Network Working Group A. Conta (Digital Equipment Corporation)
INTERNET-DRAFT S. Deering (Xerox PARC)
November 1995
Generic Packet Tunneling in IPv6
Specification
draft-ietf-ipngwg-ipv6-tunnel-00.txt
Status of this Memo
This document is an Internet Draft. Internet Drafts are working
documents of the Internet Engineering Task Force (IETF), its Areas,
and its Working Groups. Note that other groups may also distribute
working documents as Internet Drafts.
Internet Drafts are draft documents valid for a maximum of six
months. Internet Drafts may be updated, replaced, or obsoleted by
other documents at any time. It is not appropriate to use Internet
Drafts as reference material or to cite them other than as a "working
draft" or "work in progress."
To learn the current status of any Internet-Draft, please check the
``1id-abstracts.txt'' listing contained in the Internet- Drafts
Shadow Directories on ds.internic.net (US East Coast), nic.nordu.net
(Europe), ftp.isi.edu (US West Coast), or munnari.oz.au (Pacific
Rim).
Distribution of this memo is unlimited.
Abstract
This document defines the model and mechanisms for IPv6 tunneling of
various Internet or lower layer protocol packets, such as IPv6, IPv4,
AppleTalk, IPX, etc.
Conta & Deering Expires in six months [Page 1]
INTERNET-DRAFT Tunneling in IPv6 November 22, 1995
Table of Contents
Status of this Memo.................................1
Table of Contents...................................2
1. Introduction........................................3
2. Terminology.........................................3
3. IPv6 Tunneling......................................6
3.1 IPv6 Encapsulation.............................8
3.2 IPv6 Decapsulation.............................9
3.3 IPv6 Tunnel Protocol Engine...................10
4. Recursive Encapsulation............................12
4.1 Limiting Recursive Encapsulation.............12
4.1.1 Tunnel Encapsulation Limit.............13
4.1.2 Loopback Recursive Encapsulation.......14
4.1.3 Routing Loop Recursive Encapsulation...15
5. Tunnel IPv6 Header.................................16
5.1 Tunnel IPv6 Extension Headers.................17
6. IPv6 Tunnel State Variables........................19
6.1 IPv6 Tunnel Entry-Point Node..................19
6.2 IPv6 Tunnel Exit-Point Node...................19
6.3 IPv6 Tunnel Hop Limit.........................20
6.4 IPv6 Tunnel Packet Priority...................20
6.5 IPv6 Tunnel Flow Label........................21
6.6 IPv6 Tunnel Encapsulation Limit...............21
6.7 IPv6 Tunnel MTU...............................21
7. IPv6 Tunnel Error Reporting and Processing.........23
7.1 Tunnel ICMP Messages..........................25
7.2 ICMP Messages for IPv6 Original Packets.......26
7.3 ICMP Messages for IPv4 Original Packets.......27
8. Open Issues........................................28
9. References.........................................28
10. Acknowledgements..................................29
11. Security Considerations...........................29
Authors' Addresses....................................30
Fig. 1.................................................7
Fig. 2.................................................8
Fig. 3.................................................8
Fig. 4.................................................9
Fig. 5................................................10
Fig. 6................................................12
Fig. 7................................................23
Fig. 8................................................25
Conta & Deering Expires in six months [Page 2]
INTERNET-DRAFT Tunneling in IPv6 November 22, 1995
1. Introduction
This document specifies a method and generic mechanisms by which a
packet may be encapsulated and carried as payload within an IPv6
packet. The technique of encapsulating packets within IPv6 packets,
called also tunneling in IPv6, is recommended for use in various
cases of communications.
Such a case is when a node, that is not the source node of a packet,
wishes to exert explicit routing control over the packet - such as
causing the packet to be forwarded to a particular destination on the
way to the final destination identified in the original header. The
control is exerted by prepending to the packet a new IPv6 header,
with a new source and destination address (see section 3.1).
The tunnel IPv6 packet is forwarded to the tunnel IPv6 header
destination which is the IPv6 tunnel exit-point node. The IPv6 tunnel
exit-point node removes the IPv6 tunnel header, and forwards the IPv6
packet further towards the original IPv6 header destination node (see
section 3.2).
The sections of this document describe generic mechanisms for IPv6
tunneling that apply to tunneling of various Internet or lower layer
protocol packets. section 7.2, and 7.3 are an exception; they
describe a specific IPv6 tunneling mechanism for IPv6 packets and
respectively IPv4 packets.
2. Terminology
node
a device that implements IPv6.
upper-layer
a protocol layer immediately above IPv6. Examples are transport
protocols such as TCP and UDP, control protocols such as ICMP, and
internet or lower-layer protocols being "tunneled" over (i.e.,
encapsulated in) IPv6 such as IPX, AppleTalk, or IPv6 itself.
link
a communication facility or medium over which nodes can
communicate at the link layer, i.e., the layer immediately below
IPv6. Examples are Ethernets (simple or bridged); PPP links;
Conta & Deering Expires in six months [Page 3]
INTERNET-DRAFT Tunneling in IPv6 November 22, 1995
X.25, Frame Relay, or ATM networks; and internet (or higher) layer
"tunnels", such as tunnels over IPv4 or IPv6 itself.
interface
a node's attachment to a link.
address
an IPv6-layer identifier for an interface or a set of interfaces.
packet
an IPv6 header plus payload.
link MTU
the maximum transmission unit, i.e., maximum packet size in
octets, that can be conveyed in one piece over a link.
path MTU
the minimum link MTU of all the links in a path between a source
node and a destination node.
original packet
an Internet or lower layer packet that undergoes encapsulation in
a tunnel packet.
original header
the header of an original packet.
tunnel
a virtual link between two nodes, on which an Internet or lower
layer protocol packet travels after the entry-point node
encapsulates that packet in a tunnel protocol packet.
