Chaosnet

2.1 The Ether ¶

The transmission medium of Chaosnet is called the ether.
Physically it is a coaxial cable, of the semi-rigid 1/2 inch low-loss type used for cable TV,
with 75-ohm termination at both ends. At each
network node a cable transceiver is attached to the cable. A 10-meter
flat cable connects the transceiver to an interface which is attached
to a computer’s I/O bus. A network node consists of this transceiver and
interface and a computer which executes a certain piece of software called
the Network Control Program (NCP), which manages and controls Chaosnet, in
addition to whatever application software justifies its existence in the
first place.

One network node at a time may seize the ether and transmit a
packet, which arrives at all other nodes; each node decides in hardware
whether to ignore the packet or to receive it. A single ether must be a
linear cable; it may not contain branches nor stubs, and the ends may not
be joined into a circle. The maximum length of an ether cable is about 1
kilometer. This is determined by dispersion and by DC attenuation. The
maximum number of nodes on a single ether is probably a few dozen. This is
determined by degradation of the electrical properties of the cable by the
connectors used to attach the transceivers.

The maximum length of an ether could be increased by using
repeaters (bidirectional digital amplifiers joining two pieces of cable).
In practice this is not done. Instead, the protocol provides for multiple
ethers, joined together by nodes called bridges which relay packets
from one ether to another. A bridge is typically a pdp11 computer with two
or more network interfaces attached to it. A bridge node usually
performs other tasks as well, such as interfacing terminals. Bridges attach
other network media as well as ethers; some computers connect to the network
through their manufacturer’s high speed computer-to-computer interface to
a nearby bridge, rather than being interfaced directly to an ether.
Asynchronous lines have also been used as Chaosnet media.

The form of information on the ether, the transceiver and
interface hardware, and the protocols which control who may seize the
ether are described in the following sections.

2.2 Packets ¶

The basic unit of transmission is called a packet. This is a
sequence of up to 4032 data bits, plus 48 bits of header information
used by the hardware. Packets’ bits are normally grouped into 16-bit
words. The division of a transmitted bit stream into packets provides a
conveniently-sized unit for resource allocation and error control. The job
of the hardware is to deliver a packet from one node to another. Software
protocols define the meaning of the data bits in a packet, manage the hardware,
compensate for imperfections of the hardware, and provide more useful services
than simple transmission of packets from one computer to another.

The hardware header consists of three 16-bit words, called
destination, source, and check. The source identifies the node
which transmitted this packet onto this ether. This is not necessarily the
original source of the message, since it may have originated on a
different ether. The destination identifies the node which is intended
to receive this packet from this ether. This is not necessarily the final
destination of the message; it may be a bridge which should relay the
packet to another ether, whence it will eventually reach its final destination.
The check word is a cyclic redundancy checksum, generated and checked
by hardware, which detects errors in transmission through the ether,
entirely-spurious packets created by noise on the cable, and memory
errors in the transmitting and receiving packet buffers.

The software protocol is also based on packets, taking 128 of the data bits
as a software header. This is described in Packets.

2.3 The Transceiver ¶

Everyone who connects to the ether does so through a transceiver,
which is a small box which mounts directly on the cable via a UHF connector
and a T-joint. All nodes use identical transceivers (the interface varies
depending on what computer it is designed to interface to). The
transceiver contains the analog portion of the interface logic, provides
ground isolation between the ether cable and the computer, and contains
some protective circuitry designed to prevent a malfunctioning program or
interface from continuously jamming the ether. (If it were to be
redesigned, it ought to contain even more protective circuitry, since there
are some possible interface failures that can get past the protection and
render the ether unusable.)

The transceiver receives a differential digital signal from the
computer interface and impresses it onto the cable as a level of about 8
volts for a 1, or 0 volts (open circuit) for a 0, through a very fast VMOS
power FET. When the cable is idle it is held at 0 volts by the
terminations. This simple-minded unipolar scheme is adequate for the
medium cable lengths and transmission speeds we are using. The transceiver
monitors the cable by comparing it against a reference voltage, and returns
a differential signal to the interface. In addition, it detects
interference (another transceiver transmitting at the same time as this
one) and informs the interface.

The transceiver includes indicators (light-emitting diodes) for
power OK, transmitted data, received data, and interface attempting to jam
the ether. A test button simulates an input of continuous 1’s from the
interface, which should light all the lights (dimly) if the transceiver is
working. These indicators and the test button are useful for rapidly
tracking down network problems. The transceiver requires its own power
supply mounted nearby; one power supply can service three transceivers if
they are all adjacent. High-voltage isolation between the cable and the
computer is provided by optical isolators within the transceiver.

2.4 The Interface ¶

The interface is typically a wire-wrap board containing about 120 TTL
logic chips, which plugs into the I/O bus of a computer and connects it
to the ether (through a transceiver.) The interface implements the hardware
protocols described in the next section, buffers incoming and outgoing
packets, generates and checks checksums, and interrupts the host computer
when a packet is to be read out of the receive packet buffer or stored
into the transmit packet buffer. These packet buffers shield the host computer
from the high speed of data transmission on the cable. Instead of having to
produce bits at a high rate, the host can produce them at a lower rate,
collect them into a packet, and then tell the interface to transmit the packet
in a single burst of high-speed transmission.

Interfaces currently exist for Lisp machines, for DEC LSI-11
micro-computers, and for the DEC Unibus [UNIBUS], which allows Chaosnet to
be attached to pdp11’s, VAX’s, and the peripheral processors of most
pdp10’s. Programming documentation for this compatible family of
interfaces appears later in this paper.

Interfaces for other computers are likely to exist in the future. The
existing interface design does not make any unusual special demands of
its host computer and should be easily adaptable to other architectures.

2.5 Hardware Protocols ¶

The purpose of these protocols is to deliver packets intact from
one node to another node on the same ether, with fairly high probability of
success, and to guarantee to give an error indication or lose the packet
entirely if it is not delivered intact. An additional purpose is to
provide high performance and not to bog down when subjected to a heavy
load.

Bits are represented on the ether using a technique which is called
Upright Biphase NRZI. Each bit cell, which is approximately 250
nanoseconds long, begins with a transition in state, from high to low or
from low to high. This transition marks the beginning of a bit cell and
provides self-clocking. 3/4 of the way through the bit cell, the state of
the cable is sampled; high represents a 1 and low represents a 0. If the
bit being represented is the same as the previous bit, there will be one
transition at the beginning of the bit cell and a second in the middle of
the bit cell. If the bit being represented is the opposite of the previous
bit, there will be no transition in the middle of the bit cell since the
clock transition will have set the cable to the desired state. The
AC frequency of the signal on the cable varies between 1/2 the bit rate and
the full bit rate. The information bit-rate is 4 million bits per second.

The self-clocking feature allows for slight variations in transmission
and cable propagation speed. However, since the 3/4 of a bit cell delay
is a fixed delay, only modest variations in speed can be tolerated. A crystal
clock is used as the source of the transmit timing in the interface.

Since there is always at least one state-transition per bit cell,
the states where the ether remains high or low for an appreciable time
are available for non-data uses. If the ether remains low for more than
about two bit cells, it is considered to be not-busy. This condition marks
the end of a packet and allows someone else to transmit. Note that if no
transceivers are active, the terminations will hold the ether low.

