OSI network addressing scheme. Use of the Local format for these
non-standard address ensures that they cannot be confused with standard
OSI network addresses. This capability will be useful in the evolution
of existing networks to OSI, and for the accommodation of non-OSI
addressing schemes that may be used in proprietary network
architectures or for testing and other interim purposes. It should be
emphasized that
the Local format is not intended to give non-OSI schemes a permanent
place in OSI, but rather to permit the OSI network addressing sheme to
be used wherever possible without risk of conflict with other schemes
(which can be encapsulated safely under the Local format).
C.2 RESERVATION OF AFI VALUES 00-09 (Table 8-2)
The reservation of AFI values beginning with the digit 0 is intended to
allow for the use of an initial 0 to handle special cases, such as:
a) as an escape to some other addressing scheme;
b) as a technique for the optimization of NSAP address encoding in
Network Layer protocols, when the different parts of the NSAP
address semantics are encoded in different fields of the protocol
header;
c) as a way to indicate, in a protocol header, that a field that
ordinarily contains a full NSAP address in fact contains something
less than a full address (for example, a shorthand form that omits
specification of the higher-order domains, which might be used for
communication within a particular subdomain environment).
There may be other cases in which the use of an initial 0 digit is
found to be useful. This Addendum merely reserves the AFI values 00-09,
and does not specify how they might be used; all such uses are outside
the scope of this Addendum.
C.3 DERIVATION OF THE CONCRETE SYNTAXES (Clause 8.3)
In describing the two "preferred" concrete syntaxes of the NSAP
address, Clauses 8.3.1 and 8.3.2 introduce two types of padding:
padding with zero digits at the beginning of an IDI, and padding with a
semi-octet with the value 1111 at the end of the binary encoding of an
IDI with an odd number of decimal digits.
The first type of padding is necessary because some of the IDI formats
allow the IDI to consist of a variable number of digits. Since there is
no explicit syntactic marker between the IDI and the DSP, the only way
to find the end of the IDI is to know how long it is. The AFI, which
identifies which IDI format is used, allows only the maximum length of
that IDI to be determined. Rather than introduce either a specific
syntactic marker or a new field containing the length of the IDI
(either of which would have greatly complicated the encoding and
parsing of NSAP addresses), the Addendum specifies that for encoding
purposes the IDI must first be padded out to its maximum length. Note
that this does not apply to the DSP; only to the IDI.
The second type of padding is necessary to ensure that a binary
encoding of the IDI consists of an integral number of binary octets.