Conta & Deering Expires in six months [Page 4]
INTERNET-DRAFT Tunneling in IPv6 November 22, 1995
tunnel header
the header prepended to the original packet during encapsulation.
It specifies the tunnel end-points as source and destination.
tunnel packet
an original packet encapsulated by prepending tunnel header(s) to
the original header(s).
tunnel entry-point
the tunnel end-node where an original packet is encapsulated, and
where a tunnel packet originates; the source node of a tunnel
packet identified in the tunnel header.
tunnel exit-point
the tunnel end-node where a tunnel packet is decapsulated, and
processed further as original packet based on the original header;
the destination node of a tunnel packet, identified in the tunnel
header.
tunnel MTU
the Path MTU in the tunnel, i.e. between the tunnel entry-point
node and the tunnel exit-point node.
tunnel hop limit
the maximum number of hops that a tunnel packet is allowed to
travel in a tunnel, i.e. between the tunnel entry-point node and
the tunnel exit-point node.
inner tunnel
a tunnel which serves as one (virtual) hop in another tunnel.
outer tunnel
a tunnel in which one or more hops are themselves tunnels.
Conta & Deering Expires in six months [Page 5]
INTERNET-DRAFT Tunneling in IPv6 November 22, 1995
recursive tunnel IPv6 packet
a tunnel IPv6 packet encapsulated within a tunnel IPv6 packet, by
prepending IPv6 tunnel header(s) to previously prepended IPv6
tunnel header(s).
recursive encapsulation
encapsulation of a tunnel packet, i.e. encapsulation of an
encapsulated packet.
tunnel encapsulation limit
the maximum number of recursive encapsulations of a packet.
3. IPv6 Tunneling
Tunneling in IPv6 is a technique which consists in establishing a
"virtual link" between two IPv6 nodes and using that "link" for
transmitting data packets. The two IPv6 nodes view the IPv6 "virtual
link", also called an "IPv6 tunnel", as a "link", on which the IPv6
protocol acts like a "link layer" protocol. Consequently the two
nodes at the two ends of the IPv6 tunnel exchange an Internet or
lower layer protocol packet by encapsulating in and decapsulating
from an IPv6 packet, as they would encapsulate in and decapsulate
from a link layer protocol packet.
The two IPv6 nodes which are at the two ends of the IPv6 tunnel, the
IPv6 tunnel entry point node and the IPv6 tunnel exit point node play
specific roles:
The tunnel entry-point node encapsulates an original packet within an
IPv6 packet by prepending an IPv6 header (and optionally a set of
IPv6 extension headers) to the original packet and then transmits the
resulting tunnel packet towards the tunnel exit-point node. The
tunnel headers (IPv6 header and IPv6 extension headers) are used to
control the packet's processing (forwarding) through the tunnel (see
section 3.1).
The tunnel exit-point node decapsulates a tunnel packet and then it
processes further the resulting original packet like any original
packet (see section 3.2).
Conta & Deering Expires in six months [Page 6]
INTERNET-DRAFT Tunneling in IPv6 November 22, 1995
A tunnel entry point node can be seen as an original packet source -
the source of the IPv6 tunnel packet is the tunnel entry point node.
An IPv6 tunnel main header and its extension headers, if any, are
processed by the IPv6 layer similarly to processing the headers of an
original IPv6 packet. Additionally, an IPv6 tunnel packet resulting
from encapsulation is an IPv6 packet that may be the object of a
subsequent IPv6 encapsulation, similar to any original packet.
A tunnel exit point node can be seen as an original packet
destination - the destination of the IPv6 tunnel packet is the tunnel
exit-point node. The tunnel exit point node processes the IPv6 main
header and its extension headers, if any, of an IPv6 tunnel packet
destined to it similarly to processing the IPv6 headers of an
original packet destined to it.
Tunnel from node B to node C
<------------------------------------->
Tunnel Tunnel
Entry-Point Exit-Point
Node Node
+-+ +-+ +-+ +-+
|A|->-//->-|B|=========>=====//========>=========|C|->-//-->-|D|
+-+ +-+ +-+ +-+
Original Original
Packet Packet
Source Destination
Node Node
Fig. 1. Tunnel
An IPv6 tunnel is a unidirectional mechanism - tunnel packet flow
takes place in one direction between the IPv6 tunnel entry point node
and exit point node (see Fig. 1).
A bidirectional tunneling mechanism effect can be achieved by merging
two unidirectional mechanisms, by defining two tunnels that are each
one in opposite direction to the other - the entry point node of one
tunnel is the exit point node of the other tunnel (see Fig. 2).
A tunnel entry-point node can be the original packet source, and
similarly a tunnel exit-point node can be the original packet
destination, i.e. the beginning or ending of a tunnel can coincide
with the source or destination of an original packet.
Conta & Deering Expires in six months [Page 7]
INTERNET-DRAFT Tunneling in IPv6 November 22, 1995
Tunnel from node B to node C
<--------------------------------------->
Tunnel Tunnel
Original Entry-Point Exit-Point Original
Packet Node Node Packet
Source Destination
Node Node
+-+ +-+ +-+ +-+
| |-->-//->--| |=========>=====//========>=========| |-->-//-->-| |
|A| |B| |C| |D|
| |--<-//-<--| |=========<=====//========<=========| |--<-//--<-| |
+-+ +-+ +-+ +-+
Original Original
Packet Packet
Destination Tunnel Tunnel Source
Node Exit-Point Entry-Point Node
Node Node
<------------------------------------->
Tunnel from node C to node B
Fig. 2. Bidirectional tunneling mechanism effect
3.1 IPv6 Encapsulation
The IPv6 encapsulation of a packet consists in prepending to the
original packet, an IPv6 header and optionally a set of IPv6
extension headers, called tunnel IPv6 headers, as graphically shown
below in Fig. 3:
+----------------------------------//-----+
| Original | |
| | Original Packet Payload |
| Headers | |
+----------------------------------//-----+
< Original packet >
|
v
<Tunnel IPv6 headers> < Original Packet >
+---------+ - - - - - +-------------------------//--------------+
| IPv6 | IPv6 | |
| | Extension | Original Packet |
| Header | Headers | |
+---------+ - - - - - +-------------------------//--------------+
< Tunnel IPv6 packet >
Fig. 3. Encapsulating a packet
Conta & Deering Expires in six months [Page 8]
INTERNET-DRAFT Tunneling in IPv6 November 22, 1995
The IPv6 encapsulation takes place in an IPv6 tunnel entry-point
node, when transmitting an original packet through the tunnel that
begins at that entry-point node. If the transmitting of the original
packet through the tunnel is the result of a forwarding operation,
the original packet header is processed before encapsulation
according to the forwarding rules. For instance if the original
packet is an:
(a) IPv6 packet, and it is tunneled as result of an IPv6 forwarding
operation, the IPv6 original packet hop limit is decremented by
one during forwarding.
