[{"id":"networkingintermediate-001","courseId":"networking-intermediate","domain":"ip-layer","topic":"IP fragmentation fields","prompt":"IP must split a datagram so it fits a network with small packets. The receiving host must later rebuild the original datagram. Where does IP keep the information it needs to do that?","hint":"Consider which header the router reads and edits when it cuts a datagram.","explain":"Each layer wraps the data it receives in its own header. A device that cuts a datagram must mark each piece so the receiver can rebuild the whole, and the marks must sit where that layer can read them.","source":"https://www.rfc-editor.org/rfc/rfc791.html","interaction":{"type":"single"},"options":[{"id":"a","text":"Fields in the IP header","why":"The internet modules fragment and reassemble using fields in the internet header, which each piece carries."},{"id":"b","text":"The TCP header inside the data","whyNot":"The TCP header is part of the data IP carries; IP does not read it to fragment or reassemble."},{"id":"c","text":"A separate reassembly message sent before the pieces","whyNot":"IP does the job with fields in its own header, not with an extra message sent ahead of the pieces."},{"id":"d","text":"The local network's frame header","whyNot":"The local network wraps the datagram for one hop, but the fragmentation fields sit in the internet header."}],"answer":"a","explanation":"The internet protocol uses fields in its own header to fragment and reassemble datagrams for small-packet networks; a TCP segment is just data to IP.","concept":"The internet protocol uses fields in its own header to fragment and reassemble datagrams for small-packet networks; a TCP segment is just data to IP.","rev":1,"updatedAt":"2026-10-26","kind":"scenario","takeaways":["Fragmentation bookkeeping lives in the IP header, not in the TCP segment that IP carries as data."]},{"id":"networkingintermediate-002","courseId":"networking-intermediate","domain":"ip-layer","topic":"Don't Fragment flag","prompt":"A router must send a 1,500-octet datagram, with the Don't Fragment flag clear, onto a network whose MTU is 576 octets. In what order does RFC 791's example procedure build the first fragment?","hint":"Each step needs a value that an earlier step has already fixed.","explain":"Splitting a datagram is bookkeeping. Each piece needs a header that describes exactly the data it carries, and some header values depend on how much data the piece holds.","source":"https://www.rfc-editor.org/rfc/rfc791.html","interaction":{"type":"sort","items":[{"text":"Copy the original internet header","why":"The first fragment starts from a copy of the header; the fields that change are corrected later."},{"text":"Compute NFB, the number of 8-octet blocks that fit: (MTU - IHL*4)/8","why":"The block count sets how much data the piece can carry, so it must be known before data is attached."},{"text":"Attach the first NFB*8 data octets","why":"Only once the block count is known can the right amount of data be taken from the datagram."},{"text":"Set MF to 1 and TL to the header plus NFB*8 octets, then recompute the checksum","why":"The header can only describe the piece correctly once its data size and the more-fragments flag are set."},{"text":"Submit this fragment to the next step in datagram processing","why":"The piece is handed on last, once its header is correct."}]},"options":[],"answer":"o:abcde","explanation":"RFC 791's example procedure builds the first fragment by copying the header, computing how many 8-octet blocks fit, attaching that data, correcting MF and TL, then submitting it.","concept":"RFC 791's example procedure builds the first fragment by copying the header, computing how many 8-octet blocks fit, attaching that data, correcting MF and TL, then submitting it.","rev":1,"updatedAt":"2026-10-26","kind":"scenario","takeaways":["Fragment data comes in 8-octet blocks, and the header is corrected after the data is attached."]},{"id":"networkingintermediate-003","courseId":"networking-intermediate","domain":"ip-layer","topic":"ICMP Time Exceeded message","prompt":"A host is rebuilding a fragmented IPv4 datagram. Its reassembly time limit runs out with pieces still missing, and the first fragment, fragment zero, never arrived. What should the host do?","hint":"Think about what a report would need to point at, given which pieces are missing.","explain":"A receiver holding part of a split message works under a time limit. When the limit passes, it decides what to throw away and whether it can still report the failure, and the pieces it holds matter.","source":"https://www.rfc-editor.org/rfc/rfc792.html","interaction":{"type":"single"},"options":[{"id":"a","text":"Discard the pieces and send no time exceeded message","why":"RFC 792 says that if fragment zero is not available, no time exceeded message need be sent at all."