If the ether remains high for about two bit cells, this is an
“abort signal”. Abort signals are used for two purposes. If the
transceiver detects a collision (two nodes trying to transmit at the same
time), each transmitting interface ceases to transmit and sends an abort
signal (four bit cells long), which tells all receivers to ignore the aborted
packet and ensures that the other transmitter also aborts. Thus when a
collision occurs, the ether is cleared as soon as possible to help prevent
long tie-ups under conditions of heavy load. The other use for the abort
signal is in hardware flow-control. When a receiving interface determines
that an incoming packet is addressed to it, but its receive buffer already
contains a packet, it sends an abort signal which causes the transmitter to
stop. This serves the dual purpose of immediately informing the
transmitter that its message did not get through, and preventing the ether
from being tied up while a long packet is transmitted which the receiver
cannot receive.

Packets are transmitted over the ether in reverse bit-order, for
hardware convenience. The three header words, which to the software appear
to be at the end of the packet, are transmitted first, in the order check,
source, destination. The data words, in reverse order, follow. Words are
transmitted least-significant bit first. Of course, the software need not
be aware of this reversal; packets arrive at the destination in the same
form as they were created by the source. At the end of the packet, an
extra zero bit is appended to bring the ether to the low state so that an
extra spurious clock-transition will not be generated when it goes idle.
This bit is stripped off by the interface and is never seen by software.

The check word is used for error detection, as described above.
The source word is made available to the software, which ignores it in
most cases, and also serves to synchronize the clocks in the
collision-avoidance mechanism which is described below. The
destination word is compared by each receiver against its own address.
If they match, or if the destination is zero, or if the software
selects the “match any destination” mode, the packet is placed in the
receive packet buffer and the host computer is interrupted. The zero
destination feature is used for “broadcast” messages. Note that a
receiver whose packet buffer is full will only generate an abort signal
if the packet was specifically addressed to it.

2.6 Ether Contention ¶

Chaosnet has no centralized control element; when a network node
has a message to transmit, its interface seizes the ether and transmits
a packet. The time when it seizes the ether is determined only by state
inside that particular interface and by the local state of the cable at
the point where that interface’s transceiver is attached.

If two interfaces should decide to seize the ether and transmit
at the same time, their transmissions will interfere and no useful
information will be transmitted. This is called a collision.
Collisions are the principal limitation on the bandwidth of a
heavily-loaded ether-type network, and should be avoided. (However,
neither PARC’s network nor MIT’s network has yet been operated with a
heavy enough load to make collisions really significant.)

Chaosnet uses a novel collision avoidance technique. First of
all, an interface will never initiate transmission unless the ether is
seen to be not busy, i.e. it has been in the low state for some time.
This ensures that collisions can only occur near the beginning of a
packet. Once transmission of a packet has gotten well started, the
ether is effectively “seized” (all interfaces realize that it is busy)
and transmission will continue successfully through to the end of the
packet. The amount of ether transmission time wasted by a collided
packet is therefore limited to the round-trip cable propagation delay.
This technique is called carrier sense.

Secondly, the hardware uses a time-division technique to attempt
to prevent two interfaces from initiating transmission at the same time.
This technique should prevent essentially all collisions while imposing
only a modest delay in the initiation of transmission. It is designed
so that it works better as the load on the ether increases; the wasted time
between packets and the relative rate of collisions both decrease.

The basic idea is that each interface is assigned a time-slot,
or turn, according to its address. It may only initiate
transmission during its turn. The turns are spaced far enough apart
that if one interface initiates transmission, every other interface
will perceive that the ether is busy by the time its own turn arrives,
and will not initiate an interfering transmission. Each interface
contains a time-slot counter which counts while the ether is not busy,
keeping track of whose turn it is. Each packet synchronizes the
counters in all of the interfaces by setting them from the source
address of that packet; at the time the packet was transmitted, it must
have been the turn of the interface that transmitted it.

Another way to think of this is to make an analogy with ring
networks. One can imagine a virtual token which passes down the cable
until it gets to the end, then jumps to the beginning of the cable and
repeats. An interface may only initiate transmission at the instant the
token passes by it. When an interface transmits, the token stops moving
and remains at that interface until the end of the packet, whereupon it
continues down the cable, passing every other interface, giving them
each a chance to transmit before letting the first interface transmit a second packet.

The token is not represented by any physical transmission on the
cable. That would constitute a form of centralized control, and would
lead to reliability problems if the token was lost or duplicated.
Instead, every interface contains a time-slot counter which keeps
track of where the token is thought to be. Every time a packet is
transmitted these counters are brought up-to-date. The token cannot be
lost because a counter by its nature eventually returns to all previous
states. It does not matter if the token is duplicated (i.e. the
counters lose synchronization) occasionally, since this will only cause
collisions, which we know how to detect and deal with, and since the first successful
transmission will resynchronize all counters. The basic mechanism
of the ether network with contention and collisions is known to work,
and the collision-avoidance scheme is an added-on optimization which
improves performance without changing the basic mechanism.

There is a finite propagation delay time between interfaces,
and this time is not small compared with the bit-rate of Chaosnet, nor
when compared with the desirable length of a time slot.
This time consists of the delay in the cable, about 5 nanoseconds per meter,
and the delay through the two transceivers, about 220 nanoseconds.
This propagation delay means that the time-slot counters in two different
interfaces cannot be exactly synchronized, and that when interface A
initiates transmission interface B will not instantaneously see that the
ether is busy. Special relativity tells us that in fact the concept
“exactly synchronized” is meaningless. Since the two time-slot counters are
not in the same place, the only way we can compare them is to send a
message from one to the other, through the ether, containing the reading of
the counter. But this message takes non-zero time to get there, so the
counter-reading it contains is wrong by the time it is compared against the
other counter! We in fact do send messages containing counter readings;
the source address in a packet equals the reading of the time-slot counter in the
interface that sent it–at the time it was sent. Since the network nodes are
not in relative motion, we can measure the distance between them and use
that information to improve their synchronization.

What we are trying to do is to prevent collisions. This means that
if interface A starts transmitting a packet in its turn, then by the time
interface B thinks that its own turn has arrived, it must perceive the
ether as busy. We will assign addresses (and hence time slots) and set the
length of a time slot in such a way that this will happen. Suppose the
maximum delay through the ether between A and B is t. This would be
the delay for one of them sending a packet to the other; the delay between
A’s receipt of a third party’s packet and B’s receipt of that packet is
less, especially if the third party is between A and B on the cable. Then
the maximum perceived difference between a clock at A and a clock at B is
2t; if a message is sent from B to A synchronizing the clocks, and then
a message is sent from A to B containing A’s clock reading, at B this clock
reading will be slow by 2t.