(b) IPv4 packet, and it is tunneled as result of an IPv4 forwarding
operation through an IPv6 tunnel, the IPv4 original packet time
to live field (TTL) is decremented by one during forwarding.
3.2 IPv6 Decapsulation
The decapsulation is graphically shown below in Fig. 4:
+---------+- - - - - -+----------------------------------//-----+
| IPv6 | IPv6 | |
| | Extension | Original Packet |
| Headers | Headers | |
+---------+- - - - - -+----------------------------------//-----+
< Tunnel IPv6 packet >
|
v
+----------------------------------//-----+
| Original | |
| | Original Packet Payload |
| Headers | |
+----------------------------------//-----+
< Original packet >
Fig. 4. Decapsulating a packet from the tunnel IPv6 headers
The IPv6 protocol layer of a tunnel exit-point node receiving an IPv6
packet destined to one of the node's IPv6 addresses processes the
packet according to the IPv6 protocol. A Next Header field value set
to a Tunnel Protocol Value (value 41 for IPv6) in the IPv6 header, or
the last IPv6 extension header of the packet identifies the packet as
a tunnel packet. Upon the completion of the IPv6 protocol layer
processing of a tunnel packet, control is given to the Tunnel
Conta & Deering Expires in six months [Page 9]
INTERNET-DRAFT Tunneling in IPv6 November 22, 1995
Protocol layer. The Tunnel Protocol layer discards the tunnel headers
and passes the resulting original packet to the Internet or lower
layer protocol for further processing - for Next Header value 41, the
resulting original packet is passed to the IPv6 protocol layer.
3.3 IPv6 Tunnel Protocol Engine
The packet flow through the IPv6 Tunnel Protocol Engine is
graphically shown below in Fig. 5:
+-----------------------+ +-----------------------------------+
| Upper Layer Protocols | | IPv6 Tunnel Upper-Layer |
| | | |
| | | ---<-------------------<------- |
| | | | ---->---|------>--------- | |
| | | | | | | | | |
+-----------------------+ +-----------------------+ | | |
| | | | | | | | | v ^ |
v ^ v ^ v ^ v ^ Tunnel | | | |
| | | | | | | | packets| | | |
+---------------------------------------------+ | | | |
| | | | | / / | | | | D E |
| v ^ IPv6 | --<-3--/-/--<---- | | | | E N |
| | | Layer ---->-4-/-/--->-- | | | | | C C |
| v ^ / / | | | | | | A A |
| | | 2 1 | | | | | | P P |
| v ^ -----<---5---/-/-<---- v ^ v ^ | | S S |
| | | | -->---6---/-/-->-- | | | | | | | U U |
| v ^ | | / / 6 5 4 3 8 7 | | L L |
| | | | | / / | | | | | | | | A A |
| v ^ v ^ / / v ^ | | | | | | T T |
+---------------------------------------------+ | E E |
| | | | | | | | | | | | | | | |
v ^ v ^ v ^ v ^ v ^ v ^ Original| | | |
| | | | | | | | | | | | packets | v ^ |
+-----------------------+ +-----------------------+ | | |
| | | | | | | | | | | |
| | | | ---|----|-------<-------- | |
| | | --->--------------->------>---- |
| | | |
| Link Layer Protocols | | IPv6 Tunnel Link Layer |
+-----------------------+ +-----------------------------------+
Fig. 5. Packet Flow - IPv6 Tunneling Protocol Engine
The "tunnel-layer" acts as both an "upper-layer" and a "link layer":
Conta & Deering Expires in six months [Page 10]
INTERNET-DRAFT Tunneling in IPv6 November 22, 1995
(a) The "tunnel upper-layer" has as "input" tunnel IPv6 packets
which are going to be decapsulated. The tunnel packets are
incoming through the IPv6 layer from a link-layer - (path #1 in
Fig. 5) - or from a tunneling link-layer (the exit-point node
of an inner tunnel coincides with the exit-point node of an
outer tunnel) - (path #7 in Fig. 5). The resulting original
packets are passed back to the IPv6 layer as "tunnel link-
layer" output for further processing - see (d).
(b) The "tunnel upper-layer" has as "output" tunnel IPv6 packets
resulting from encapsulation of original packets - see (c) - or
packets resulted from previous encapsulations - when this node
is an entry-point to an outer tunnel and to an inner tunnel.
The "output" tunnel packets are passed through the IPv6 layer
down to a link-layer for transmission - (path #2 Fig. 5) - or
to a tunnel link-layer for another encapsulation (the entry-
point node in an inner tunnel coincides with the entry-point
node in an outer tunnel) - (path #8 in Fig. 5).
Implementation Note:
the "tunnel upper-layer" input and output may be implemented
similar to the other upper layer protocols input and output.
