Structure of NSAP Addresses

The Syntax of an NSAP Address

An NSAP, as defined by ISO, has several characteristics. It may be one to twenty octets (bytes) in length. It is composed of two parts: the Initial Domain Part (IDP) and the Domain Specific Part (DSP). The DSP may also be partitioned into several fields. These fields and their sizes are defined by the authority that controls the IDP.

The Initial Domain Part

The IDP portion of an NSAP declares which national or international group owns the right to manage an NSAP space. This group is called an authority. Examples of authorities that presently manage NSAP spaces are AFNOR, ANSI and NIST. They have been given, by ISO, specific IDP values that they control and have the right to use for the distribution of NSAP addresses. As an example, NIST has been given the IDP value 470005 for use within the U.S. GOSIP networks. Therefore, any NSAP that begins with this value is ultimately under the control of NIST. NIST also owns the right to define how the DSP portion of its NSAP space is to be formatted and used.

The IDP contains two fields. The authority and format identifier (AFI) field identifies the type of address used in the DSP. The initial domain identifier (IDI) field identifies which domain the DSP part belongs to.

	NOTE: In this section, all numeric values are represented as hexadecimal digits unless otherwise specified.

The Domain Specific Part

The purpose of the DSP is to

This is accomplished by dividing the DSP into a number of fields, each with its own meaning. As an example, consider the format defined by ANSI:

The AFI value of 39 indicates that the IDI portion will be a Data Country Code (DCC), and the DSP will be encoded in binary. The value 840 (padded with F) is the DCC for the United States.

ANSI has defined the first three octets of the DSP to contain an Organization Identifier (Org Id). This number, assigned by ANSI, allows other organizations to control a subset of ANSI's NSAP space. ANSI has left the rest of the DSP undefined so that the organization will have the ability to define its own address structure. For instance, if Joe's Grommet Shop petitions ANSI for an Org Id value and is assigned the value 010101 (three octets), then Joe's Grommet Shop controls the prefix 39840F010101. Any NSAP that begins with this value is under the control of Joe's Grommet Shop. Joe's Grommet Shop may also define the format for the rest of the DSP to suit their needs.

The fields of an NSAP create a hierarchy where each field further divides the NSAP space into smaller, more manageable spaces. Also, the leftmost portion of an NSAP deals with the administration of NSAP spaces, not specifically with the identification of real systems. In the above example, the leftmost part of the NSAP space is ``publicly'' administered, that is, administered by national and international standards bodies, whereas the right part is left up to private organizations to define and use. For convenience, these two parts will be called the administrative prefix (AP), and the privately allocated part (PAP).

This separation between what a network administrator is given and what the administrator controls is very important. It is the key to understanding how an NSAP space is managed. Also, some companies may have several different APs, or may start with one AP and have to migrate over to a different one at some point in time. Only through careful management of the PAP can a network administrator minimize the problems that this may cause.

The Privately Allocated Part

Whereas the AP is used to administer NSAP spaces, the PAP is used to define a routing strategy. The two portions, when combined together, yield a unique NSAP for a given system.

The ISO model of routing contains the following levels:

Subnetwork - A subnetwork is an autonomous collection of equipment and medias used to interconnect systems. Examples of types of subnetworks are: IEEE 802.3 LANs and X.25 PDNs.

Area - An area is a group of end systems and intermediate systems interconnected by one or more subnetworks. They have been grouped together by way of an autonomous routing mechanism. An autonomous routing mechanism is either a set of statically configured intermediate systems, or a set of intermediate systems that support a dynamic routing protocol, such as the ISO IS-IS protocol.

Routing Domain - A Routing Domain is also a group of end systems and intermediate systems interconnected by one or more subnetworks. They have been grouped together by way of some routing or security policy defined by the network administrator. Typically, a number of areas are grouped together to form a routing domain.

An area may be composed of one or more subnetworks. The criteria for grouping subnetworks into areas can be based upon topology, reachability, and network traffic. The criteria for grouping areas into routing domains is based upon policy. For instance, a company might want to create an area for their factory, and area for their product design lab and put them in the same routing domain. However, the company may choose to create a separate routing domain for their accounting and payroll departments.

To facilitate routing strategy, a network administrator will often define the PAP to contain one or more of the following fields:

The order they are listed here is the order, left to right, in which they would be defined in a PAP.

Other fields that often find their way into the PAP are:

Version Number or DSP Format Identifier - The purpose of these fields is to allow a network administrator to redefine the PAP structure at a later point in time, or allow the network administrator to define several PAP formats. Each format would be identified by a different value in the Version Number or DSP Format Identifier field.

Reserved field - This is simply a portion of the PAP that is not in use, but has been reserved in case it is needed some time in the future. If the network administrator defines a PAP with a reserved field, a default value (usually all 00s) should be defined that the reserved field is always set to.

NSAP selector - A one octet field at the rightmost portion of the PAP. It identifies the entity attached to the network layer. In OTS the entity is always the transport protocol. HP recommends that the NSAP selector be set to 01.

The section on current allocation formats, and the section on recommended PAP structures show examples of how these fields are used to manage NSAP spaces and routing topologies.

ISO IS-IS Routing protocol - ISO is in the process of defining a protocol standard for a dynamic Intermediate System to Intermediate System routing protocol, known as the IS-IS protocol. This protocol, when complete, will allow Intermediate Systems (the ISO term for network layer routers) to automatically route traffic between each other across concatenated subnetworks and areas. It will not, however, automate routing traffic between routing domains. This will still require static configuration of route information.

Though the protocol is still in development, ISO has defined a format for the rightmost portion of the DSP to be used by the IS-IS protocol for its automated routing. Network administrators should use this format whenever possible so that it is easier to support the IS-IS protocol once it becomes available.

The format defined by ISO is the following:

This format is commonly called the 2/6/1 structure because of the sizes of the three defined fields.

This is the format that NSAP addresses should obey in order to get the most use out of the automated routing protocols. See the section "A Recommended PAP Structure" later in this chapter for a possible PAP structure that network administrators may use to define NSAP formats. This recommended structure is compatible with the IS-IS protocol format.