Network Layer Protocol Practice Questions

The subnet mask $255.255.31.0$ is $11111111.11111111.00011111.00000000$ in binary. Two IP addresses belong to the same network if their bitwise AND with the mask yields the same network ID. For Option D ($128.8.129.43$ and $128.8.161.55$), we analyze the third octets: $129$ ($10000001_2$) and $161$ ($10100001_2$). Applying the mask ($31 = 00011111_2$):
$129 \text{ AND } 31 = 10000001_2 \text{ AND } 00011111_2 = 00000001_2 = 1$
$161 \text{ AND } 31 = 10100001_2 \text{ AND } 00011111_2 = 00000001_2 = 1$
Since both calculations result in the same network ID ($1$), the addresses belong to the same network.

Option A is true because a host can have multiple network interfaces, each assigned a distinct $IP$ address. Option B is true as routers independently process packets, allowing dynamic routing based on network conditions. Option C is true because the $TTL$ (Time to Live) field in the header ensures packets are discarded if they exceed the maximum hop count, preventing infinite loops. Option D is false because the $IP$ protocol includes "Source Routing" options (e.g., $Strict Source$ and $Loose Source$ routing), which allow the source node to explicitly specify the sequence of routers the packet must traverse. Thus, the assertion that the route is determined only by routing tables is incorrect.

RIP (RFC 2453) characteristics: A: TRUE - metric = hop count (number of routers to destination); max = 15 hops; 16 = infinity (unreachable). This limits RIP to small networks. B: TRUE - RIP sends complete routing table updates (UDP port 520) to all directly connected neighbors every 30 seconds. C: FALSE - RIP uses Bellman-Ford (distributed distance vector), not Dijkstra's (which is used by OSPF). D: TRUE - count-to-infinity: when a route fails, incorrect information can propagate slowly through the network (counting up to 16). Mitigations: split horizon (don't advertise a route back to the neighbor you learned it from), split horizon with route poisoning (advertise failed routes with metric 16), hold-down timers.

Dijkstra's with a binary min-heap runs in $O((e + n) \log n)$ which simplifies to $O(e \log n)$ for connected graphs. With an adjacency matrix it's $O(n^2)$, and with a Fibonacci heap it's $O(e + n \log n)$.

Split horizon prevents a router from advertising a route back on the same interface through which it learned the route, reducing routing loops and speeding up convergence.

Routing protocol comparison: A: RIP - distance vector, metric = hop count (max 15), periodic updates every 30s, slow convergence, count-to-infinity problem, suitable for small networks only. B: OSPF - link-state, metric = cost (inversely proportional to bandwidth: cost = 10^8 / interface bandwidth), triggered updates (fast convergence), hierarchical areas, Dijkstra's SPF, supports VLSM/CIDR, authentication. Widely used in enterprise and ISP networks. C: BGP - path vector, exterior gateway protocol (between ASes), policy-based routing (LOCAL_PREF, MED, AS_PATH), loop prevention via AS_PATH, TCP port 179 (reliable transport), currently BGP-4 (RFC 4271) with MP-BGP extensions for IPv6 and VPNs. D is FALSE - RIP and OSPF are interior gateway protocols (IGPs); BGP is the exterior gateway protocol (EGP).

IPv6 uses 128-bit addresses (compared to IPv4's 32-bit), providing $2^{128}$ possible addresses. They are written as 8 groups of 4 hexadecimal digits separated by colons.

In CIDR notation, /26 means 26 bits for the network prefix, leaving 32 - 26 = 6 bits for hosts. Total addresses = 2^6 = 64. Usable host addresses = 2^6 - 2 = 62 (subtract 1 for network address: all host bits = 0, and 1 for broadcast address: all host bits = 1). Subnet details: Network address: 192.168.10.0, Broadcast: 192.168.10.63, Usable hosts: 192.168.10.1 to 192.168.10.62 (62 hosts).

Using the Bellman-Ford equation: D(X, A) = min over neighbors {cost(X, neighbor) + D(neighbor, A)}. Via Y: cost(X, Y) + D(Y, A) = 1 + 7 = 8. Via Z: cost(X, Z) + D(Z, A) = 2 + 2 = 4. D(X, A) = min(8, 4) = 4, via neighbor Z. Note: the table entries given are what Y and Z report as their costs to the destinations. X computes its own cost by adding the link cost to Y or Z respectively. Best path to A = 4 through Z.