(c) The "tunnel link-layer" has as "input" original IPv6 packets
which are going to be encapsulated. The original packets are
incoming through the IPv6 layer from an upper-layer (original
packets originated on this node) - (path #4 in Fig. 5) - or
from a link layer (original packets originated on a different
node) - (path #6 in Fig. 5). The resulting tunnel packets are
passed through the IPv6 layer down to a link-layer for
transmission - see (b).
(d) The "tunnel link-layer" has as "output" original IPv6 packets
resulting from decapsulation - see (a) - which are passed
through the IPv6 layer up to an upper-layer (the original
packet is destined to this node) - (path #3 in Fig. 5) - or
down to a link-layer (the original packet is destined to
another node, so it is forwarded) - (path #5 in Fig. 5).
Implementation Note:
the "IPv6 tunnel link layer" input and output may be
implemented similar to other link layer protocols input and
output, for instance by associating an interface or pseudo-
Conta & Deering Expires in six months [Page 11]
INTERNET-DRAFT Tunneling in IPv6 November 22, 1995
interface with the IPv6 tunnel.
The selection of the "IPv6 tunnel link" over other links
results from the packet forwarding decision taken based on the
content of the node's routing table.
4. Recursive Encapsulation
Recursive IPv6 encapsulation takes place when portions of an IPv6
tunnel are IPv6 tunnels themselves, i.e. a tunnel contains inner
tunnels - see Fig. 6.
The entry-point node of an "inner IPv6 tunnel" may receive and
encapsulate packets that are tunnel IPv6 packets encapsulated at an
"outer IPv6 tunnel" entry-point node. These packets are original
packets for the "inner IPv6 tunnel" entry-point node, the result of
their encapsulation at the "inner IPv6 tunnel entry-point node" is a
"tunnel IPv6 packet" for the "inner IPv6 tunnel", and a "recursive
tunnel IPv6 packet" for the "outer IPv6 tunnel".
Outer Tunnel
<------------------------------------->
<--------------> Inner Tunnel
Outer Tunnel Outer Tunnel
Entry-Point Exit-Point
Node Node
+-+ +-+ +-+ +-+ +-+ +-+
| | | | | | | | | | | |
| |->-//->-| |=>=//=>=| |**>**//**>**| |=>=//=>==| |->-//-->-| |
| | | | | | | | | | | |
+-+ +-+ +-+ +-+ +-+ +-+
Original Inner Tunnel Inner Tunnel Original
Packet Entry-Point Exit-Point Packet
Source Node Node Destination
Node Node
Fig. 6. Recursive Encapsulation
4.1 Limiting Recursive Encapsulation
There is a practical limit on how many "inner IPv6 tunnels" an "outer
IPv6 tunnel" may have which results from the maximum number of
possible IPv6 encapsulations of a packet.
Conta & Deering Expires in six months [Page 12]
INTERNET-DRAFT Tunneling in IPv6 November 22, 1995
Each encapsulation adds to the size of a tunnel packet the size of
the tunnel IPv6 headers. Consequently, in the case of inner tunnels,
the number of recursive encapsulations is practically limited by the
number of tunnel IPv6 headers that can be prepended to the original
packet before the resulting tunnel packet reaches the maximum IPv6
datagram size [IPv6].
The increase in the size of a tunnel IPv6 packet due to repeated
recursive encapsulation may require fragmentation at a tunnel entry
point node [IPv6] - see section 6.7.
Each next recursive encapsulation of a tunnel IPv6 fragment may
result in a new fragmentation and thus the addition of one more
fragment to the previous existing fragments. Therefore, it is highly
probable that once the fragmentation of a tunnel IPv6 packet was
triggered, each next recursive encapsulation may result in additional
fragmentation, and thus IPv6 fragment multiplication. Therefore, it
is recommended to keep recursive encapsulation to a minimum.
The proposed mechanism for controlling excessive recursive
encapsulation is a "tunnel encapsulation limit" that is carried in an
IPv6 Destination Option header.
4.1.1 Tunnel Encapsulation Limit
The "Tunnel Encapsulation Limit" destination option header is built
only by tunnel entry-point nodes, it is discarded only by tunnel
exit-point nodes, and it is used to carry optional information [IPv6]
that need be examined only by tunnel entry-point nodes.
It is defined according to [IPv6] as follows:
Option Type Opt Data Len Tun Encap Lim
0 1 2 3 4 5 6 7
+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
|0 0 0 0 0 1 0 0| 1 | u_8int |
+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
Option Type decimal value 4 - the highest-order two bits - set
to 00 - specify "skip over this option if the
option is not recognized". The third-highest-order
bit - set to 0 - specifies that the option data in
this option does not change en-route to the
packet's destination [IPv6].
Conta & Deering Expires in six months [Page 13]
INTERNET-DRAFT Tunneling in IPv6 November 22, 1995
Opt Data Len 1 - the data portion of the Option is one byte
long.
Tun Encap Lim the Tunnel Encapsulation Limit - 8-bit unsigned
integer (u_8int).
To avoid excessive recursive encapsulation, an IPv6 tunnel entry-
point node may prepend, as part of the IPv6 tunnel headers, a "Tunnel
Encapsulation Limit - Destination Extension Header" - with the "Tun
Encap Lim - tunnel encapsulation limit" - field set to:
(a) a pre-configured value - if the packet has no previous
"Tunnel Encapsulation Limit" header - see section 6.6.
(b) a value resulting from a pre-existent value - if the packet
has a "Tunnel Encapsulation Limit" header, the value is
copied into the newly prepended "Tunnel Encapsulation
Limit" header and then decremented by one.
This is an exception from the rule of processing
destination extension headers, in that although the entry-
point node is not a destination node, during the
encapsulation of a packet, the IPv6 tunneling protocol
engine looks ahead in the extant tunnel headers of that
packet for the "Tunnel Encapsulation Limit" destination
option header.
