ICMPv6

from Wikipedia, the free encyclopedia
ICMPv6 (Internet Control Message Protocol Version 6)
Family: Internet protocol family
Operation area: Mandatory addition to IPv6 , error messages, diagnostics, autoconfiguration, routing
Internet protocols on the TCP / IP protocol stack
Internet ICMPv6
IPv6
Network access Ethernet Token
bus 		IEEE
802.11a / b / g / n 		FDDI ...
Standards:

RFC 4443 ( 2006 )

The Internet Control Message Protocol for the Internet Protocol Version 6 ( ICMPv6 ) is the version of the Internet Control Message Protocol used together with IPv6 . As with IPv4 , it is used in networks to exchange error and information messages. It is also used in the Neighbor Discovery Protocol , the replacement of the Address Resolution Protocol .

In contrast to ICMP with IPv4, ICMPv6 is absolutely necessary for the operation of IPv6. A general blocking of ICMPv6 on the firewall means that IPv6 does not work (see RFC 4890 ).

ICMPv6 serves as an auxiliary protocol for IPv6, is located in the same OSI layer 3 as this and uses the IPv6 protocol to send ICMP messages. 58 is inserted as the protocol number in the next header field of the IPv6 header.

ICMPv6 header

ICMPv6 header
+ Bits 0-7 Bits 8-15 Bits 16-23 Bits 24-31
0 Type code Checksum
... ICMPv6 message ...

The Type field indicates the class of the ICMP message, which can be specified more precisely with the Code field . The checksum is used to verify the validity of the ICMPv6 packet. The rest of the content of the ICMP message is determined by its type. In the case of error messages, as much of the error-causing package as possible is appended after the possible additional fields.

ICMPv6 types

The message types are divided into two groups. The first 128 types (0-127) with the most significant bit (engl. Most significant bit ) to 0, are error messages. The second 128 types (128-255), with the most significant bit at 1, are informational messages.

Error messages
Type description RFC
1 Destination unreachable RFC 4443
2 Packet Too Big RFC 4443
3 Time Exceeded RFC 4443
4th Parameter problem RFC 4443
100 Private experimentation
101 Private experimentation
Information messages
Type description RFC
128 Echo request RFC 4443
129 Echo Reply RFC 4443
130 Multicast Listener Query RFC 2710 and RFC 3810
131 Version 1 Multicast Listener Report RFC 2710
132 Multicast Listener Done RFC 2710
133 Router solicitation RFC 4861
134 Router Advertisement RFC 4861
135 Neighbor solicitation RFC 4861
136 Neighbor Advertisement RFC 4861
137 Redirect RFC 4861
138 Router renumbering
139 ICMP Node Information Query RFC 4620
140 ICMP Node Information Response RFC 4620
141 Inverse Neighbor Discovery Solicitation Message RFC 3122
142 Inverse Neighbor Discovery Advertisement Message RFC 3122
143 Version 2 Multicast Listener Report RFC 3810
144 Home Agent Address Discovery Request Message RFC 3775
145 Home Agent Address Discovery Reply Message RFC 3775
146 Mobile Prefix Solicitation RFC 3775
147 Mobile Prefix Advertisement RFC 3775
148 Certification Path Solicitation Message RFC 3971
149 Certification Path Advertisement Message RFC 3971
150 ICMP messages utilized by experimental mobility protocols such as Seamoby RFC 4065
151 Multicast Router Advertisement RFC 4286
152 Multicast router solicitation RFC 4286
153 Multicast router termination RFC 4286
155 RPL control message RFC 6550
200 Private experimentation
201 Private experimentation
255 Reserved for expansion of ICMPv6 informational messages

Checksum

Checksum scheme
+ Bits 0-7 Bits 8-15 Bits 16-23 Bits 24-31
0 IPv6 sender address
32
64
96
128 IPv6 destination address
160
192
224
256 IPv6 payload size
288 Check sum 0 Next header 58

The checksum (engl. Checksum ) of an ICMPv6 packet is a 16-bit one's complement of the one's complement sum of the entire ICMPv6 message. In addition to the message, an IPv6 pseudo header is appended to the front. The checksum field is set to 0 to calculate the checksum. The pseudo header used to calculate the checksum looks like the diagram next to it.

This is one of the innovations of ICMPv6 compared to ICMP , where the checksum was only calculated via the ICMP header.

ICMPv6 processing

The following rules apply to the processing of ICMPv6 messages:

  • Unknown ICMPv6 error messages must be passed on to the network layer above.
  • Unknown ICMPv6 informational messages must be discarded without notifying the sender.
  • Each error message is appended to the end of as much as possible of the package that caused the error.
  • The protocol number for forwarding unknown error messages is taken from the attached original package.
  • No error messages are sent on the following packages:
    • Error messages
    • Packets to multicast, link-level multicast or link-level broadcast addresses with the following exceptions:
      • Packet Too Big Messages
      • Parameter problem messages with code 2 - unknown IPv6 option
  • The network must not be flooded with ICMPv6 error messages.