Link State Routing requires each router to store the complete network topology (link state database). With $n$ routers and up to $O(n^2)$ links, storage is $O(n^2)$. Distance Vector only stores $O(n)$ next-hop entries.

DHCP uses UDP: client-to-server port 67, server-to-client port 68. DORA process: (1) DHCP Discover: client broadcasts (src IP 0.0.0.0, dst 255.255.255.255) to find DHCP servers. (2) DHCP Offer: server(s) respond with an offered IP address, subnet mask, gateway, DNS, and lease duration. (3) DHCP Request: client broadcasts its selection (accepting one offer, notifying others). (4) DHCP Acknowledge: selected server confirms the IP assignment with lease time. After DORA, the client has: IP address, subnet mask, default gateway, DNS server addresses, and lease duration. DHCP Relay agents forward DHCP broadcasts across router boundaries.

Routers use longest prefix match. 192.168.5.64/26 covers 192.168.5.64-127 (26-bit prefix), while 192.168.5.0/24 covers the entire /24. Since 73 is in the /26 range and /26 is more specific, it is selected.

Flooding characteristics: A: TRUE - flooding sends packets on ALL paths simultaneously; if any path exists to the destination, flooding will find it. B: TRUE - flooding is maximally robust - even if multiple routers fail, as long as one path survives the packet will be delivered. Used in military networks and link-state routing (controlled flooding of LSAs). C: TRUE - flooding generates exponentially many duplicate packets. Without countermeasures (hop limit, sequence numbers, reverse path forwarding), packets loop forever. Solutions: hop count limit in the packet header, sequence number per source so routers discard duplicates. D: FALSE - flooding requires NO routing table; it simply forwards on all links except the input - this is its key advantage (no routing computation) and disadvantage (massive redundancy).

IS-IS vs OSPF: A: TRUE - IS-IS runs directly over the Data Link Layer (it is not encapsulated in IP packets). This makes IS-IS immune to IP routing errors during convergence - an advantage in large ISP networks. OSPF encapsulates its messages in IP packets (protocol number 89). B: TRUE - IS-IS has Level 1 (intra-area routing) and Level 2 (inter-area/backbone routing); Level 1/2 routers handle both. Analogous to OSPF's area structure and ABRs. C: TRUE - both IS-IS and OSPF use Dijkstra's Shortest Path First algorithm on their respective link-state databases. D: FALSE - IS-IS supports IPv6 through extension TLVs (Multi-Topology IS-IS); it is widely used for IPv6 routing in ISP networks. OSPFv3 is the OSPF variant for IPv6.

Dijkstra's from A: Initialize: d[A]=0, d[B]=∞, d[C]=∞, d[D]=∞. Visit A (d=0): update d[B]=1, d[C]=4. Visit B (d=1): update d[C]=min(4, 1+2)=3, d[D]=1+5=6. Visit C (d=3): update d[D]=min(6, 3+1)=4. Visit D (d=4): done. Shortest path A to D = 4, via A→B→C→D (cost 1+2+1=4).

Special IPv4 addresses: 127.0.0.1 (loopback): traffic sent to 127.x.x.x never leaves the host - used for testing local network stack. 255.255.255.255 (limited broadcast): broadcast to all hosts on the local network; not forwarded by routers. 0.0.0.0 (unspecified): used as source IP during DHCP Discover (before an IP is assigned), also means 'any address' in socket binding. 224.0.0.1 (all hosts multicast): all hosts on the local subnet; Class D range (224.0.0.0-239.255.255.255) is reserved for multicast. Additional special: 169.254.0.0/16 (APIPA/link-local, auto-assigned when DHCP fails).

/30 subnets have block size 4. 100.10.5.2 belongs to subnet 100.10.5.0/30, while 100.10.5.5 and 100.10.5.6 belong to subnet 100.10.5.4/30.

IPv6 abbreviation rules: (1) Leading zeros in each group can be omitted: 0db8 → db8, 0000 → 0. (2) One or more consecutive groups of all zeros can be replaced with :: (only once per address). So 2001:0db8:0000:0000:0000:0000:0000:0001 → 2001:db8::1 (option A, using ::). Option B is also valid (removing leading zeros but not using ::). Option C: ::1 = 0000:...:0001 (128 bits) = IPv6 loopback (equivalent to IPv4's 127.0.0.1). Option D: fe80::/10 is the link-local prefix - automatically assigned, not routable, used for neighbor discovery and router discovery on local links.