If the Tunnel Encapsulation Limit is decremented to zero,
the packet undergoing encapsulation is discarded. Upon
dicarding the packet, a Parameter Problem ICMP message
[IPv6ICMP] is returned to the packet originator - the
Parameter Problem ICMP message points into the Tunnel
Encapsulation Limit destination header of the packet, to
the Tun Encap Lim field, which has the value one.
Two cases of recursive encapsulation that should be avoided are
described below:
4.1.2 Loopback Recursive Encapsulation
A particular case of recursive encapsulation which must be avoided is
the loopback recursive encapsulation. Loopback recursive
encapsulation takes place when a tunnel IPv6 entry-point node
encapsulates tunnel IPv6 packets originated from itself, and destined
Conta & Deering Expires in six months [Page 14]
INTERNET-DRAFT Tunneling in IPv6 November 22, 1995
to itself. This may generate an infinite processing loop in the
entry-point node.
To avoid such an infinite processing loop in the tunnel entry-point
node IPv6 protocol engine, the tunneling engine should not
encapsulate packets that have the pair of tunnel entry-point and
exit-point addresses the same as the pair of original packet source
address and final destination address. To avoid the additional
processing in the tunneling protocol engine that the above validation
mechanism would require, it is recommended that the validation be
done at tunnel configuration time: a node should not permit
configuring tunnels where both the entry-point node and exit-point
node addresses belong to the entry-point node.
4.1.3 Routing Loop Recursive Encapsulation
In case of a packet path involving mulitple levels of inner tunnels,
a routing loop from an inner tunnel to an outer tunnel is
particularly dangerous when the loop is such that a tunnel IPv6
packet encapsulated at a certain tunnel entry-point node reaches
again that tunnel entry-point node before reaching that tunnel exit-
point node.
Because there is no relationship between the tunnel IPv6 header hop
limit and the original packet hop limit, a tunnel packet reaching its
originator - a tunnel entry-point - in a routing loop may expire only
after an excessively large number of encapsulations.
Additionally, in such a routing loop case, because the prepending of
the tunnel IPv6 headers adds to the size of the packets, a tunnel
packet may grow to the maximum packet size limit [IPv6], resulting in
tunnel packet fragmentation, and fragment multiplication.
When the path of a packet from source to final destination includes
tunnels, the maximum number of hops that the packet can traverse is
controlled by:
(a) the original IPv6 packet hop limit, at each forwarding
operation performed on the orirginal packet, including at
the points where it is encapsulated and decapsulated.
(b) the tunnel IPv6 packet encapsulation limit, at each
recursive encapsulation of the packet.
The two mechanisms mentioned above are used together to avoid the
negative effects of routing loops in tunnels.
Conta & Deering Expires in six months [Page 15]
INTERNET-DRAFT Tunneling in IPv6 November 22, 1995
5. Tunnel IPv6 Header
The tunnel entry point node fills out a tunnel IPv6 main header
[IPv6] as follows:
Version:
6
Priority:
Depending on the entry-point node tunnel configuration, the
priority may be set to that of the original packet, or to a
pre-configured value - see section 6.3.
Flow label:
Depending on the entry-point node tunnel configuration, the
flow label may be set to a pre-configured value. The tipical
value is zero - see section 6.4.
Payload Length:
The original packet length.
In case the packet is prepended with tunnel IPv6 extension
headers, this value is set to the original packet length
plus the length of the encapsulating IPv6 extension headers.
Next header:
The next header value according to [IPv6] from the Assigned
Numbers RFC [RFC-1700].
If the original packet is an IPv6 packet, and there are no
intermediate headers, the next header protocol value is set
to 41 (Assigned payload type number for IPv6).
If a hop by hop option header is immediately following the
tunnel IPv6 header, then the next header protocol value in
this field is set to 0 (Assigned payload type number for
IPv6 Hop by Hop Options header).
If a Tunnel Encapsulation Limit destination option header is
Conta & Deering Expires in six months [Page 16]
INTERNET-DRAFT Tunneling in IPv6 November 22, 1995
immediately following the tunnel IPv6 header, then the next
header protocol value in this field is set to 60 (Assigned
payload type number for IPv6 Destination Options header).
Hop limit:
The tunnel IPv6 header hop limit is set to a pre-configured
value - see Section 6.3.
The default value for hosts is the neighbor discovery
advertised hop limit [IPv6ND]. The default value for routers
is the default IPv6 Hop Limit [TBD] value from the Assigned
Numbers RFC.
Source Address:
IPv6 address of the outgoing interface of the tunnel entry-
point node. This address is configured as entry-point node
address - see section 6.1.
Destination Address:
IPv6 address of the tunnel exit-point node. If the tunnel
is configured with an unspecified exit-point node address,
then the IPv6 address of the destination from the original
IPv6 header - see section 6.2.
5.1 Tunnel IPv6 Extension Headers
Depending on various IPv6 node configuration parameters a tunnel
entry-point node may append to the tunnel IPv6 main header, one or
more IPv6 extension headers that are not directly related to
tunneling, such as hop by hop, routing,...etc, and therefore they are
not discussed here.
To limit the number of recursive encapsulations of a packet, if it
was configured to do so - see section 6.6 - a tunnel entry-point
appends after the tunnel IPv6 main header, or after the hop by hop
extension header, if any, a Tunnel Encapsulation Limit destination
option header with fields set as follows:
Conta & Deering Expires in six months [Page 17]
INTERNET-DRAFT Tunneling in IPv6 November 22, 1995
+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
| Next Header |Hdr Ext Len = 0| Opt Type = 4 |Opt Data Len=1 |
+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
| Tun Encap Lim |PadN Opt Type=1|Opt Data Len=1 | 0 |
+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
Next Header:
Identifies the type of header immediately following the
Tunnel Encapsulation Limit destination option header [IPv6].
If the original packet is an IPv6 packet, and there are no
intermediate headers, the next header protocol value is set
to 41 (Assigned payload type number for IPv6).