When a packet transmitted by A arrives at B, B’s clock may read as
much as 2t earlier or later than A’s turn, depending on the transmission
direction of the last synchronizing message. In order to guarantee that B’s
turn has not yet happened, the time between any of A’s turns and any of B’s
turns must always be at least 2t, twice the maximum propagation delay
through the ether between A and B. This is the important idea! We cause
this to be true by assigning addresses starting at one end of the cable;
each node’s address is the previous node’s address plus twice the
propagation delay between them, divided by the length of a turn. It is
easy to see that if this is done for all adjacent pairs, the condition will
automatically be true for non-adjacent pairs as well. When we get to the
end of the cable, we must assign a number of empty slots equal to twice the
propagation delay of the full length of cable, to provide
the necessary separation between the turns of the two nodes at the ends
of the cable.

The virtual token travels through the network at a substantially
slower speed than a real signal such as a packet; in the fastest case, when nodes are very
far apart, it travels at just half the speed of a real signal. Since a
Chaosnet ether has the geometry of a line, as compared to the ring net’s
circle, the virtual token is also slowed down by the need to return from
the end of the cable to the beginning. This slower speed of the token is
the price one pays for the increased robustness of Chaosnet as compared
with a ring network. In a real system, we slow the token down even more to
provide a margin of safety. The speed of the network is not significantly
affected by the slow token, since the interval between packet transmissions
by a single node is much longer than the round-trip time of the token.
Indeed, if the network is being used primarily for file transfer, and hence
the packets are large, the transmission time alone for a typical packet is
several times the round-trip time of the token. A typical value for the
token’s round-trip time is 64 microseconds.

In spite of all this, sometimes collisions will occur anyway. If
the cable has been idle for a long time, the various clocks will have lost
synchronization. If a source address is corrupted by a transmission error,
any interface that sees that source address will set its clock to an
incorrect value. Sometimes a packet will collide with random noise rather
than another legitimate packet. In addition, the transmitter does not
distinguish receiver-busy aborts from real collisions.

When a collision does occur, we recover from it (in software) by
retransmitting the packet again a couple of times, hoping that we will be
lucky enough not to have another collision, or that the receiver will soon
clear its packet buffer. This retransmission is done by the software, not
the hardware, since the hardware destroys the packet in its packet buffer
in the process of transmitting it. But if collisions continue to occur, we
give up and let somebody else have the ether. The packet is lost. A
higher level of protocol will soon realize that it has been lost and
retransmit it. We assume that there is enough randomness in the
higher-level software that the two nodes which originally collided will not
collide again on the retransmission by deciding to retransmit at precisely
the same instant.

3.1 Connections ¶

The principal service provided by Chaosnet is a connection between
two user processes. This is a full-duplex reliable packet-transmission
channel. The network undertakes never to garble, lose, duplicate, or
resequence the packets; in the event of a serious error it may break
the connection off entirely, informing both user processes. User
programs may either deal in terms of packets, or ignore packet
boundaries and treat the connection as two uni-directional streams of
8-bit or 16-bit bytes.

On top of the connection facility “user” programs build other
facilities, such as file access, interactive terminal connections, and
data in other byte sizes, such as 36 bits. The meaning of the packets
or bytes transmitted through a connection is defined by the particular
higher-level protocol in use.

In addition to reliable communication, the protocol provides flow
control, includes a way by which prospective communicants may get in
touch with each other (called contacting or rendezvous), and
provides various network maintenance and housekeeping facilities. These are
discussed later.

3.2 Contact Names ¶

When first establishing a connection, it is necessary for the two
communicating processes to contact each other. In addition, in the
usual user/server situation, the server process does not exist beforehand
and needs to be created and made to execute the appropriate program.

We chose to implement contacting in an asymmetric way. (Once the
connection has been established everything is completely symmetric.)
One process is designated the user, and the other is designated the
server. The server has some contact name to which it
listens. The user process requests its local operating system to
connect it to the server, specifying the network node and contact name
of the server. The local operating system sends a message (a Request
for Connection) to the remote operating system, which examines the
contact name and creates a connection to a listening process, creates
a new server process and connects to it, or rejects the request.

Automatically discovering to which host to connect in order to obtain a
particular service is a subject for higher-level protocols and for
further research. It is not dealt with by Chaosnet.

Once a connection has been established, there is no more need for the
contact name and it is discarded. Indeed, often the contact name
is simply the name of a service (such as “TELNET“) and several
users should be able to have simultaneous connections to separate
instances of that service, so contact names must be reusable.

In the case where two existing processes that already know about each
other want to establish a connection, we arbitrarily designate one as
the listener (server) and the other as the requester (user). The
listener somehow generates a “unique” contact name, somehow
communicates it to the requester, and listens for it. The requester
requests to connect to that contact name and the connection is
established. In the most common case of establishing a second
connection between two processes which are already connected, the index
number (see below) of the first connection can serve as a unique
contact name.

Contact names are restricted to strings of upper-case letters, numbers,
and ASCII punctuation. The maximum length of a contact name is limited
only by the packet size, although on ITS hosts the names of
automatically-started servers are limited by the file-system to six characters.

See Connection Establishment for complete details of how to establish
a connection.

3.3 Addresses and Indices ¶

Each node (or host) on the network is identified by an address,
which is a 16-bit number. These addresses are used in the routing of
packets. There is a table (the system hosts table, SYSBIN;HOSTS2, in the case
of ITS) which relates symbolic host names to numeric host addresses.

An address consists of two fields. The most-significant 8 bits
identify a subnet, and the least-significant 8 bits identify a host
within that subnet. Both fields must be non-zero. A subnet
corresponds to a single transmission path. Some subnets are physical
Chaosnet cables (ethers), while others are other media, for
instance an interface between a pdp10 and a pdp11. The significance
of subnets will become clear when routing is discussed (see Routing).

When a host is connected to an ether, the host’s hardware address on that
ether is the same as its software address, including the subnet field.

A connection is specified by the names of its two ends. Such a name consists
of a 16-bit host address and a 16-bit connection index, which is assigned
by that host, as the name of the entity inside the host which owns the connection.
The only requirements placed by the protocol on indices
are that they be non-zero and that they be unique within a particular
host; that is, a host may not assign the same index number to two different
connections unless enough time has elapsed between the closing of the first
connection and the opening of the second connection that confusion between
the two is unlikely.

Typically the least-significant n bits of an index are used as a
subscript into the operating system’s tables, and the most-significant
16-n bits are incremented each time a table slot is reused, to
provide uniqueness. The number of uniquizing bits must be sufficiently
large, compared to the rate at which connection-table slots are reused,
that if two connections have the same index, a packet from the old
connection cannot sit around in the network (e.g. in buffers inside
hosts or bridges) long enough to be seen as belonging to the new connection.

It is important to note that packets are not sent between hosts
(physical computers). They are sent between user processes; more exactly,
between channels attached to user processes. Each channel has a 32-bit
identification, which is divided into subnet, host, index, and uniquization
fields. From the point of a view of a user process using the network, the
Network Control Program section of his host’s operating system is part of
the network, and the multiplexing and demultiplexing it performs is no
different from the routing performed by other parts of the network.
It makes no difference whether two communicating processes run in the same
host or in different hosts.

Certain control packets, however, are sent between hosts rather than users.
This is visible to users when opening a connection; a contact name is only
valid with respect to a particular host. This is a compromise in the design
of Chaosnet, which was made so that an operational system could be built
without first solving the research and engineering problems associated with
making a diverse set of hosts into a uniform, one-level name space.

