How to use these computer networks practice questions
These ten single-best-answer questions cover core computer networking concepts: addressing, common protocols, switching, routing and congestion control. They are useful for checking understanding, whether you are revising for a course assessment or strengthening your foundations. They are not tied to one institution’s exam format.
Read each question, choose an option, then check the explanation directly below it. For every miss, identify whether the issue was a forgotten fact, a calculation error or confusion between related concepts. That distinction tells you what to review next. A question answered correctly by guessing is also worth revisiting.
The questions move from straightforward facts towards applying several ideas at once. MySummaries displays each practice question with its answer and reasoning like this:
An IPv4 subnet is 192.168.10.0/26. How many usable host addresses does it contain?
A /26 leaves 32 − 26 = 6 host bits, giving 2⁶ = 64 addresses. In a conventional IPv4 subnet, the network and broadcast addresses are not assigned to hosts, so 62 are usable. The strongest distractor is 64: it counts every address, including those two reserved addresses.
A /26 fixes 26 of the 32 bits as the network prefix. For a conventional IPv4 subnet, subtract the network and broadcast addresses from the total. Do not apply that calculation uncritically to IPv6: IPv6 does not use a broadcast address.
Questions
1. IPv4 subnet capacity
Use this first question to check that you can move from CIDR notation to a usable-host count, rather than confusing total addresses with host addresses.
An IPv4 subnet is 192.168.10.0/26. How many usable host addresses does it contain?
A /26 leaves 6 host bits, so the subnet contains 2⁶ = 64 addresses. In a conventional IPv4 subnet, subtract the network and broadcast addresses: 62 usable hosts. The strongest distractor is 64, which counts all addresses rather than usable host addresses.
A /27 has 5 host bits and 32 total addresses, so 30 usable hosts; that is why 30 is a tempting but incorrect alternative here. For subnet questions, write down the host-bit count before choosing.
2. DNS resolution
This tests the difference between a client’s recursive resolver and the authoritative server for a domain.
A laptop asks its configured recursive DNS resolver for the address of www.example.net. What is the resolver expected to do if it has no cached answer?
A recursive resolver follows the resolution process on the client's behalf, querying other DNS servers as needed and returning the result. The strongest distractor is the first option: an authoritative nameserver can provide an authoritative answer, but the recursive resolver's task is to obtain and return the requested record, not merely hand back a nameserver address.
A resolver may use cached records, so it does not necessarily contact other servers for every request. DNS also does not translate names by encryption; encryption can protect some DNS transport, but it is a separate concern.
3. TCP connection setup
The sequence matters here: each message responds to the previous step in establishing a TCP connection.
Which sequence is the standard TCP three-way handshake?
The client sends SYN, the server replies with SYN-ACK, and the client completes the handshake with ACK. The strongest distractor is ACK, SYN, SYN-ACK: it contains the same kinds of flags but puts them in the wrong order. The handshake synchronises sequence numbers in both directions.
A FIN is used in connection termination, not as one of the three handshake messages. When revising packet sequences, learn both the order and what each step establishes.
4. UDP behaviour
This question checks the service UDP provides at the transport layer—and what it does not promise.
Which statement about UDP is correct?
UDP is connectionless and does not itself guarantee delivery, retransmission or in-order delivery. The strongest distractor is the first option, which describes reliability and ordering that an application may build using other mechanisms, but UDP itself does not provide.
Applications choose UDP when its properties suit their needs, or implement additional reliability at another layer. Do not infer that UDP packets always arrive out of order or are always lost; the point is that UDP does not guarantee delivery or ordering.
5. Ethernet switching
Focus on how a learning switch builds its forwarding table when it receives a frame.
A learning Ethernet switch receives a frame on port 4. Which information does it use to learn that a source MAC address is reachable through port 4?
A learning switch records the source MAC address and the port on which the frame arrived. It can then use the destination MAC address of a later frame to select a forwarding port. The strongest distractor is source IP plus incoming port: ordinary Ethernet switch learning is based on MAC addresses, not IP addresses.
If the destination MAC address is unknown, a switch may flood the frame within the relevant VLAN. That forwarding decision is distinct from learning, which uses the source address of the received frame.
