Internet Control Message Protocol

Header

The ICMP header starts after the IPv4 header and is identified by its protocol number, 1.^[6] All ICMP packets have an eight-byte header and variable-sized data section. The first four bytes of the header have fixed format, while the last four bytes depend on the type and code of the ICMP packet.^[3]

Type: 8 bits

ICMP type, see § Control messages.

Code: 8 bits

ICMP subtype, see § Control messages.

Checksum: 16 bits

Internet checksum^[7] for error checking, calculated from the ICMP header and data with value 0 substituted for this field.

Rest of Header: 32 bits

Four-byte field, contents vary based on the ICMP type and code.

Data

ICMP error messages contain a data section that includes a copy of the entire IPv4 header, plus at least the first eight bytes of data from the IPv4 packet that caused the error message. The length of ICMP error messages should not exceed 576 bytes.^[1] This data is used by the host to match the message to the appropriate process. If a higher-level protocol uses port numbers, they are assumed to be in the first eight bytes of the original datagram's data.^[2]

The variable size of the ICMP packet data section has been exploited. In the "Ping of death", large or fragmented ICMP packets are used for denial-of-service attacks. ICMP data can also be used to create covert channels for communication. These channels are known as ICMP tunnels.

Source quench

Source Quench requests that the sender decrease the rate of messages sent to a router or host. This message may be generated if a router or host does not have sufficient buffer space to process the request, or may occur if the router or host buffer is approaching its limit.

Data is sent at a very high speed from a host or from several hosts at the same time to a particular router on a network. Although a router has buffering capabilities, the buffering is limited to within a specified range. The router cannot queue any more data than the capacity of the limited buffering space. Thus, if the queue gets filled up, incoming data is discarded until the queue is no longer full. But as no acknowledgement mechanism is present in the network layer, the client does not know whether the data has reached the destination successfully. Hence, some remedial measures should be taken by the network layer to avoid these kinds of situations. These measures are referred to as source quench.

In a source quench mechanism, the router sees that the incoming data rate is much faster than the outgoing data rate and sends an ICMP message to the clients, informing them that they should slow down their data transfer speeds or wait for a certain amount of time before attempting to send more data. When a client receives this message, it automatically slows down the outgoing data rate or waits for a sufficient amount of time, which enables the router to empty the queue. Thus, the source quench ICMP message acts as flow control in the network layer.

Since research suggested that "ICMP Source Quench [was] an ineffective (and unfair) antidote for congestion",^[10] routers' creation of source quench messages was deprecated in 1995 by RFC 1812. Furthermore, forwarding of and any kind of reaction to (flow control actions) source quench messages was deprecated from 2012 by RFC 6633.

Where:

• Type must be set to 4
• Code must be set to 0
• IP header and additional data is used by the sender to match the reply with the associated request

Redirect

Redirect requests data packets be sent on an alternative route. ICMP Redirect is a mechanism for routers to convey routing information to hosts. The message informs a host to update its routing information (to send packets on an alternative route). If a host tries to send data through a router (R1) and R1 sends the data on another router (R2) and a direct path from the host to R2 is available (that is, the host and R2 are on the same subnetwork), then R1 will send a redirect message to inform the host that the best route for the destination is via R2. The host should then change its route information and send packets for that destination directly to R2. The router will still send the original datagram to the intended destination.^[15] However, if the datagram contains routing information, this message will not be sent even if a better route is available. RFC 1122 states that redirects should only be sent by gateways and should not be sent by Internet hosts.

Where:

• Type must be set to 5.
• Code specifies the reason for the redirection, and may be one of the following:

  Code	Description
  0	Redirect for Network
  1	Redirect for Host
  2	Redirect for Type of Service and Network
  3	Redirect for Type of Service and Host

• IP address is the 32-bit address of the gateway to which the redirection should be sent.
• IP header and additional data is included to allow the host to match the reply with the request that caused the redirection reply.

Time exceeded

Time Exceeded is generated by a gateway to inform the source of a discarded datagram due to the time to live field reaching zero. A time exceeded message may also be sent by a host if it fails to reassemble a fragmented datagram within its time limit.

Time exceeded messages are used by the traceroute utility to identify gateways on the path between two hosts.

Where:

• Type must be set to 11
• Code specifies the reason for the time exceeded message, include the following:

  Code	Description
  0	Time-to-live exceeded in transit.
  1	Fragment reassembly time exceeded.

• IP header and first 64 bits of the original payload are used by the source host to match the time exceeded message to the discarded datagram. For higher-level protocols such as UDP and TCP the 64-bit payload will include the source and destination ports of the discarded packet.