},{"id":"b","text":"Discard the pieces and send a time exceeded message with code 0","whyNot":"Code 0 reports a time to live that reached zero in transit, not a reassembly timeout."},{"id":"c","text":"Keep the pieces until the missing ones arrive","whyNot":"Missing fragments within the time limit mean the host discards the datagram; it cannot wait indefinitely."},{"id":"d","text":"Send a destination unreachable message to the source","whyNot":"Destination unreachable reports a network, host, or protocol failure, not a reassembly timeout."}],"answer":"a","explanation":"A host that cannot finish reassembly within its time limit discards the datagram, and no time exceeded message is needed when fragment zero is absent.","concept":"A host that cannot finish reassembly within its time limit discards the datagram, and no time exceeded message is needed when fragment zero is absent.","rev":1,"updatedAt":"2026-10-26","kind":"scenario","takeaways":["If fragment zero never arrived, the host discards the pieces and sends no time exceeded message."]},{"id":"networkingintermediate-004","courseId":"networking-intermediate","domain":"ip-layer","topic":"Longest-prefix match forwarding","prompt":"A router holds routes for 10.0.0.0/8 and 10.1.0.0/16. For a packet to 10.1.2.3, it forwards using the 10.0.0.0/8 route, because that route matches more addresses.","hint":"Ask which of the matching routes is the most specific.","explain":"Several routes can cover one destination, since a short prefix covers a larger block that includes smaller ones. The router must pick one, and the choice depends on how the matching prefixes compare.","source":"https://www.rfc-editor.org/rfc/rfc4632.html","interaction":{"type":"true-false"},"options":[{"id":"a","text":"True","whyNot":"Matching more addresses does not decide it; forwarding picks the longest matching prefix, which is the /16 route."},{"id":"b","text":"False","why":"Forwarding is done on a longest-match basis, so the more specific 10.1.0.0/16 route is used for 10.1.2.3."}],"answer":"b","explanation":"Internet forwarding is done on a longest-match basis: when several routes cover a destination, the longest, most specific prefix is chosen.","concept":"Internet forwarding is done on a longest-match basis: when several routes cover a destination, the longest, most specific prefix is chosen.","rev":1,"updatedAt":"2026-10-26","kind":"scenario","takeaways":["A broad route still matches the address, but a more specific route covering it wins."]},{"id":"networkingintermediate-005","courseId":"networking-intermediate","domain":"ip-layer","topic":"Gratuitous ARP announcements","prompt":"A host has just configured address 192.0.2.7 and wants to announce it to every host on the link. Under RFC 5227, which ARP packet type does that announcement use?","hint":"Consider which ARP packet type the RFC prefers, and why older implementations accept it.","explain":"An announcement tells the whole link about an address that nobody asked about. Each ARP message type has a usual role, and using one unsolicited depends on what existing implementations accept.","source":"https://www.rfc-editor.org/rfc/rfc5227.html","interaction":{"type":"single"},"options":[{"id":"a","text":"ARP Request packets","why":"RFC 5227 says announcements use ARP Request packets, which are always broadcast, so every host on the link receives them."},{"id":"b","text":"ARP Reply packets","whyNot":"The name gratuitous reply sounds right, but the RFC keeps the Request form for historical and interoperability reasons."},{"id":"c","text":"Unicast ARP Reply packets to the gateway","whyNot":"A unicast reply reaches one host and normally answers a broadcast request, so it does not announce to the link."},{"id":"d","text":"ICMP echo request packets","whyNot":"ICMP echo tests reachability; an ARP announcement is an ARP packet, and the RFC names ARP Request packets."}],"answer":"a","explanation":"RFC 5227 says ARP announcements use ARP Request packets rather than ARP Reply packets, for historical precedent and better interoperability.","concept":"RFC 5227 says ARP announcements use ARP Request packets rather than ARP Reply packets, for historical precedent and better interoperability.","rev":1,"updatedAt":"2026-10-26","kind":"scenario","takeaways":["Gratuitous ARP is sent as an ARP Request even though it answers no question."]}]