ICMP standard types

Destination Unreachable - Type 1

Destination-unreachable scheme
+ Bits 0-7 Bits 8-15 Bits 16-23 Bits 24-31
0 Type code Checksum
32 Unused
... Bad package

Destination-unreachable messages should be generated by the router if a packet could not be delivered. If the package was dropped due to overload, no Destination Unreachable needs to be sent.

If the package was not delivered due to missing routes, code 0 is set. If delivery is administratively prohibited ( firewall ), code 1 is set. If the router cannot resolve the IPv6 address or has a problem with the link, code 3 is set. If a destination host does not have a listener for a UDP packet, it should send a Destination Unreachable with Code 4.

If a Destination Unreachable is received, it must be passed on to the layer above.

Packet Too Big - Type 2

Packet-Too-Big-Scheme
+ Bits 0-7 Bits 8-15 Bits 16-23 Bits 24-31
0 Type code Checksum
32 MTU
... Bad package

A packet too big message has to be generated by the router if a packet cannot be forwarded because it is larger than the maximum MTU of the link over which it is to be sent. Packet-Too-Big messages are used by Path MTU Discovery to determine the path-dependent MTU.

The code should be set to 0 by the sender and ignored by the receiver.

When a packet too big is received, it has to be passed on to the layer above.

Time Exceeded - Type 3

Time-Exceeded Scheme
+ Bits 0-7 Bits 8-15 Bits 16-23 Bits 24-31
0 Type code Checksum
32 Unused
... Bad package

If a router receives a packet with a hop limit of 0 or reduces the time-to-live value to 0, it must discard the packet and send a Time Exceeded with Code 0 to the sender. This indicates either an infinite loop in the routing or an initial hop limit that is too small.

If not all fragments of a fragmented message arrive within a certain time, the packet is discarded and a Time Exceeded with Code 1 must be sent.

Parameter problem - Type 4

Parameter problem scheme
+ Bits 0-7 Bits 8-15 Bits 16-23 Bits 24-31
0 Type code Checksum
32 pointer
... Bad package

If a host detects a problem in a field while processing an IPv6 packet and cannot continue processing, it must discard the packet and send a parameter problem message.

The code describes the nature of the problem in more detail.

0 Bad header field found
1 Unknown next header type found
2 Unknown IPv6 option
3 Incomplete IPv6 header chain in the first IPv6 fragment

The pointer points to the point in the packet where the problem occurred.

Echo Request - Type 128

Echo request scheme
+ Bits 0-7 Bits 8-15 Bits 16-23 Bits 24-31
0 Type code Checksum
32 ID Sequence number
... Data

A response is requested with an echo request . An echo request is nothing more than a simple ping . The data field can be enlarged with data to produce larger packages. For example, you can determine the MTU .

Each system needs to RFC on Echo Request s react with Echo Replies respond. Every system should also have an application for sending and receiving echo requests / replies . In practice, however, this is often deviated from, for example the Windows firewall blocks ICMPv6 echo request requests by default.

Received echo requests can be forwarded to applications that listen for ICMP messages.

Echo Reply - Type 129

Echo reply scheme
+ Bits 0-7 Bits 8-15 Bits 16-23 Bits 24-31
0 Type code Checksum
32 ID Sequence number
... Data

An echo request message must be responded to with an echo reply . The package is the same except for the type field. Echo reply messages should only be sent to unicast addresses.

Using the identification and the sequence number, the recipient will be able to assign the answers to his inquiries.

Received echo reply messages must be passed on to the application that sent the associated echo request . It can be passed on to the rest of the applications listening for ICMP.

Multicast Listener Discovery - Type 130

MLD is the implementation of IGMP (IPv4) in IPv6. So it is used to manage multicast subscriptions. MLDv1 corresponds to IGMPv2 and MLDv2 corresponds to IGMPv3 . With the newer versions, you can determine which source addresses are acceptable for multicast streams. Linux has supported it since 2003 (2.5.68), Windows since 2006 (Vista), FreeBSD since 2009 (8.0)

Web links

  • RFC 4861 - Neighbor Discovery for IP Version 6 (IPv6)
  • RFC 4443 - Internet Control Message Protocol (ICMPv6) for the Internet Protocol Version 6 (IPv6) Specification
  • RFC 3122 - Extensions to IPv6 Neighbor Discovery for Inverse Discovery Specification
  • IANA ICMP Parameters - full list of ICMPv6 types and codes
  • RFC 4890 - Recommendations for Filtering ICMPv6 Messages in Firewalls
  • RFC 7112 - Implications of Oversized IPv6 Header Chains

Individual evidence

  1. ↑ https://tools.ietf.org/html/rfc4604
  2. ↑ https://lwn.net/Articles/29489/