3.4 Packet Numbers ¶

There are two kinds of packets, controlled and uncontrolled. Controlled
packets are subject to error-control and flow-control protocols, described
below (see Flow and Error Control), which guarantee that each controlled
packet is delivered to its destination exactly once, that the controlled
packets belonging to a single connection are delivered in the same order
they were sent, and that a slow receiver is not overwhelmed with packets
from a fast sender. Uncontrolled packets are simply transmitted; they will
usually but not always arrive at their destination exactly once. The
protocol for using them must take this into account.

Each controlled packet is identified by an unsigned 16-bit packet number.
Successive packets are identified by sequential numbers, with wrap-around
from all 1’s to all 0’s. When a connection is first opened, each end
numbers its first controlled packet (RFC or OPN) however it likes, and
that sets the numbering for all following packets.

Packet numbers should be compared modulo 65536 (2 to the 16th), to ensure correct
handling of wrap-around cases.
On a pdp11, use the instructions

Do not use the BLT or BLO instruction.
On a pdp10, use the instructions

SUB A,B
TRNE A,100000
 JRST A_is_less

Do not use the CAMGE instruction.
On a Lisp machine, use the code

(IF (BIT-TEST #o100000 (- A B))
    )

Do not use the LESSP (or <) function.

3.5 Packets ¶

A packet consists of a header, which is 8 16-bit words, and zero or
more 8-bit or 16-bit bytes of accompanying data. In addition there are
three words put on by the hardware, described earlier in this paper.

The following are the 8 header words:

Operation

The most-significant 8 bits of this word are the Opcode of the packet,
a number which tells what the packet means. The 128 opcodes with high-order
bit 0 are for the use of the network itself. The 128 opcodes with high-order
bit 1 are for use by users. The various opcodes are described in
Software Protocol–Details.

The least-significant 8 bits of this word are reserved for future use,
and must be zero.

Count

The most-significant 4 bits of this word are the forwarding count, which tells
how many times this packet has been forwarded by bridges. Its use is explained
in the Routing section.

The least-significant 12 bits of this word are the data byte count,
which tells the number of 8-bit bytes of data in the packet. The
minimum value is 0 and the maximum value is 488. Note that the count
is in 8-bit bytes even if the data are regarded as 16-bit bytes.

The byte count must be consistent with the actual length of the
hardware packet. Since the hardware cyclic redundancy check algorithm
is not sensitive to extra zero bits, packets whose hardware length
disagrees with their software length are discarded as hardware errors.

Destination Address

This word contains the network address of the destination host
to which this packet should be sent.

Destination Index

This word contains the connection index at the destination host
of the connection to which this packet belongs, or 0 if this packet
does not belong to any connection.

Source Address

This word contains the network address of the source host which
originated this packet.

Source Index

This word contains the connection index at the source host
of the connection to which this packet belongs, or 0 if this
packet does not belong to any connection.

Packet Number

If this is a controlled packet, this word contains its identifying number.

Acknowledgement

The use of this word is described in Flow and Error Control.

3.6 Data Formats ¶

Data transmitted through Chaosnet generally follow Lisp Machine
standards. Bits and bytes are numbered from right to
left, or least-significant to most-significant. The first 8-bit byte
in a 16-bit word is the one in the arithmetically least-significant
position. The first 16-bit word in a 32-bit double-word is the one
in the arithmetically least-significant position.

The character set used is dictated by the higher-level protocol in use.
Telnet and Supdup, for example, each specifies its own ASCII-based character set.
The “default” character set–used for new protocols and for text that
appears in the basic Chaosnet protocol, such as contact names–is
the Lisp Machine character set [CHINUAL]. This is basically ASCII,
augmented with additional printing characters and a different set of
format-effector (or “control”) characters.

Because the rules for bit numbering conflict with the native byte-ordering
in pdp10s, and because it is quite expensive to rearrange the bytes using
the pdp10 instruction set, pdp11s which act as front-ends for pdp10s must
reformat packets passing through them, and pdp10s interfaced directly to
the network must have interfaces capable of rearranging the bytes. This
requires that the network protocols explicitly specify which portions of
each type of packet are 8-bit bytes and which are 16-bit bytes. In general
the header is 16-bit bytes and the data field is 8-bit bytes, but certain
packet types (OPN, STS, RUT, and opcodes 300 through 377) have 16-bit bytes
in the data field. Use of 32-bit data is rare, so no provision is made for
putting 32-bit data into the standard format for pdp10s. On our current
network pdp10s are the only hosts which require this packet reformatting
assistance, because most modern computers number their bits and bytes from
least-significant to most-significant.

The effect of this is that user programs that use the Chaosnet always see the
data in a packet and its header in the native form of the machine they are
running on, and the necessary conversions are automatically applied by the
network. This statement applies to the order of bits and bytes within a word,
but not to the character set (when packets contain textual data) which is
dictated by protocols.

Unlike some other network protocols, Chaosnet does not use any software
checksumming. Because of the diversity of hosts with different
architectures attached to the Chaosnet, it is impossible to devise a
checksumming algorithm which can be executed compatibly and efficiently on
all hosts. Instead, Chaosnet relies on error-checking hardware in the
network interfaces, and assumes that other sources of packet damage that
checksums could detect, such as software bugs in a Network Control Program,
either do not occur or will produce symptoms so obvious that they will be
detected and fixed immediately.

3.7 Routing ¶

Routing consists of deciding how to deliver a packet to the network
node specified by the destination address field of the packet.
Having reached that node, the packet can trivially be delivered to the
destination user process via the destination index.
In general routing may be a multi-step process involving transmission
through several subnets, since there may not be a direct hardware
connection between the source and the destination. Note that the
routing decision is made separately for each packet, with no reference
to the concept of connections.

Any host that is connected to more than one subnet acts as a bridge
and forwards packets from one subnet to another when necessary.
There could also be hardware bridges which are not hosts, although we have
not yet designed any such device. Since routing does not depend on connections,
a bridge is a very simple device (or program) which does not need much state.
This makes the bridge function inexpensive to piggyback onto a computer which
is also performing other functions, and makes reliable bridge software easy
to implement.

The difference between a bridge and a gateway, in our terminology,
is that a bridge forwards packets from one sub-Chaosnet to another, without
modifying the packets or understanding them other than to look at the
destination address and increment the forwarding count, and does not deal
with connections nor with flow control, while a gateway interconnects two
networks with differing protocols and must understand and translate the
information passing through it. Gateways may also have to deal with flow
and error control because they connect networks with slow or differing
speeds. Bridges are suitable for local networks while gateways are
suitable for long-distance networks and for connecting networks not
produced by the same organization.

To prevent routing loops, each packet contains a forwarding-count field.
Each bridge that forwards the packet increments this count; if the count reaches
its maximum value the packet is discarded. The error-control protocol will
recover discarded packets, or decide that no viable connection can be established
between the two hosts.

