The router only updates if the new cost is lower than the current best. No update is sent to neighbors (no change in the router's distance table). If the new cost were 6 (< 7), the router would update: route to D via N, cost 6, and then send triggered updates to its neighbors about this improvement.

IPv4 fragmentation: any router along the path can fragment a packet if it exceeds the outgoing link's MTU. Each fragment carries: original IP header with same Identification field, Fragment Offset (position in original datagram in 8-byte units), More Fragments (MF) flag set to 1 for all but the last fragment, Don't Fragment (DF) flag. Reassembly occurs ONLY at the destination - intermediate routers do NOT reassemble. If DF bit is set and packet exceeds MTU, the router discards the packet and sends ICMP 'Fragmentation Needed' (Type 3, Code 4). IPv6 does NOT support router fragmentation - only the source can fragment; routers use ICMP 'Packet Too Big' messages to signal Path MTU.

Original packet: 4000 bytes total = 20-byte header + 3980 bytes data. MTU = 1500 bytes → max data per fragment = 1500 - 20 = 1480 bytes. Fragment 1: 1480 bytes data, offset = 0, MF = 1. Fragment 2: 1480 bytes data, offset = 1480/8 = 185, MF = 1. Fragment 3: 3980 - 1480 - 1480 = 1020 bytes data, offset = 2960/8 = 370, MF = 0. Total fragments = 3. The fragment offset field is in units of 8 bytes (to fit in 13 bits). Second fragment offset = 1480/8 = 185.

BGP loop prevention using AS_PATH attribute: every BGP route advertisement carries the AS_PATH - an ordered sequence of AS numbers that the route has traversed. When a BGP router receives an advertisement: it checks if its own AS number (ASN) appears in the AS_PATH. If yes → the route would create a loop → it is rejected (not used or propagated). If no → the router prepends its own ASN to the AS_PATH and propagates the route. Example: AS1 → AS2 → AS3 route has AS_PATH [AS1, AS2]. When AS3 sends this back toward AS1, AS1 sees its own ASN [AS1] in the path and rejects it. This is analogous to poison reverse but at the AS level.

OSPF (RFC 2328) is a link-state routing protocol: (1) Each router discovers its neighbors using Hello packets. (2) Each router creates a Link State Advertisement (LSA) describing its links and their costs. (3) LSAs are flooded to ALL routers in the OSPF area (using reliable flooding). (4) Each router builds an identical Link State Database (LSDB) - a complete map of the network topology. (5) Each router independently runs Dijkstra's Shortest Path First (SPF) algorithm on the LSDB to compute the shortest path tree rooted at itself. Metric: cost (typically inversely proportional to bandwidth). OSPF supports hierarchical routing with areas, reducing LSA flooding overhead.

To divide /24 into 4 subnets, we need 2 additional bits (since $2^2 = 4$), giving /26. The subnet mask for /26 is 255.255.255.192 (11000000 in the last octet, i.e., 192).

Class B addresses range from 128.0.0.0 to 191.255.255.255. 172.16.0.0 starts with 172 (128-191 range), making it a Class B address. It is also a private IP address range.

Count-to-infinity: suppose router A reaches destination D via router B (cost 1). If the A-D link fails: B still believes it can reach D via A (cost 2 from B's perspective). A updates its table: it can reach D via B at cost 3. B updates: via A at cost 4. This ping-pong continues, incrementing the metric slowly toward infinity (16 in RIP). Mitigation techniques: Split horizon - don't advertise a route back to the neighbor from which you learned it. Split horizon with route poisoning - advertise failed routes with metric = infinity (16 in RIP). Holddown timers - ignore updates about a failed route for a period. Triggered updates - send updates immediately on change rather than waiting for the periodic timer.

RIP uses hop count as its sole routing metric. The maximum hop count is 15; a hop count of 16 is considered infinity (unreachable). This limits RIP to small networks.

BGP best path selection (Cisco implementation, commonly tested): (1) Highest Weight (Cisco-proprietary, local to router). (2) Highest LOCAL_PREF (local to AS, higher = preferred; used for outbound routing policy). (3) Locally originated routes preferred. (4) Shortest AS_PATH length. (5) Lowest Origin type (IGP < EGP < Incomplete). (6) Lowest MED (Multi-Exit Discriminator - hint to neighboring AS about preferred entry point). (7) eBGP routes preferred over iBGP routes. (8) Lowest IGP metric to BGP next-hop. (9) Oldest eBGP route (for stability). (10) Lowest BGP Router ID (tie-breaker). The mnemonic: 'We Love Oranges As Oranges Mean Pure Refreshment' for Weight, LOCAL_PREF, Originated, AS_PATH, Origin, MED, Prefer eBGP, Router ID.