6. IP route selection
This tests the longest-prefix-match rule used when a router has more than one matching route.
A router has routes for 10.0.0.0/8 and 10.20.0.0/16. Which route is selected for a packet destined for 10.20.5.7, assuming both routes are otherwise eligible?
Both prefixes match 10.20.5.7, but /16 is longer and more specific than /8, so longest-prefix match selects 10.20.0.0/16. The strongest distractor is the /8 route: it does cover the destination, but a less-specific match loses to a more-specific eligible route.
The destination does not need to equal the network address; it needs to fall within the route’s prefix. A default route is the least-specific route and is used only when no more-specific eligible route applies.
7. IPv6 subnetting
This checks a common IPv6 subnet convention and the distinction between the network prefix and interface identifier.
In a typical IPv6 subnet configured for SLAAC, what prefix length is used for the subnet?
A typical IPv6 subnet configured for Stateless Address Autoconfiguration (SLAAC) uses a /64 prefix. The strongest distractor is /128, which identifies a single IPv6 address rather than the usual subnet prefix for SLAAC. /48 and /56 can be allocation sizes from a provider, not the individual SLAAC subnet length.
Keep allocation prefixes and subnet prefixes separate in your notes. An organisation may receive a larger block and divide it into /64 subnets; the provider allocation length is not the same thing as the prefix length of each subnet.
8. HTTP conditional requests
The useful clue is that the client already has a cached representation and is asking whether it remains valid.
A client sends a conditional GET with an If-None-Match header. The selected representation has not changed. Which HTTP status code can the server return?
A 304 Not Modified response tells the client that its cached representation can be reused for this conditional request. The strongest distractor is 200 OK: a server can return a full representation with 200, but 304 is the specific response when the condition indicates no change.
A 304 response is associated with conditional retrieval; it is not a general statement that a URL is valid or that every request has succeeded. Read the condition in the stem before selecting a status code.
9. Network address translation
This question focuses on how port translation allows multiple private hosts to share one public IPv4 address.
A home router uses one public IPv4 address for outbound connections from several private hosts. What lets it distinguish their return traffic when using port address translation?
Port address translation tracks mappings that include transport-layer ports, allowing return traffic to be associated with the relevant internal connection. The strongest distractor is a separate public address for every packet: that would not explain how several private hosts share one public address.
NAT changes address information and, with port address translation, uses transport ports as part of its connection state. It does not make private addresses globally routable, and it is not a replacement for DNS.
10. TCP slow start
The final question asks about the approximate growth pattern of TCP’s congestion window during slow start, not about a fixed byte count.
In the usual simplified description of TCP slow start, how does the congestion window grow while slow start continues and no loss or other congestion signal intervenes?
During slow start, the congestion window grows by roughly one MSS per acknowledgement, which is approximately exponential growth—often described as doubling each round-trip time. The strongest distractor is one MSS per round-trip time; that is closer to the simplified linear-growth description associated with congestion avoidance, not slow start.
The doubling description is approximate: the increase occurs as acknowledgements arrive, and real behaviour depends on the algorithm and network events. In an explanation, state the pattern and connect it to the per-acknowledgement increase rather than treating doubling as an exact rule under every condition.
Review the mistakes, not just the score
After working through the set, group errors by concept. A subnetting error calls for more address calculations; mixing up DNS roles calls for a resolver-versus-authoritative-server diagram; confusing TCP and UDP calls for comparing what each transport does and does not guarantee. Then try a fresh question on the weak area. Repeating the same item from memory can hide a gap in understanding.
A sample weak-area view might rank topics by recent performance and how often they have been tested:
If the same idea is missed twice, turn the specific correction into a short retrieval prompt—for example, asking which route wins when two eligible prefixes match. A remediation prompt should target the reason for the error, not merely repeat the original question.
A prompt waiting for another attempt could look like this:
You lost this mark twice: when two eligible routes match a destination, which one does longest-prefix match select?
How MySummaries helps
Build a revision board from your networking notes, then generate question practice and flashcards from the material you uploaded. Use repeated misses to choose what to revisit next, rather than treating a total score as the whole diagnosis. Explore MySummaries.