The implementation of routing in an operating system is as follows, given a
packet to be routed, which may have come in from the network or may have
been originated by the local host. First, check the packet’s destination
address. If it is this host, receive the packet. Otherwise, increment the
forwarding count and discard the packet if it has been forwarded too many
times. If the destination is some other host on a subnet to which this
host is directly connected, transmit the packet on that subnet; the
destination host should receive it. If the destination is a host on a
subnet of which this host has no knowledge, look up the subnet in the
host’s routing table to find the best bridge to that subnet, and
transmit the packet to that bridge.

Each host has a routing table, indexed by subnet number, which tells
how to get packets to hosts on that subnet.
Each entry contains: (exact details may vary depending on implementation)

Type

The type of connection between the host and this subnet. This can
be one of Direct, Bridge, or Fixed Bridge. Direct means a physical
connection such as a Chaosnet interface. Bridge means
an indirect connection, via a packet-forwarding bridge. Which bridge
is best to use is to be discovered by this routing mechanism.
Fixed Bridge is the same except that the automatic mechanism
is not to change which bridge is used. This is useful to set up
explicit routing for purposes such as network debugging.

Address

Identifies the connection to this subnet in a way which depends on the type.
For a direct connection, this identifies the piece of hardware which
implements the connection. (It might be a Unibus address.)
For a bridge or a fixed bridge, this is the network address of the bridge.

Cost

A measure of the cost of sending a packet through this route.
Costs are used to select the best route from among alternatives in a way
described below. For a direct connection, the cost is 10 for a direct
interface between two computers (e.g. between a pdp10 and its front-end pdp11),
11 for a Chaosnet ether cable, 20 for a slow medium such as an asynchronous line, etc.
For a bridge or a fixed bridge, the cost is specified by the bridge in a RUT packet
(described below).

The routing table is initialized with the number of a more or
less arbitrary existent host and a high cost, for each subnet
to which the host is not directly connected. Until the correct
bridge is discovered (which normally happens within a minute
of coming up), packets for that subnet will be bounced off of
that arbitrary host, which probably knows the right bridge to forward them to.

The cost for subnets accessed via bridges is increased by 1 every 4
seconds, thus typically doubling after a minute. When the cost
reaches a “high” value, it sticks there, preventing problems with
arithmetic overflow. The purpose of the increasing cost is to discount
the value of old information. The cost for subnets accessed via direct
connections and fixed bridges does not increase.

Every 15 seconds, a bridge advertises its presence by broadcasting a
routing (RUT) packet on each subnet to which it is directly connected.
Each host on that subnet receives the RUT packet and uses it to update
its routing table. If the host’s routing table says to access a certain
subnet via bridges, and the RUT packet says that this is the best bridge
to that subnet, the routing table is updated to say that this bridge should
be used.

Note that it is important that the rate at which the costs increase with
time be slow enough that it takes more than twice the broadcast interval to
increase the cost of one hop to be more than the cost of two hops. Otherwise
the routing algorithm is not well-behaved. Suppose subnet A has two
bridges (alpha and beta) on it, and bridge alpha is connected to subnet
B but bridge beta is not (it goes to some other irrelevant subnet). Then
if the costs increase too fast and bridges alpha and beta do not
broadcast their RUT packets exactly simultaneously, sometimes packets for
subnet B may be sent to bridge beta because its cost appears lower.
Bridge beta will then send them to bridge alpha, where they should have
gone directly. In more complicated situations packets can go around in a
circle some of the time.

The source address of a RUT packet must be the hardware address of the
bridge on the particular subnet on which the packet is broadcast. The
destination address of a RUT packet must be zero; RUT packets are
not forwarded onto other subnets. The byte count of a RUT packet is a
multiple of 4 and the packet contains up to 122 pairs of 16-bit words:

word 1

The subnet number of a subnet which this bridge can get to,
directly or indirectly, right-adjusted.

word 2

The cost of sending to that subnet via this bridge. This is the
current cost from the bridge’s routing table, plus the cost
for the subnet on which the routing packet is being broadcast.
Adding the subnet cost eliminates loops, and prefers one-hop
paths over two-hop paths.

When a host receives a RUT packet, it processes each 2-word entry by
comparing the cost for that subnet against its current cost; if it is less or equal
the cost and the address of the bridge are entered into the routing table,
provided that that subnet’s routing table entry is not of the Direct
or Fixed Bridge type.

When there are multiple equivalent bridges, the traffic is spread among them only by
virtue of their RUT packets being sent at different times, so that sometimes one
bridge has the lower cost, and sometimes the other. If this isn’t adequate,
hosts could have hairier routing tables which remember more than one possible
route and use them according to their relative costs, but so far
this has not been necessary since the network traffic is not so high as
to saturate any one bridge.

The design of this routing scheme is predicated on the assumption that
the network geometry is simple, there are few multiple paths,
and the length of any path is quite short. This makes more sophisticated
schemes unnecessary.

An important feature of this routing scheme is that the size of the
table is proportional to the number of subnets, not to the number of hosts.
Thus it does not take up an inordinate amount of memory in a small computer,
and no complicated dynamic allocation schemes are required.

In the case of a pdp10 which accesses the Chaosnet through a front-end pdp11, we define
the interface between the two computers to be a subnet, and regard the pdp11
as a bridge which forwards packets between the network and the pdp10.
This gives the pdp10 and the pdp11 separate addresses so that we can
choose to talk to either one, even though they are part of the same
computer system. This is occasionally useful for maintenance purposes.
It becomes more useful when the front-end pdp11 has peripherals which
are to be accessed through the Chaosnet, since they can simply look like
hosts on that pdp11’s private subnet.

In the case of a host which is attached to more than one subnet,
it is undesirable for the host to have more than one address, since this
would complicate user programs which use addresses. Instead, one of
the host’s network attachments is designated as primary, and that address
is used as the host’s single address. The other attachments are regarded
as bridges which can forward to that host. Sometimes, we simplify the
routing by inventing a new subnet which contains only that host and
has no physical realization. The host’s address is an address on that
fake subnet. All of the host’s network attachments are regarded as bridges
which know how to forward packets to that subnet.

The ITS host table allows a host to have multiple addresses on multiple networks,
but when you ask for the address of a certain host on a certain network
you only get back the primary address. All packets coming from that host have
that as their source address.

3.8 Flow and Error Control ¶

The Network Control Programs (NCPs) conspire to ensure that data
packets are sent from user to user with no garbling, duplications, omissions, or
changes of order. Secondarily, the NCPs attempt to achieve a maximum
rate of flow of data, and a minimum of overhead and retransmission.

The fundamental basis of flow-control and error-control in Chaosnet is
retransmission. Packets which are damaged in transmission, which won’t
fit in buffers, which are duplicated or out-of-sequence, or which otherwise
are embarrassing are simply discarded. Packets are periodically retransmitted
until an indication that they have been successfully received is returned.
This retransmission is end-to-end; any intermediate bridges do not participate
in flow-control and error-control, and hence are free to discard any packets
they wish.

There are actually two kinds of packets, controlled and
uncontrolled. Controlled packets are retransmitted and delivered
reliably; most packets, including all packets used by the user (except for
UNC packets), are of this type. Uncontrolled packets are not
retransmitted; these are used for certain lower-level functions of the
protocol such as the implementation of flow and error control. The usage
of these packets is designed so that they need not be delivered reliably.