Hdr Ext Len:
Length of the Tunnel Encapsulation Limit destination option
header in 8-octet units, not including the first 8 octets.
Set to value 0.
Option Type:
4 - see 4.1.1.
Opt Data Len:
1 - see 4.1.1.
Tun Encap Lim:
8 bit unsigned integer - see 4.1.1.
Option Type:
1 - PadN option, to align the header following this header.
Opt Data Len:
1 - one octet of option data.
Option Data:
Conta & Deering Expires in six months [Page 18]
INTERNET-DRAFT Tunneling in IPv6 November 22, 1995
One zero-valued octet.
6. IPv6 Tunnel State Variables
The IPv6 tunnel state variables, some of which are or may be
configured, are:
(a) the tunnel entry-point node address - is configured
(b) the tunnel exit-point node address - is configured
(c) the tunnel hop limit - is configured
(d) the tunnel packet priority - is configured
(e) the tunnel packet flow label - is configured
(f) the tunnel encapsulation limit - may be configured
(g) the tunnel MTU
6.1 IPv6 Tunnel Entry-Point Node Address
The tunnel entry-point node address is one of the valid IPv6 unicast
addresses belonging to the entry-point node - it is recommended to
validate the address at configuration time.
The tunnel entry-point node address is copied to the source address
field in the tunnel IPv6 header during packet encapsulation.
6.2 IPv6 Tunnel Exit-Point Node Address
The tunnel exit-point node address is used as the IPv6 destination
address for the tunnel IPv6 header. The tunnel exit-point node
address may be configured with a specific IPv6 address, in which case
the tunnel acts like a virtual point to point link between the
entry-point node and exit-point node, or with the unspecified IPv6
address, in which case the tunnel acts like a virtual multi-point
link, between the entry-point node and many exit-point nodes, in
which case the destination address from each original packet header
Conta & Deering Expires in six months [Page 19]
INTERNET-DRAFT Tunneling in IPv6 November 22, 1995
is used as tunnel IPv6 exit-point node for encapsulating that packet.
The tunnel exit-point node address is copied to the destination
address field in the tunnel IPv6 header during packet encapsulation.
6.3 IPv6 Tunnel Hop Limit
An IPv6 tunnel is modeled as a "single-hop virtual link" tunnel, in
which regardless of the number of hops in the IPv6 tunnel, the
original packet's passing through the tunnel is like the original
packet's passing over a one hop link.
The "single-hop" mechanism should be implemented by having the tunnel
entry point node set a tunnel IPv6 header hop limit independently of
the original headers.
The "single-hop" mechanism hides to the original IPv6 packets the
IPv6 topology of the tunnel.
The tunnel hop limit is configured into the tunnel entry-point node
with a value that:
(a) ensures that tunnel IPv6 packets reach the tunnel exit-
point node
(b) ensures a quick expiration of the tunnel packet if a
routing loop occurs within the IPv6 tunnel.
The tunnel hop limit default value for hosts is the neighbor
discovery advertised hop limit [IPv6ND]. The tunnel hop limit default
value for routers is the default IPv6 Hop Limit [TBD] value from the
Assigned Numbers RFC.
The tunnel hop limit is copied into the hop limit field of the tunnel
IPv6 header of each packet encapsulated by the tunnel entry-point
node.
6.4 IPv6 Tunnel Packet Priority
The IPv6 Tunnel Packet Priority indicates the value that a tunnel
entry-point node sets in the priority field of a tunnel header. The
default value is zero. The Packet Priority may also indicate whether
the value of the priority field in the tunnel header is copied from
the original header, or it is set to the configured value.
Conta & Deering Expires in six months [Page 20]
INTERNET-DRAFT Tunneling in IPv6 November 22, 1995
6.5 IPv6 Tunnel Flow Label
The IPv6 Tunnel Flow Label indicates the value that a tunnel entry-
point node sets in the flow label of a tunnel header. The default
value is zero.
6.6 IPv6 Tunnel Encapsulation Limit
The IPv6 Tunnel Encapsulation Limit may be configured in an entry
point node as the maximum number of encapsulations permitted for
original packets entering the tunnel at that entry-point node.
Recommended default value is 5.
An entry-point node configured to limit the number of encapsulations,
prepends a Tunnel Encapsulation Limit Destination header to an
original packet undergoing encapsulation - see section 5.1.
6.7 IPv6 Tunnel MTU
The tunnel MTU is set dynamically to the Path MTU of the tunnel - the
maximum size of a packet that can be sent through the tunnel without
fragmentation - see [IPv6]. The tunnel entry point node performs Path
MTU discovery on the tunnel [IPv6PMTU], [IPv6ICMP].
The tunnel's entry-point - the encapsulating node - should be able to
send an encapsulated IPv6 packet of any valid size over an IPv6
tunnel.
If the tunnel entry-point node determines that a packet does not
exceed the tunnel MTU after encapsulation, then the tunnel entry-
point node encapsulates and sends that packet.
If the tunnel entry-point node determines that a packet exceeds the
tunnel MTU after encapsulation, then the tunnel entry-point node does
one of the following depending on the size of the original packet:
(a) if the original packet is larger than 576 octets then the
entry-point node returns an ICMP "packet too big" message
to the packet originator. The IPv6 node source of the
original IPv6 packet resizes - makes packets of smaller
size - and retransmits the resulting smaller packets.
Conta & Deering Expires in six months [Page 21]
INTERNET-DRAFT Tunneling in IPv6 November 22, 1995
(b) if the original packet is equal or smaller than 576 octets
then the tunnel entry-point node, after encapsulating the
original packet, breaks the tunnel packet into fragments
that do not exceed the tunnel MTU.
The IPv6 packet encapsulation is considered an IPv6 packet
originating operation, therefore an IPv6 node that is an IPv6 tunnel
entry-point must support fragmentation to encapsulate a packet of any
size.
