Read a sample: 1. Reading Neighbor Discovery as a Conversation
A packet capture can look like a collection of isolated facts: an IPv6 source, a multicast destination, an ICMPv6 type, and a target address. The exam challenge is to turn those facts into a conversation. Who is asking a question? Who is expected to answer? What is the question actually about?
IPv6 Neighbor Discovery provides the messages for several conversations that take place on a local link. A host can ask whether routers are present. It can seek the link-layer address associated with an IPv6 next hop. Before assigning an address to an interface, it can also check whether another node is already using that address.
These exchanges share a small set of message types, but they do not all have the same purpose. A Neighbor Solicitation, for example, can appear in ordinary address resolution or in duplicate address detection. The message name alone does not settle which conversation you are reading. The source address, destination address, target address, and surrounding packets provide the context.
This chapter establishes a method you can use throughout the book: identify the message, read its addresses by role, connect it to nearby packets, and limit your conclusion to what the evidence supports.
Neighbor Discovery is a set of functions carried by ICMPv6 messages. Its scope includes discovering routers, learning link parameters and prefixes, resolving a neighbor’s IPv6 address to a link-layer address, tracking neighbor reachability, and receiving redirect information. This book concentrates on three exam-oriented exchanges:
- Router discovery: a host learns about routers on its link.
- Address resolution: a node obtains the link-layer address needed to send traffic to an IPv6 neighbor.
- Duplicate address detection, or DAD: a node checks whether a tentative address is already in use on the link.
The emphasis on the link matters. Neighbor Discovery messages are intended for communication among nodes on the same link, not for a conversation carried across ordinary routed hops. A host may ultimately want to reach a remote server, but its Neighbor Discovery exchange for that traffic concerns a local next hop, typically a default router.
Consider a host sending to an off-link destination. The IPv6 packet’s final destination remains the remote host. To transmit the packet on Ethernet, however, the sender needs a suitable next-hop link-layer address. If the selected next hop is a router and the necessary mapping is not already available, the host resolves the router’s IPv6 address—not the remote server’s address.
That distinction separates two questions that are easy to blur:
1. Where is the IPv6 packet ultimately going?
2. Which neighbor receives the next frame on this link?
In a Neighbor Discovery trace, the target of address resolution answers the second question.
Neighbor Discovery uses five ICMPv6 message types. Learn their names and numbers as a vocabulary for reading exchanges, but connect each one to its job.
The first two names form an obvious pair: a host may send a Router Solicitation, and a router may respond with a Router Advertisement. Routers can also send advertisements without a preceding solicitation. Therefore, an RA in a trace does not prove that an RS was sent.
The next pair requires more care. An NS identifies a target address, but the reason for asking about that target depends on the exchange. A node resolving an address wants to learn how to reach a neighbor. A node performing DAD is testing a tentative address before using it. An NA can supply information in response to a solicitation, but it can also appear without an immediately preceding NS in the visible trace.
A Redirect has a different role. It is not the normal answer to an NS and does not serve as a Router Advertisement. Recognizing it keeps an unfamiliar type 137 packet from being forced into the wrong conversation.
The examples use short, labeled lines rather than full protocol-analyzer output. A typical line looks like this:
Host A → solicited-node multicast: NS | Src: 2001:db8:1::10 | Dst: ff02::1:ff00:20 | Target: 2001:db8:1::20
Read the labels in order:
- The name before the arrow identifies the sender as described by the scenario.
- The text after the arrow describes the intended recipient or destination category.
- The message abbreviation identifies the ICMPv6 message.
- Src and Dst are the IPv6 source and destination addresses.
- Target is the target IPv6 address carried inside an NS or NA.
- Additional labels identify options, flags, or other fields when they matter.
The arrow is a reading aid, not an additional packet field. The address after Dst is the IPv6 destination; it may be unicast or multicast. A multicast destination identifies a group of potential receivers, not a single named device.
The target address is also distinct from the IPv6 destination. In the example, the NS is sent to ff02::1:ff00:20 and asks about 2001:db8:1::20. The target is not required to be the packet’s destination address. Confusing those two fields can turn a straightforward solicitation into a misleading one.
For compactness, an example may omit fields that are not needed for the question. Omission means the field or option is not shown; it does not automatically mean that the field or option was absent on the wire. When absence is important, the scenario will say so. Likewise, a short trace may omit packets that occurred before or after the lines displayed.
IPv6 addresses beginning with 2001:db8::/32 are used as documentation examples. Names such as Host A…