Timestamp

Timestamp is used for time synchronization. The originating timestamp is set to the time (in milliseconds since midnight) the sender last touched the packet. The receive and transmit timestamps are not used.

Where:

• Type must be set to 13
• Code must be set to 0
• Identifier and Sequence Number can be used by the client to match the timestamp reply with the timestamp request.
• Originate timestamp is the number of milliseconds since midnight Universal Time (UT). If a UT reference is not available, the most-significant bit can be set to indicate a non-standard time value.

Timestamp reply

Timestamp Reply replies to a Timestamp message. It consists of the originating timestamp sent by the sender of the Timestamp as well as a receive timestamp indicating when the Timestamp was received and a transmit timestamp indicating when the Timestamp reply was sent.

Where:

• Type must be set to 14
• Code must be set to 0
• Identifier and Sequence number can be used by the client to match the reply with the request that caused the reply.
• Originate timestamp is the time the sender last touched the message before sending it.
• Receive timestamp is the time the echoer first touched it on receipt.
• Transmit timestamp is the time the echoer last touched the message on sending it.

All timestamps are in units of milliseconds since midnight UT. If the time is not available in milliseconds or cannot be provided with respect to midnight UT, then any time can be inserted in a timestamp, provided the high-order bit of the timestamp is also set to indicate this non-standard value.

The use of Timestamp and Timestamp Reply messages to synchronize the clocks of Internet nodes has largely been replaced by the UDP-based Network Time Protocol and the Precision Time Protocol.^[16]

Address mask request

Address mask request is normally sent by a host to a router in order to obtain an appropriate subnet mask.

Recipients should reply to this message with an Address mask reply message.

Where:

• Type must be set to 17
• Code must be set to 0
• Address mask can be set to 0

ICMP Address Mask Request may be used as a part of reconnaissance attack to gather information on the target network, therefore ICMP Address Mask Reply is disabled by default on Cisco IOS.^[17]

Address mask reply

Address mask reply is used to reply to an address mask request message with an appropriate subnet mask.

Where:

• Type must be set to 18
• Code must be set to 0
• Address mask should be set to the subnet mask

Destination unreachable

Destination unreachable is generated by the host or its inbound gateway^[2] to inform the client that the destination is unreachable for some reason. Reasons for this message may include: the physical connection to the host does not exist (distance is infinite); the indicated protocol or port is not active; the data must be fragmented but the 'don't fragment' flag is on.^[18] Unreachable TCP ports notably respond with TCP RST rather than a destination unreachable type 3 as might be expected. Destination unreachable is never reported for IP multicast transmissions.

With the following field contents:

Type: 8 bits; Type == 3

A value of 3 indicates 'Destination unreachable'.

Code: 8 bits

This specifies the type of error, and can be any of the following:^[8]

Code	Description
0	Network unreachable error.
1	Host unreachable error.
2	Protocol unreachable error (the designated transport protocol is not supported).
3	Port unreachable error (the designated protocol is unable to inform the host of the incoming message).
4	The datagram is too big. Packet fragmentation is required but the 'don't fragment' (DF) flag is on.
5	Source route failed error.
6	Destination network unknown error.
7	Destination host unknown error.
8	Source host isolated error.
9	The destination network is administratively prohibited.
10	The destination host is administratively prohibited.
11	The network is unreachable for Type Of Service.
12	The host is unreachable for Type Of Service.
13	Communication administratively prohibited (administrative filtering prevents packet from being forwarded).
14	Host precedence violation (indicates the requested precedence is not permitted for the combination of host or network and port).
15	Precedence cutoff in effect (precedence of datagram is below the level set by the network administrators).

Unused: 8 - 32 bits; Unused == 0

Unused, must be set to zero. If Length or Next-hop MTU are not used, they are considered part of this field.

Length: 8 bits

Optional. The Length field indicates the length of the original datagram data, in 32-bit words. This allows this ICMP message to be extended with extra information. If used, the original datagram data must be padded with zeroes to the nearest 32-bit boundary.

Next-hop MTU: 16 bits

Optional. Contains the MTU of the next-hop network if a code 4 error occurs.

IP header and data: 20 - 568 bytes

The IP header (20 bytes) and at most 548 bytes of the start of the original datagram (as not to exceed the minimum IPv4 reassembly buffer size). If this message is extended, then this field must contain at least 128 bytes of original datagram data (padded with zeroes if necessary).
These data are included to allow the client to match the reply with the request that caused the Destination unreachable reply.