RFC 796 HISTORIC ADDRESSING · IP · HISTORIC
Address mappings
A 1981 chart showing how a 32-bit internet address was split into network and local parts, and how each early network's own addresses fitted inside.
RFC EDITOR
Postel, J., "Address mappings", RFC 796, DOI 10.17487/RFC0796, September 1981, <https://www.rfc-editor.org/info/rfc796>.
BIBTEX
@misc{rfc796,
series = {Request for Comments},
number = 796,
howpublished = {RFC 796},
publisher = {RFC Editor},
doi = {10.17487/RFC0796},
url = {https://www.rfc-editor.org/info/rfc796},
author = {J. Postel},
title = {{Address mappings}},
pagetotal = 7,
year = 1981,
month = sep,
} RFC text
Contents
Network Working Group J. Postel
Request for Comments: 796 ISI
Replaces: IEN 115 September 1981
ADDRESS MAPPINGS
----------------
Internet Addresses
------------------
This memo describes the relationship between address fields used in
the Internet Protocol (IP) [1] and several specific networks.
An internet address is a 32 bit quantity, with several codings as
shown below.
The first type (or class a) of address has a 7-bit network number and
a 24-bit local address.
1 2 3
0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1
+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
|0| NETWORK | Local Address |
+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
Class A Address
The second type (or class b) of address has a 14-bit network number
and a 16-bit local address.
1 2 3
0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1
+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
|1 0| NETWORK | Local Address |
+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
Class B Address
The third type (or class c) of address has a 21-bit network number
and a 8-bit local address.
1 2 3
0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1
+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
|1 1 0| NETWORK | Local Address |
+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
Class C Address
The local address carries information to address a host in the
network identified by the network number. Since each network has a
RFC 796 Address Mappings
particular address format and length, the following section describes
the mapping between internet local addresses and the actual address
format used in the particular network.
Internet to Local Net Address Mappings
--------------------------------------
The following transformations are used to convert internet addresses
to local net addresses and vice versa:
AUTODIN II
----------
The AUTODIN II has 16 bit subscriber addresses which identify
either a host or a terminal. These addresses may be assigned
independent of location. The 16 bit AUTODIN II address is
located in the 24 bit internet local address as shown below.
The network number of the AUTODIN II is 26 (Class A).
+----------------+
| HOST/TERMINAL | AUTODIN II
+----------------+
16
+--------+--------+--------+--------+
| 26 | ZERO | HOST/TERMINAL | IP
+--------+--------+--------+--------+
8 8 16
RFC 796 Address Mappings
ARPANET
-------
The ARPANET (with 96 bit leaders) has 24 bit addresses. The 24
bits are assigned to host, logical host, and IMP leader fields
as illustrated below. These 24 bit addresses are used directly
for the 24 bit local address of the internet address. However,
the ARPANET IMPs do not yet support this form of logical
addressing so the logical host field is set to zero in the
leader.
The network number of the ARPANET is 10 (Class A).
+--------+--------+--------+
| HOST | ZERO | IMP | ARPANET
+--------+--------+--------+
8 8 8
+--------+--------+--------+--------+
| 10 | HOST | LH | IMP | IP
+--------+--------+--------+--------+
8 8 8 8
DCNs
----
The Distributed Computing Networks (DCNs) at COMSAT and UCL use
16 bit addresses divided into an 8 bit host identifier (HID),
and an 8 bit process identifier (PID). The format locates
these 16 bits in the low order 16 bits of the 24 bit internet
address, as shown below.
The network number of the COMSAT-DCN is 29 (Class A), and of
the UCL-DCN is 30 (Class A).
+--------+--------+
| HID | PID | DCN
+--------+--------+
8 8
+--------+--------+--------+--------+
| 18 | ZERO | HID | PID | IP
+--------+--------+--------+--------+
8 8 8 8
RFC 796 Address Mappings
EDN
---
The Experimental Data Network at the Defense Communication
Engineering Center (DCEC) uses the same type of addresses as
the ARPANET (with 96 bit leaders) and has 24 bit addresses.
The 24 bits are assigned to host, logical host, and IMP leader
fields as illustrated below. These 24 bit addresses are used
directly for the 24 bit local address of the internet address.
However, the IMPs do not yet support this form of logical
addressing so the logical host field is set to zero in the
leader.
The network number of the EDN is 21 (Class A).
+--------+--------+--------+
| HOST | ZERO | IMP | EDN
+--------+--------+--------+
8 8 8
+--------+--------+--------+--------+
| 21 | HOST | LH | IMP | IP
+--------+--------+--------+--------+
8 8 8 8
LCSNET
------
The LCS NET at MIT's Laboratory for Computer Science uses 32
bit addresses of several formats. Please see [3] for more
details. The most common format locates the low order 24 bits
of the 32 bit LCS NET address in the 24 bit internet local
address, as shown below.
The network number of the LCS NET is 18 (Class A).
+--------+--------+--------+
| SUBNET |RESERVED| HOST | LCSNET
+--------+--------+--------+
8 8 8
+--------+--------+--------+--------+
| 18 | SUBNET |RESERVED| HOST | IP
+--------+--------+--------+--------+
8 8 8 8
RFC 796 Address Mappings
PRNET
-----
The Packet Radio networks use 16 bit addresses. These are
independent of location (indeed the hosts may be mobile). The
16 bit PRNET addresses are located in the 24 bit internet local
address as shown below.
