From: Howard C. Berkowitz (hcb@gettcomm.com)
Date: Thu Oct 02 2003 - 16:09:28 GMT-3
At 5:29 PM +0100 10/2/03, Ken.Farrington@barclayscapital.com wrote:
>Guys,
>
>v.quick one indeed.
>
>Lots of examples of NETs on the www. Can I just clarify a couple of points
>(MAINLY THE LENGTH OF THE AREA ID is it 1 byte of 3?) :-
>
>I have confiured this
>
>router isis ken
> net 47.1000.5000.0001.0000.00
>
>the first xx.xxxx of the net is an area ID - this is a static size and
>always 3 bytes?
>the next fields are the system ID (shortest being xxxx.xxxx (4 bytes),
>longest being xxxx.xxxx.xxxx.xxxx.xxxx.xxxx.xxxx.xxxx (16 bytes)
>the last two are NSAP xx (1 byte)
>
>Is this correct??????????????????????????????????
There really are two questions here. First, what will Cisco accept as
an area ID, and if it has to meet the true rules for NSAP formats.
Second, what is the format of an NSAP.
An NSAP starting 0x4706, for example, means that the NSAP is encoded
in binary (as opposed to decimal) and the IDP uses codes for
internationally recognized networking organizations, assigned by ISO.
0x06 happens to be the US Department of Defense.
To the first part, at least in the past (certainly for ATM), Cisco
did not actually look at an NSAP format other than in CLNP routing. I
haven't personally tested it, but I would strongly suspect that IOS
really doesn't care about the values of an ISIS area ID as long as
the areas are unique where they need to be unique.
Second, NSAPs are variable-length strings of which the first bytes
define the format and the address authority that makes the address
unique. The NSAP splits into two logical parts, the Initial Domain
Part and the Domain Specific Part. IDP information includes format,
who controls the address space (e.g., some forms have an embedded
country code), etc. The real "user address part" is the DSP.
Certain formats, as you'll see below, have a null IDP or a null DSP,
with the remaining part being enough for uniqueness (e.g., there's a
telephone number in the IDP).
The material below is from my book, Designing Addressing
Architectures for Routing and Switching. I'll need to translate some
diagrams into text. Unfortunately, there are some graphics that
would help but are a little complex to translate to ASCII. I will
have them up in a public area on certificationzone.com in the next
few days.
>Abstractly, AFI's have two decimal digits. They will variously be
>allocated by ISO, CCITT, or both in joint agreement; reserved and
>not to be used; and reserved for a family of addressing conventions
>already in use. Those existing conventions include
>o X.121 (used by X.25, X.21, and X.75 networks),
>o F.69 used for Telex,
>o E.163 used for public switched telephone numbers,
>o E.164 for ISDN and public ATM
>o International Codes defined by ISO 6523
>o Country Codes defined by ISO 3166
>Addressing authorities identified by the AFI prescribe the syntax of
>the DSP and administer the assignment of semantics. Binary octets,
>decimal digits, character, and national character syntaxes may be
>used. IDIs abstractly have a variable number of decimal digits.
>Encoding, or syntax, may be significantly different in the manner a
>user or administrator externally specifies it, and the way it is
>encoded into NPDUs.
>Numerous ways exist to encode NSAPs. One popular method, for which
>a U.S. government example is shown, allows self-configuration of
>devices on a LAN. This is also used as an ATM endpoint identifier
>in private ATM networks.
>
>(d) ATM Use of NSAP
>ATM endpoint addressing uses four variants of the NSAP address.
>o ISO Data Country Code
>o ISO International Code Designator
>o ITU-T (former CCITT) E.164 private
>o ITU-T E.164 public format
>Private networks are required to support the first three types,
>while public networks either must support the fourth, or the first
>three types. The first three are "data-oriented," in that they
>contain a six-byte end station identifier that often, in practice,
>is a MAC address. The ICD format was described above.
>(e) Data Country Code (DCC)
>Data country codes are associated with national addressing
>authorities. The IDP has an AFI is 0x39 and a two-byte IDI with a
>value from the ISO 3166 country code space. The DSP is identical to
>that used with the ICD format.
>(e) E.164 Private
>The "full NSAP" version of this format begins eginning with an AFI
>of 0x45, the IDI is an 8-byte, binary-encoded field. It is followed
>by a DSP composed of a RD and AREA field, followed in turn by the
>six-byte end system ID and a selector field.
>(e) E.164 Public
>As shown in E.164 only, there is no AFI field; the address begins
>with a four-byte country code administered by ITU-T. A
>variable-length binary-encoded-decimal DSP follows, which may be up
>to 15 decimal digits long. The DSP is a "national number," much
>like an X.121 address or E.163 telephone number, whose format is
>specified by the national authority.
***Get your CCIE and a FREE vacation: Shop.GroupStudy.com***
This archive was generated by hypermail 2.1.4 : Mon Nov 24 2003 - 07:52:55 GMT-3