Retransmission of a packet continues until stopped by a signal from the receiver to
the sender called a receipt. A receipt contains a packet
number, and indicates that all controlled packets with a packet
number less than or equal (modulo 65536) to that number have been
successfully received, and therefore need not be retransmitted any
more. A receipt does not indicate that these packets have been
processed by the destination user process; it simply
indicates that they have successfully arrived in the destination host,
and are guaranteed to be there when the user process asks for them.

There is another signal from the receiver to the sender, called
an acknowledgement. An acknowledgement also contains a packet
number, and indicates that all controlled packets with a packet number
less than or equal (modulo 65536) to that number have been read
by the destination user process. This is used to implement flow-control.
Note that acknowledgement of a packet implies receipt of that packet.
In fact, if the receiving process does not fall behind,
explicit receipts need not be sent, because the receiving host will not
have to buffer any packets, but will acknowledge them as soon as they arrive.

The purpose of flow-control is to match the speeds of the sending and
receiving processes. The extremes to be avoided are, on the one hand,
too small a “buffer size” causing the data transmission rate to be slower
than it could be, and on the other hand, large numbers of packets piling
up in the network because the sender is sending faster than the receiver
is receiving. It is also necessary to be aware that receipts and
acknowledgements must be transmitted through the network, and hence
have an associated cost.

Chaosnet flow-control operates by controlling the number of packets “in
the network”. These are packets which have been emitted by the sending
user process, but have not been acknowledged. We define a window
into the set of packet numbers. The beginning of this window is the
first packet number which has not been acknowledged, and the width of
the window is a fixed number established when the connection is opened.
The sending process is only allowed to emit packets whose packet
numbers lie within the window. Once it has emitted all of the packets
in the window, the window is said to be full. Thus, the size of the
window is the “buffer size” for the connection, and is the maximum
number of packets that may need to be buffered inside an NCP (sending
or receiving). Acknowledgements move the window, making it not full,
and allowing the sending process to emit additional packets.

We do not receipt and acknowledge every single controlled packet that is transmitted
through a connection, since that would double or triple the number of packets
sent through the network to move a given amount of data. Instead we batch
the receipts and acknowledgements. But if acknowledgements are not sent sufficiently
often, the data will not flow smoothly, because the window will often appear
full to the sender when it is not. If receipts are not sent sufficiently
often, there will be unnecessary retransmissions.

Whenever a packet is sent through a connection, an acknowledgement for the
reverse direction of that connection is “piggy-backed” onto it, using the
Acknowledgement field in the packet header. For interactive applications,
where there is much traffic in both directions, this provides all the necessary
acknowledgement and receipting with no need to send any extra packets through
the network.

When this does not suffice, STS (status) packets are generated to carry
receipts and acknowledgements. STS packets are uncontrolled, since they
are part of the mechanism that implements controlled packets. If an STS
packet is duplicated, it does no harm. If an STS packet is lost, mechanisms
exist which will cause a replacement to be generated later. An STS packet
carries separate receipt and acknowledgement packet numbers.

When a user process reads a packet from the network, if the number of packets which
should have been acknowledged but have not been is more than 1/3 the window size,
an STS is generated to acknowledge them. Thus the preferred batch size for acknowledgement
is 1/3 the window size. The advantage of this size is that if one STS is lost,
another will be generated before the window fills up (at the 2/3 point).

When a packet is received with the same packet number as one which has already been
successfully received, this is evidence of unnecessary retransmission, and
an STS is generated to carry a receipt back to the sender. If this STS is lost,
the next retransmission will stimulate another one. Thus receipts are normally
implied by acknowledgements, and only sent separately when there is evidence of
unnecessary retransmission.

Retransmission consists of sending all unreceipted controlled packets,
except those that were last sent very recently (within 1/30’th of a
second in ITS.) Retransmission occurs every 1/2 second. This interval
is somewhat arbitrary, but should be close to the response time of the
systems involved. Retransmission also occurs in response to an STS
packet, so that a receiver may cause a faster retranmission rate than
twice a second if it so desires. This should never cause useless
retransmission, since STS carries a receipt, and
very-recently-transmitted packets, which might still be in transit
through the network, are not retransmitted.

Another operation is probing, which consists of sending a SNS
packet, in the hope of eliciting either an STS or a LOS, depending on
whether the other side believes the connection exists. Probing is used
periodically as a way of testing that the connection is still open, and
also serves as a way to get STS packets retransmitted as a hedge
against the loss of an acknowledgement, which could otherwise stymie
the connection. SNS packets are uncontrolled.

We probe every five seconds on connections which have unacknowledged
packets outstanding (a non-empty window) and on connections which have not
received any packets (neither data nor control) for one minute. If a
connection receives no packets for 1 1/2 minutes, this means that at least
5 probes have been ignored, and the connection is declared to be broken;
either the remote host is down or there is no viable path through the
network between the two hosts.

The receiver can generate “spontaneous” STS’s, to stimulate
retransmission and keep things moving on fast devices with insufficient
buffering for 1/2 second worth of packets. This provides a way for the
receiver to speed up the retransmission timeout in the sender, and to
make sure that acknowledges are happening often enough.

Note that the network still functions if either or both parties to a
connection ignore the window. The window is simply an improver of
efficiency. Receipts have the same property. This allows very small
implementations to be compatible with the same protocol, which is
useful for applications such as bootstrapping through the network.

It would be possible to have dynamic adjustment of the window size in
response to observed behavior. The STS packet includes the window size so
that changes to it can be communicated. However, this has not been found
necessary in practice. Each higher-level protocol has a standard
pre-determined window size, which it establishes when it first opens a
connection, and this seems to be close enough to optimum that careful
dynamic adjustment of it wouldn’t make a big difference.

This scheme for flow-control and error-control is based on several
assumptions. It is assumed that the underlying transmission media have
their own checking, so that they discard all damaged packets, making packet
checksums unnecessary at the protocol level. The transit time through the
network is assumed to be fast, so that a fairly-small retransmission
interval is practical, and negative acknowledgements are not necessary.
The error rate is assumed to be low so that overall efficiency is not
affected by the simple-minded error recovery scheme of simply
retransmitting all outstanding packets. It is assumed that no reformatting
of packets occurs inside the network, so that flow-control and
error-control can operate on a packet basis rather than a byte basis.

4.1 Connection Establishment ¶

The following packet types are associated with creating and destroying
connections. First the packets are described and then the details of the
various connection-establishment protocols are given.

Opcode: 1 RFC Request for connection ¶

All connections are initiated by the transmission of an RFC from the user to the server.
The data field of the packet contains the contact name. The contact name can
be followed by arbitrary arguments to the server, delimited by a space character.
The destination index field of an RFC contains 0 since the destination index is
not known yet.

RFC is a controlled packet; it is retransmitted until some sort of response
is received. Because RFC’s are not sent over normal, error-controlled
connections, a special way of detecting and discarding duplicates is
required. When an NCP receives an RFC packet, it checks all pending RFC’s
and all connections which are in the Open or RFC-received state (see
Connection States), to see if the source address and index match; if so,
the RFC is a duplicate and is discarded.

