Study

IPv6 addressing

12 questions. Nothing is recorded — answers are graded and shown, not stored.

Which prefix identifies IPv6 multicast?
- ff00::/16
- ff00::/8
- fe80::/10
- fc00::/7
What is the difference between :: and ::1?
- :: is loopback and ::1 is the first address on the link
- They are the same address written two ways, since leading zeros are dropped
- ::/128 is the unspecified address, used as a source before a host has one; ::1/128 is loopback
- :: is a wildcard matching any address in a rule, and ::1 is loopback
Why may :: appear only once in an IPv6 address?
- Because two occurrences would make it impossible to work out how many zero hextets belong to each run
- Because the address would then be longer than 128 bits
- Because only the leading run of zeros may be compressed
- Because a second :: is reserved to mark the boundary between the network and interface portions
Compress 2001:0db8:0000:0000:0000:ff00:0042:8329 as RFC 5952 asks.
- 2001:db8:0:0:0:ff00:42:8329
- 2001:DB8::FF00:42:8329
- 2001:0db8::ff00:0042:8329
- 2001:db8::ff00:42:8329
A site is given a /48 and assigns a /64 to each subnet. How many subnets does that allow?
- 256
- 4,096
- 65,536
- 18,446,744,073,709,551,616
Is the /64 boundary between network and interface a property of the IPv6 address format?
- Yes — the format fixes 64 bits of network and 64 of interface identifier
- No — it is a strong convention that stateless autoconfiguration requires and most tooling assumes
- Yes for global unicast, no for link-local
- No — it is an arbitrary default that can be changed on a LAN with no consequences
EUI-64 derives an interface identifier from the hardware MAC address. Why is it no longer the default?
- It produced collisions on large segments
- It is incompatible with stateless autoconfiguration
- It embeds a hardware serial number in every packet the host sends, which is a privacy problem
- It requires a 64-bit MAC address, which most hardware does not have
Why does a link-local address often need a zone index, as in fe80::1%eth0?
- Because the same fe80::/10 prefix exists on every link, so the address alone does not say which one you mean
- Because the zone index supplies the missing prefix length
- Because link-local addresses are not unique and the index disambiguates duplicates on the same segment
- Because the kernel uses it to select a routing table entry for off-link traffic
IPv6 has no broadcast address. How does a host ask which link-layer address holds a given IPv6 address?
- It sends an ARP request to the all-ones link-layer address, as IPv4 does
- It sends a neighbor solicitation to that address’s solicited-node multicast group
- It queries the router, which keeps a table of every address on the link
- It sends a neighbor solicitation to ff02::1, the all-nodes group
Duplicate address detection sends a neighbor solicitation for an address the host is about to use. What does it use as the source address, and why?
- The address being tested, so a reply can be matched to the request
- Its link-local address, which is already known to be unique
- The unspecified address ::, because the host has not committed to the address it is asking about
- The all-nodes multicast address ff02::1
Can an IPv6-only host reach an IPv4-only server without help?
- Yes — IPv6 includes a compatibility mode that maps IPv4 addresses into the IPv6 space
- Yes, provided both ends support Happy Eyeballs
- No — they are separate protocols, and something such as NAT64 with DNS64 has to translate
- No, unless the server has a global unicast address in 2000::/3
Which range fills the role for IPv6 that RFC 1918 fills for IPv4?
- fe80::/10, since link-local traffic never leaves the network
- fc00::/7, with fd00::/8 the half actually used, where a site generates a random global ID
- 2000::/3, filtered at the site boundary
- ::/128, which is unroutable by definition