A tunnel packet being forwarded follows, like any IPv6 packet, the
rule that it must not be fragmented by any IPv6 node other than the
originating node - the tunnel entry-point node.
7. IPv6 Tunnel Error Processing and Reporting
IPv6 Tunneling follows the general rule that errors detected during
the processing of IPv6 packets are reported to the packet originator
through ICMP messages.
For errors detected by nodes that are outside an IPv6 tunnel, on a
path that includes IPv6 tunnels, the errors are reported to the
original IPv6 packet source node.
For errors detected by nodes inside an IPv6 tunnel, ICMP error
messages are sent to the tunnel IPv6 packet source node, which is the
IPv6 tunnel entry-point node. The ICMP messages sent to the tunnel
entry-point node have as ICMP payload the tunnel IPv6 packet that
includes the original packet.
The cause of an error uncovered inside a tunnel can be:
(a) the tunnel header, or
(b) the tunnel packet.
The tunnel header problems are of concern to the tunnel entry-point
node which is the tunnel IPv6 packet originator.
The tunnel packet problems that result from bad encapsulation, are of
concern also to the tunnel entry-point node.
If the tunnel packet problems are a consequence of an original packet
problem and if they can be corrected by the original packet
Conta & Deering Expires in six months [Page 22]
INTERNET-DRAFT Tunneling in IPv6 November 22, 1995
originator, then they must be reported to both the tunnel entry-point
node and the original packet originator.
To report to the original packet originator a problem detected inside
the tunnel and reported from inside the tunnel through an ICMP
message, the tunnel entry-point node must relay the ICMP message to
the original IPv6 packet originator. To relay the ICMP message, the
IPv6 tunnel entry-point node builds and sends an ICMP message to the
original packet originator, based on the tunnel ICMP message, as it
is graphically described below:
+-------+ +-------+ +-----------------------+
| Upper | | Upper | | Upper |
| Layer | | Layer | | Layer |
| Prot. | | Prot. | | IPv6 Tunnel |
| Error | | Error | | Error |
| Input | | Input | | Input |
| | | | | Decapsulate |
| | | | | -->--ICMPv6--#2->-- |
| | | | | | Payload | |
+-------+ +-------+ +--|-----------------|--+
| | | |
^ ^ ^ v
| | | |
--------------------#1-- -----Orig.Packet?--- - - - - - - - - -
error code, | #3 error code, | |
ICMPv6 payload ^ v source address, v v
| IPv6 | orig. packet | IPv4 |
+--------------+ +--------------+ +--------------+ + - - - - +
| | | | | |
| ICMP v6 | | ICMP v6 | | ICMP v4 | | |
| Input | | Error Report | | Error Report |
| - - - - +----+ - - - - | + - - - - + + - - - - +
| | | |
| IPv6 Layer | | IPv4 Layer | | |
| | | |
+----------------------------------+ +--------------+ + - - - - +
Fig. 7. Error Reporting Flow - IPv6 Tunneling Protocol Engine
For example, in case of IPv6 original packets, the tunnel entry-point
node IPv6 layer passes the received ICMP message to the ICMPv6 Input.
The ICMPv6 Input, based on the ICMP type and code [IPv6ICMP]
generates an internal "error code" which is passed with the "ICMPv6
message payload" to the upper layer protocol - in this case the IPv6
tunnel upper layer - error input (path #1 in Fig. 7, and (a) in Fig.
8).
Conta & Deering Expires in six months [Page 23]
INTERNET-DRAFT Tunneling in IPv6 November 22, 1995
The IPv6 tunnel error input decapsulates the tunnel IPv6 packet,
which is the ICMPv6 message payload, obtaining the original packet,
and thus the original headers - path #2 in Fig. 7 and (b) in Fig. 8 -
and dispatches the "error code", the source address from the original
packet header, and the original packet, down to the error report
block of the protocol identified by the Next Header field in the
tunnel header immediately preceding the original packet in the ICMP
message payload - in this example ICMPv6. The ICMPv6 error report
builds an ICMP message of a type and code according to the "error
code", containing the "original packet" as ICMP payload - - path #3
in Fig. 7 and (c) in Fig. 8. The ICMP message has the tunnel entry-
point node address as source address, and the original packet source
node address as destination address - (d) in Fig. 8. The tunnel
entry point node sends the ICMP message to the original packet
originator node.
A graphical description of the header processing taking place is the
following:
< Tunnel packet >
+--------+- - - - - -+--------+------------------------------//------+
| IPv6 | IPv6 | ICMP | Tunnel |
(a)| | Extension | | IPv6 |
| Header | Headers | Header | Packet in error |
+--------+- - - - - -+--------+------------------------------//------+
< Tunnel headers > < Tunnel ICMP message >
< ICMPv6 message payload >
|
v
< Tunnel ICMP message >
< Tunnel IPv6 packet in error >
+--------+ +---------+ +----------+--------//------+
| ICMP | | Tunnel | | Original | Original |
(b) | | <- | IPv6 | <- | IPv6 | IPv6 packet |
| Header | | Headers | | Headers | payload |
+--------+ +---------+ +----------+--------//------+
| | <Orig. IPv6 pck. in error >
----------------- | |
----------|---------------
| |
V V
+---------+ +--------+ +-------------------//------+
| New | | ICMP | | |
(c) | IPv6 | -> | | -> | Orig. IPv6 packet in error|
| Headers | | Header | | |
+---------+ +--------+ +-------------------//------+
|
Conta & Deering Expires in six months [Page 24]
INTERNET-DRAFT Tunneling in IPv6 November 22, 1995
v
+---------+--------+-------------------//------+
| New | ICMP | Original |
(d) | IPv6 | | IPv6 |
| Headers | Header | packet in error |
+---------+--------+-------------------//------+
< new ICMP message >
Fig. 8. ICMP Error reporting and processing
7.1 Tunnel ICMP Messages
The tunnel ICMP messages which are reported to the original packet
originator are:
hop limit exceeded
The tunnel is misconfigured, or contains a routing loop,
and packets do not reach the tunnel exit-point node. This
problem is of concern to the tunnel entry point node - the
tunnel hop limit must be reconfigured - and also to the
original IPv6 packet originator.
unreachable node
One of the nodes in the tunnel is not or is no longer
reachable. This is a problem of concern to the tunnel
entry-point node, which should reconfigure the tunnel with
a valid and active path between the entry and exit-point of
the tunnel.
parameter problem
A Parameter Problem ICMP message pointing to a valid Tunnel
Encapsulation Limit Destination header with a Tun Encap Lim
field value set to one is an indication that the tunnel
packet exceeded the maximum number of encapsulations
allowed.
