1. Introduction

   When experimenting with or extending protocols, it is often necessary
   to have a protocol number as part of the implementation [RFC2434].
   For example, to develop a protocol that runs directly above IP, one
   needs an IP Protocol Number to place in the Protocol field of the IP
   header [RFC791].  In some cases, obtaining a new number is
   straightforward (e.g., a well-known TCP or UDP port) or not even
   necessary (e.g., TCP and UDP port numbers for testing purposes).  In
   other cases, obtaining a number is more difficult.  For example, the
   number of available and unassigned values in a name space may be
   small enough that there is concern that all available numbers will be
   used up if assigned carelessly.  Even in cases where numbers are
   potentially plentiful, it may be undesirable to assign numbers unless
   the proposed usage has been adequately reviewed by the broader
   community.  Consequently, some number spaces specify that IANA only
   make assignments in cases where there is strong community support for
   a proposed protocol.  For example, values out of some name spaces are
   only assigned through an "IETF Standards Action" [RFC2434], which
   requires that the proposed use be in an IETF Standards Track RFC.

   In order to experiment with a new protocol, an experimental value may
   be needed that won't collide with an existing or future usage.

   One approach is to allow IANA to make temporary assignments for such
   purposes.  The idea is that a protocol value can be assigned to allow
   experimentation, but after the experiment ends, the number would be
   returned to IANA.  There are several drawbacks to this approach,
   however.  First, experience has shown that it can be difficult to
   reclaim numbers once assigned.  For example, contact information
   becomes outdated and it can become difficult to find out what the
   status of an experiment actually is.  Second, should deployment with
   the temporarily assigned number take place (e.g., it is included as
   part of a product), it becomes very difficult to determine whether or
   not reuse of that number would lead to adverse impact with regards to
   deployed devices.  Finally, it can be difficult to determine when an
   experiment has ended and whether the number needs to be returned.

   An alternate approach, and the one recommended in this document, is
   to assign a range of numbers specifically earmarked for testing and
   experimentation purposes.  Mutually consenting devices could use
   these numbers for whatever purposes they desire, but under the
   understanding that they are reserved for generic testing purposes,
   and other implementations may use the same numbers for different
   experimental uses.

   Numbers in the experimentation range are similar to those called
   "Private Use" in RFC 2434 [IANA-CONSIDERATIONS].  They are not
   intended to be used in general deployments or be enabled by default
   in products or other general releases.  In those cases where a
   product or release makes use of an experimental number, the end user
   must be required to explicitly enable the experimental feature and
   likewise have the ability to chose and assign which number from the
   experimental range will be used for a specific purpose (i.e., so the
   end user can ensure that use of a particular number doesn't conflict
   with other on-going uses).  Shipping a product with a specific value
   pre-enabled would be inappropriate and can lead to interoperability
   problems when the chosen value collides with a different usage, as it
   someday surely will.

   From the above, it follows that it would be inappropriate for a group
   of vendors, a consortia, or another Standards Development
   Organization to agree among themselves to use a particular value for
   a specific purpose and then agree to deploy devices using those
   values.  By definition, experimental numbers are not guaranteed to be
   unique in any environment other than one where the local system
   administrator has chosen to use a particular number for a particular
   purpose and can ensure that a particular value is not already in use
   for some other purpose.

   Once an extension has been tested and shown to be useful, a permanent
   number could be obtained through the normal assignment procedures.

   Most implementations will not do anything special with numbers
   assigned for testing purposes.  In particular, unless a packet or
   other Protocol Data Unit (PDU) is specifically directed at a device,
   that device will not even look at the field while processing the PDU.
   For example, IP routers do not need to examine or understand the
   Protocol Type field of IP datagrams in order to know how to correctly
   forward them.  In those cases where a packet or PDU is directed at a
   device, and that device has not been configured to recognize the
   extension, the device will either ignore the PDU, discard it, or
   signal an error, depending on the protocol-specific rules that
   indicate how to process unknown options or features.  In those cases
   where a protocol has different ways of handling unrecognized
   extensions (e.g., silently discard vs. signal an error), that
   protocol needs to reserve values for testing purposes from all the
   appropriate ranges.  Only those implementations specifically enabled
   or configured to make use of an extension or feature that is being
   experimented with would process the data further.

1.1. Recommendation for Protocols

   To make it possible to experiment with protocol extensions safely,
   protocol documents should consider reserving a small set of protocol
   numbers for experimentation.  Such reservations can be made through
   an explicit reservation in an IANA Considerations section.

   The exact number of values to reserve for experimentation will depend
   on the specific protocol and factors specific to that protocol.  For
   example, in cases where the values of a field are subdivided into
   ranges that are treated differently (e.g., "silently ignore" vs.
   "return an error" if the value is not understood), one or more values
   from each sub-range may need to be reserved.

   For protocols that return error codes, it may also be appropriate to
   reserve a small number of experimental error values that can be used
   in conjunction with possible experimental uses.  For example, an
   experimental message might result (even under normal conditions) in
   an error, with a special error code (or sub-code) indicating the type
   of error condition.

   In many, if not most cases, reserving a single value for experimental
   use will suffice.  While it may be tempting to reserve more in order
   to make it easy to test many things at once, reserving many may also
   increase the temptation for someone using a particular value to
   assume that a specific experimental value can be used for a given
   purpose exclusively.  Values reserved for experimental use are never
   to be made permanent; permanent assignments should be obtained
   through standard processes.  As described above, experimental numbers
   are intended for experimentation and testing and are not intended for
   wide or general deployments.

   When protocols that use experimental numbers are included in
   products, the shipping versions of the products must disable
   recognition of protocol experimental numbers by default -- that is,
   the end user of the product must explicitly "turn on" the
   experimental protocol functionality.  In most cases, a product
   implementation must require the end user to configure the value
   explicitly prior to enabling its usage.  Should a product not have a
   user interface for such end user configuration, the product must
   require explicit re-programming (e.g., a special firmware download,
   or installation of a feature card) to configure the experimental
   number(s) of the protocol(s) implicitly.