Opcode: 12 LSN Listen ¶

A server process informs the local NCP of the contact name to which it is
listening by sending a LSN packet, with the contact name in the data field.
This packet is never transmitted anywhere through the network. It simply
serves as a convenient buffer to hold the server’s contact name. When an
RFC and a LSN containing the same contact name meet, the LSN is discarded
and the RFC is given to the server, putting its connection into the
RFC-received state (see Connection States). The server reads the RFC
and decides whether or not to open the connection.

Opcode: 2 OPN Open connection ¶

OPN is the usual positive response to RFC. The source index field conveys the
server’s index number to the user; the user’s index number was conveyed in the RFC.
The data field of OPN is the same as that of STS (see below); it serves mainly to convey
the server’s window-size to the user. The Acknowledgement field of the OPN
acknowledges the RFC so that it will no longer be retransmitted.

OPN is a controlled packet; it is retransmitted until it is acknowledged.
Duplicate OPN packets are detected in a special way; if an OPN is received
for a connection which is not in the RFC-sent state (see Connection States),
it is simply discarded
and an STS is sent. This can happen if the connection is opened while a
retransmitted OPN packet is in transit through the network, or if the STS
which acknowledges an OPN is lost in the network.

Opcode: 3 CLS Close connection ¶

CLS is the negative response to RFC. It indicates that no server was
listening to the contact name, and one couldn’t be created, or for
some reason the server didn’t feel like accepting this request for a
connection, or the destination NCP was unable to complete the
connection (e.g. connection table full.)

CLS is also used to close a connection after it has been open for a
while. Any data packets in transit may be lost. Protocols which
require a reliable end-of-data indication should use the mechanism for
that (see End-of-Data) before sending CLS.

The data field of a CLS contains a character-string explanation of the
reason for closing, intended to be returned to a user as an error message.

CLS is an uncontrolled packet, so that the program which sends it may
go away immediately afterwards, leaving nothing to retransmit the
CLS. Since there is no error recovery or retransmission mechanism for
CLS, the use of CLS is necessarily optional; a process could simply
stop responding to its connection. However, it is desirable to send a
CLS when possible to provide an error message for the user.

Opcode: 4 FWD Forward a request for connection ¶

This is a response to RFC which indicates that the desired service is
not available from the process contacted, but may be available at a
possibly-different contact name at a possibly-different host. The data
field contains the new contact name and the Acknowledgement
field–exceptionally–contains the new host number. The issuer of the
RFC should issue another RFC to that address. FWD is an uncontrolled
packet; if it is lost in the network, the retransmission of the RFC
will presumably stimulate an identical FWD.

Opcode: 5 ANS Answer to a simple transaction ¶

This is another kind of response to RFC. The data field contains
the entirety of the response, and no connection is established.
ANS is an uncontrolled packet; if it is lost in the network, the
retransmission of the RFC will presumably stimulate an identical ANS.

There are several connection-initiation protocols implemented with
these packet types. In addition to those described here, there is also
a broadcast mechanism; see Broadcast.

When an RFC arrives at a host, the NCP finds a user process that is
listening for this RFC’s contact name, or creates a server process to
provide the desired service, or responds to the RFC itself if it knows how
to provide the requested service, or refuses the request for connection.
The process that serves the RFC chooses which connection-initiation
protocol to follow. This process is given the RFC as data, so that it can
look at the contact name and any arguments that may be present.

A stream connection is initiated by an RFC, transmitted from user to
server. The server returns an OPN to the user, which responds with an STS.
These three packets convey the source and destination addresses, indices,
initial packet numbers, and window sizes between the two NCP’s. In
addition a character-string argument can be conveyed from the user to the
server in the RFC.

The OPN serves to acknowledge the RFC and extinguish its retransmission.
It also carries the server’s index, initial packet number, and window size.
The STS serves to acknowledge the OPN and extinguish its retransmission.
It also carries the user’s window size; the user’s index and initial packet
number were carried by the RFC. Retransmission of the RFC and the OPN provides
reliability in the face of lost packets. If the RFC is lost, it will
be retransmitted. If the STS is lost, the OPN will be retransmitted. If
the OPN is lost, the RFC will be retransmitted superfluously and the OPN
will be retransmitted since no STS will be sent.

The exchange of an OPN and an STS tells each side of the connection that the
other side believes the connection is open; once this has happened data may
begin to flow through the connection. The user process may begin
transmitting data when it sees the OPN. The server process may begin
transmitting data when it sees the STS. These rules ensure that data
packets cannot arrive at a receiver before it knows and agrees that the
connection is open. If data packets did arrive before then, the receiver
would reject them with a LOS (see below), believing them to be a violation
of protocol, and this would destroy the connection before it was ever fully
established.

Once data packets begin to flow, they are subject to the flow and error control
protocol described in Flow and Error Control. Thus a stream connection
provides the desired reliable, bidirectional data stream.

A refusal is initiated by an RFC in the same way, but
the server returns CLS rather than OPN. The data field of the CLS contains
the reason for refusal to connect.

A forwarded connection is initiated by an RFC in the same way,
but the server returns a FWD, telling the user another place to look
for the desired service.

A simple transaction is initiated by an RFC from user to server, and completed
by an ANS from server to user. Since a full connection is not established
and the reliable-transmission mechanism of connections is not used, the user
process cannot be sure how many copies of the RFC the server saw, and the server
process cannot be sure that its answer got back to the user. This means that
simple transactions should not be used for applications where it is important
to know whether the transaction was really completed, nor for applications
in which repeating the same query might produce a different answer.
Simple transactions are a simple efficient mechanism for applications such as extracting
a small piece of information (e.g. the time of day) from a central data-base.

4.2 Status ¶

Opcode: 7 STS Status ¶

STS is an uncontrolled packet which is used to convey status
information between NCPs. The Acknowledgement field in the packet
header contains an acknowledgement, that is, the packet number of the
last packet given to the receiving user process. The first 16-bit byte
in the data field contains a receipt, that is, a packet number such
that all controlled packets up to and including that one have been
successfully received by the NCP. The second 16-bit byte in the data
field contains the window size for packets sent in the opposite direction
(to the end of the connection which sent the STS).
The byte count is presently always 4.
This will change if the protocol is revised to add additional items to
the STS packet.

Opcode: 6 SNS Sense status ¶

SNS is an uncontrolled packet whose sole purpose is to cause the
other end of the connection to send back an STS. This is used by
the probing mechanism described above (see Probing).

Opcode: 11 LOS Lossage ¶

LOS is an uncontrolled packet which is used by one NCP to inform another of
an error. The data field contains a character-string explanation of the
problem. The source and destination addresses and indices are simply the
destination and source addresses and indices, respectively, of the
erroneous packet, and do not necessarily correspond to a connection. When
an NCP receives a LOS whose destination corresponds to an existent
connection and whose source corresponds to the supposed other end of that
connection, it breaks the connection and makes the data field of the
LOS available to the user as an error message. Other LOS’s, that don’t
correspond to connections, are simply ignored.

LOS is sent in response to situations such as: arrival of a data packet
or an STS for a connection that does not exist or is not open, arrival
of a packet from the wrong source for its destination, arrival of a packet
containing an undefined opcode or too large a byte count, etc.