The network numbers of the PRNETs are:
BBN-PR 1 (Class A)
SF-PR-1 2 (Class A)
SILL-PR 5 (Class A)
SF-PR-2 6 (Class A)
BRAGG-PR 9 (Class A)
DC-PR 20 (Class A)
+--------+--------+
| HOST | PRNET
+--------+--------+
16
+--------+--------+--------+--------+
| net | ZERO | HOST | IP
+--------+--------+--------+--------+
8 8 16
RFC 796 Address Mappings
SATNET
------
The Atlantic Satellite Packet Network has 16 bit addresses for
hosts. These addresses may be assigned independent of location
(i.e., ground station). It is also possible to assign several
addresses to one physical host, so the addresses are logical
addresses. The 16 bit SATNET address is located in the 24 bit
internet local address as shown below.
The network number of the SATNET is 4 (Class A).
+--------+--------+
| HOST | SATNET
+--------+--------+
16
+--------+--------+--------+--------+
| 4 | ZERO | HOST | IP
+--------+--------+--------+--------+
8 8 16
WBCNET
------
The Wideband Communication Satellite Packet Network (WBCNET)
Host Access Protocol (HAP) has 16 bit addresses for hosts. It
is possible to assign several addresses to one physical host,
so the addresses are logical addresses. The 16 bit WBCNET
address is divided into a HAP Number field and a Local Address
field, and is located in the 24 bit internet local address as
shown below. Please see [2] for more details.
The network number of the WBCNET is 28 (Class A).
+--------+--------+
| HAP NUM| LCL ADD| WBCNET
+--------+--------+
8 8
+--------+--------+--------+--------+
| 28 | HAP NUM| ZERO | LCL ADD| IP
+--------+--------+--------+--------+
8 8 8 8
RFC 796 Address Mappings
References
----------
[1] Postel, J. (ed.), "Internet Protocol - DARPA Internet Program
Protocol Specification," RFC 791, USC/Information Sciences
Institute, September 1981.
[2] Pershing J., "Addressing Revisited," Bolt Beranek and Newman
Inc., W Note 27, May 1981.
[3] Noel Chiappa, David Clark, David Reed, "LCS Net Address
Format," M.I.T. Laboratory for Computer Science Network
Implementation, Note No.5, IEN 82, February 1979.
AI ELI5
On this page
PROBLEMAn internet address has a network number and a local address, but every network has its own address format and length. Gateways needed to know how to turn one into the other.
MECHANISMThe memo shows the three address codings (class a, b and c) and then, for each of several networks, draws exactly where that network's native address sits inside the local part of the internet address.
STATUS TODAYThe unit metadata lists this RFC as Historic. It is a snapshot of the networks and network numbers of September 1981 and replaces IEN 115.
§1Internet Address Codings
This memo describes the relationship between the address fields used in the Internet Protocol (RFC 791) and several specific networks. An internet address is a 32 bit quantity, and it has several codings.
Each coding splits the 32 bits into a network number and a local address. The local address carries the information needed to address a host inside the network that the network number names.
An internet address is like a street address with two parts: which neighborhood, and which house inside it.
§2Class A, B and C Addresses
The first type, class a, has a 7-bit network number and a 24-bit local address. The address starts with a leading 0 bit.
The second type, class b, has a 14-bit network number and a 16-bit local address, and starts with the bits 1 0. The third type, class c, has a 21-bit network number and an 8-bit local address, and starts with the bits 1 1 0.
There are three sizes of neighborhood. Some have a huge number of houses, some a medium number and some only a few.
081624
| Field | Name | Size | Value | Description |
|---|---|---|---|---|
CLASS-BIT | Class bit | 1 bits | 0 | A leading zero marks a class a address. |
NETWORK | Network | 7 bits | The 7-bit network number. | |
LOCAL | Local Address | 3 bytes | Addresses a host within the network. |
§3Mapping Local Network Addresses
Because each network has its own address format and length, the memo gives a transformation for each one. In every case the network's own address is placed inside the local part of the internet address. All the networks listed are class A.
- ARPANET (network number 10): 24-bit addresses made of host, logical host and IMP fields are used directly as the 24-bit local address. The IMPs did not yet support logical addressing, so the logical host field is set to zero in the leader.
- AUTODIN II (26): a 16-bit host or terminal address sits in the low 16 bits, with a zero octet before it.
- DCNs: 16-bit addresses split into an 8-bit host identifier and an 8-bit process identifier. COMSAT-DCN is network 29 and UCL-DCN is network 30.
- EDN (21): the same layout as the ARPANET.
- LCSNET (18): the low order 24 bits of the 32-bit LCS NET address are used, as subnet, reserved and host octets.
- PRNET: 16-bit addresses, independent of location. Six packet radio networks are listed, with network numbers 1, 2, 5, 6, 9 and 20.
- SATNET (4): 16-bit logical host addresses, so several can belong to one physical host.
- WBCNET (28): a 16-bit address split into an 8-bit HAP number and an 8-bit local address.
Each network gets its own slot in the big address and the memo shows which boxes its own numbers fill, padding the rest with zeros.
§4References and Scope
The memo is a plain reference table with no procedures beyond the layouts above. It points to RFC 791 for the Internet Protocol, to a BBN note for the WBCNET address format, and to IEN 82 for the several LCS NET address formats.
It is only a lookup chart. It tells you where to find more details if you need them.
§5References
[1]RFC 791Postel, J. (ed.), "Internet Protocol - DARPA Internet Program Protocol Specification," RFC 791, USC/Information Sciences Institute, September 1981.September 1981
[2]Pershing J., "Addressing Revisited," Bolt Beranek and Newman Inc., W Note 27, May 1981.
[3]Noel Chiappa, David Clark, David Reed, "LCS Net Address Format," M.I.T. Laboratory for Computer Science Network Implementation, Note No.5, IEN 82, February 1979.