The three above problems detected inside the tunnel, which are a
tunnel configuration and a tunnel topology problem, are reported to
the original IPv6 packet originator, as a tunnel generic
"unreachable" problem caused by a "link problem" - see section 7.2
Conta & Deering Expires in six months [Page 25]
INTERNET-DRAFT Tunneling in IPv6 November 22, 1995
and 7.3.
packet too big
The tunnel packet exceeds the tunnel Path MTU. The original
packet originator is notified - see section 6.3. If the
original packet is an IPv6 packet, the ICMP message sent to
the packet originator has the same ICMP type/code as the
ICMP message sent from inside the tunnel to the tunnel
entry-point node - see section 7.2, and 7.3.
This ICMP message is used by the tunnel entry point node to
determine the tunnel MTU.
7.2 ICMP Messages for IPv6 Original Packets
The tunnel entry-point node builds the ICMP and IPv6 headers of the
new ICMP message as follows:
IPv6 Fields:
Source Address
A valid unicast IPv6 address of the outgoing interface.
Destination Address
Copied from the Source Address field of the Original
IPv6 header.
ICMP Fields:
For tunnel ICMP error message "hop limit exceeded", or "unreachable
node", or "parameter problem" pointing to a valid Tunnel
Enacpsulation Limit destination header with the Tun Encap Lim field
set to a value one:
Conta & Deering Expires in six months [Page 26]
INTERNET-DRAFT Tunneling in IPv6 November 22, 1995
Type 1 - unreachable node
Code 3 - address unreachable
For tunnel ICMP error message "packet too big":
Type 2 - packet too big
Code 0
MTU The MTU field from the tunnel ICMP message minus
the length of the tunnel headers.
7.3 ICMP Messages for IPv4 Original Packets
The tunnel entry-point node builds the ICMP and IPv4 header of the
new ICMP message as follows:
IPv4 Fields:
Source Address
A valid unicast IPv4 address of the outgoing interface.
Destination Address
Copied from the Source Address field of the Original
IPv4 header.
ICMP Fields:
For tunnel ICMP error message "hop limit exceeded", or "unreachable
node", or "parameter problem" pointing to a valid Tunnel
Enacpsulation Limit destination header with the Tun Encap Lim field
set to a value one:
Conta & Deering Expires in six months [Page 27]
INTERNET-DRAFT Tunneling in IPv6 November 22, 1995
Type 3 - destination unreachable
Code 1 - host unreachable
For tunnel ICMP error message "packet too big":
Type 3 - destination unreachable
Code 4 - datagram too big
MTU The MTU field from the tunnel ICMP message minus
the length of the tunnel headers.
8. Open Issues
Open issues are:
(a) Multicast exit point
Does it make sense to have an IPv6 multicast address as tunnel
exit-point node address?
(b) Anycast exit point
Does it make sense to have an IPv6 anycast address as tunnel
exit-point node address?
(c) Tunnel default hop limit value
At this time, there is no definition for an IPv6 hop limit
default value. The Assigned Numbers [RFC-1700] IPv4 TTL default
value could be used instead.
9. References
[IPv6]S. Deering, R. Hinden, "Internet Protocol Version 6
Specification"
[IPv6ICMP]
Conta & Deering Expires in six months [Page 28]
INTERNET-DRAFT Tunneling in IPv6 November 22, 1995
A. Conta, and S. Deering "Internet Control Message Protocol for
the Internet Protocol Version 6 (IPv6)"
[IPv6ND]
T. Narten, E. Nordmark, "IPv6 Neighbor Discovery Specification"
[IPv6PMTU]
J. McCann and S. Deering, "IPv6 Path MTU Discovery"
[RFC-1700]
J. Reynolds, J. Postel, "Assigned Numbers", 10/20/1994
10.Acknowledgements
The document is partially derived from several ideas about tunneling,
from several discussions about IPv6 in IPv6 tunneling on the IPng
Working Group Mailing List, and from feedback from an IPv6 tunneling
focused IPng Working Group session at the 33th IETF, in Stockholm,
July 1995.
Additionally, several documents focused on tunneling or encapsulation
were used as a reference: "Transition Mechanisms for IPv6 Hosts and
Routers" document by Robert Gilligan and Erik Nordmark, RFC 1241 (R.
Woodburn, and D. Mills), RFC 1326 (P. Tsuchiya), RFC 1701, and RFC
1702 (S. Hanks, D. Farinacci, P. Traina), RFC 1834 (W. Simpson), and
Mobile IP working Group drafts (C. Perkins).
Also an important contribution was made by the reviewers of this
document: [TBD]
11.Security Considerations
Security considerations are not discussed in this memo.
Conta & Deering Expires in six months [Page 29]
INTERNET-DRAFT Tunneling in IPv6 November 22, 1995
Authors' Addresses:
Alex Conta Stephen Deering
Digital Equipment Corporation Xerox Palo Alto Research Center
110 Spitbrook Rd 3333 Coyote Hill Road
Nashua, NH 03062 Palo Alto, CA 94304
+1-603-881-0744 +1-415-812-4839
email: conta@zk3.dec.com email: deering@parc.xerox.com