LOS’s are given to the user process so that it may read the error message.

No LOS is given in response to an OPN to a connection not in the RFC-Sent
state, nor in response to a SNS to a connection not in the Open state,
nor in response to a LOS to a non-existent or broken connection. These
rules are important to make the protocols work without timing errors.
An OPN or a SNS to a non-existent connection elicits a LOS.

4.3 Data ¶

Opcode: 200-277 DAT 8-bit Data ¶

Opcodes 200 through 277 (octal) are controlled packets with user data
in 8-bit bytes in the data field. The NCP treats all 64 of these opcodes
identically; some higher-level protocols use the opcodes for their
own purposes. The standard default opcode is 200.

Opcode: 300-377 DAT 16-bit Data ¶

Opcodes 300 through 377 (octal) are controlled packets with user data
in 16-bit bytes in the data field. The NCP treats all 64 of these
opcodes identically; some higher-level protocols use the opcodes for their
own purposes. The standard default opcode for 16-bit data is 300.

Opcode: 15 UNC Uncontrolled Data ¶

This is an uncontrolled packet with user data in 8-bit bytes in
the data field. It exists so that user-level programs may bypass
the flow-control mechanism of Chaosnet protocol. Note that
the NCP is free to discard these packets at any time, since they
are uncontrolled. Since UNC’s are not subject to flow control, discarding
may be necessary to avoid running out of buffers. A connection may not
have more input packets queued awaiting the attention of the user program
than the window size of the connection, except that you are always allowed
to have one UNC packet queued. If no normal data packets are in use,
up to one more UNC packet than the window size may be queued.

UNC packets are also used by the standard protocol for encapsulating
packets of foreign protocols for transmission through Chaosnet
(see Using Foreign Protocols in Chaosnet).

4.4 End-of-Data ¶

Opcode: 14 EOF End of File ¶

EOF is a controlled packet which serves as a “logical end of data” mark
in the packet stream. When the user program is ignoring packets and
treating a Chaosnet connection as a conventional byte-stream I/O device,
the NCP uses the EOF packet to convey the notion of conventional end-of-file
from one end of the connection to the other. When the user program
is working at the packet level, it may transmit and receive EOF’s.

It is illegal to put data in an EOF packet; in other words, the byte
count should always be zero. Most Chaosnet implementations will simply
ignore any data that is present in an EOF.

EOF packets are used in the following protocol which is the recommended
way to reliably determine that all data have been transferred before
closing a connection (in applications where that is an important
consideration).

The important issue is that neither side may send a CLS until both sides
are sure that all the data have been transmitted. After sending all the
data it is going to send, including an EOF packet to mark the end, the
sending process waits for all packets to be acknowledged. This ensures
that the receiver has seen all the data and knows that no more data are to
come. The sending process then closes the connection. When the receiving
process sees an EOF, it knows that there are no more data. It does not
close the connection until it sees the sender close it, or until a brief
timeout elapses. The timeout is to provide for the case where the sender’s
CLS gets lost in the network (CLS cannot be retransmitted). The timeout is
long enough (a few seconds) to make it unlikely that the sender will not
have seen the acknowledgement of the EOF by the time the timeout is over.

To use this protocol in a bidirectional fashion, where both parties to the
connection are sending data simultaneously, it is necessary to use an
asymmetrical protocol. Arbitrarily call one party the user and the other
the server. The protocol is that after sending all its data, each party
sends an EOF and waits for it to be acknowledged. The server, having seen
its EOF acknowledged, sends a second EOF. The user, having seen its EOF
acknowledged, looks for a second EOF and then sends a CLS and goes
away. The server goes away when it sees the user’s CLS, or after a brief
timeout has elapsed. This asymmetrical protocol guarantees that each side
gets a chance to know that both sides agree that all the data have been
transferred. The first CLS will only be sent after both sides have waited
for their (first) EOF to be acknowledged.

4.5 Broadcast ¶

Chaosnet includes a generalized broadcast facility, intended to satisfy needs such as:

• Locating services when it is not known what host they are on.
• Internal communications of other protocols using Chaosnet as a transmission
  medium, such as routing in their own address spaces.
• Reloading and remote debugging of Chaosnet bridge computers.
• Experiments with radically different protocols.

Opcode: 16 BRD Broadcast ¶

A BRD packet works much like an RFC packet; it contains the name of a
server to be communicated with, and possibly some arguments. Unlike an
RFC, which is delivered to a particular host, a BRD is broadcast to all
hosts. Only hosts which understand the service it is looking for will
respond. The response can be anything which is valid as a response to RFC.
Typically BRD will be used in a simple-transaction mode, and the response
will be an ANS packet. Actually it can be any number of ANS packets since
multiple hosts may respond. BRD can also be used to open a full
byte-stream connection to a server whose host is not known. In this case
the response will be an OPN packet; only the first OPN succeeds in opening
a connection. CLS is also a valid response, but only as a true negative
response; BRD’s for unrecognized or unavailable services should be ignored
and no CLS should be sent, since some other host might be able to provide
the service.

The TIME and STATUS protocols (see Higher-Level Protocols) will work
through BRD packets as well as RFC packets. I don’t think there are any
other standard protocols that need to be able to work with BRD packets.

The data field of a BRD contains a subnet bit map followed by a contact
name and possible arguments. The subnet bit map has a “1” for each subnet
on which this packet is to be broadcast to all hosts; these bits are turned
off as the packets flow through the network, to avoid loops. The sender
initializes the bit map with 1’s for whichever subnets he desires (often
all of them).

In the packet header, the destination host and index are 0. The source
host and index are who to send the reply (ANS or OPN) to. The
acknowledgement field contains the number of bytes in the bit map (this
would normally be 32, but may be changed in the future). The number of
bytes in the bit map is required to be a multiple of 4. Bits in the bitmap
are numbered from right to left within a byte and from earlier to later
bytes; thus the bit for subnet 1 is the bit with weight 2 in the first byte
of the data field. Bits that lie outside of the declared length of the
bit map are considered to be zero; thus the BRD is not transmitted to those
subnets.

After the subnet bit map there is a contact name and arguments, exactly as
in an RFC. Operating systems should treat incoming BRD packets exactly
like RFC, even to the extent that a contact name of STATUS must retrieve
the host’s network throughput and error statistics. BRD packets will never
be refused with a “CLS”, however; broadcast requests to nonexistent servers
should simply be ignored, and no CLS reply should be sent. Most operating
systems will simplify incoming BRD handling for themselves and their users
by reformatting incoming BRD packets to look like RFC’s; deleting the subnet
bit map from the data field and decreasing the byte count. For consistency
when this is done the bit map length (in the acknowledgement field) should
be set to zero. The packet opcode will remain BRD (rather than RFC).

Operating systems should handle outgoing BRD packets as follows. When a user
process transmits a BRD packet over a closed connection, the connection
enters a special “Broadcast Sent” state. In this state, the user process
is allowed to transmit additional BRD packets. All incoming packets
other than OPN’s should be made available for the user process to read,
until the allowed buffering capacity is exceeded; further incoming packets
are then simply discarded. These incoming packets would normally be expected
to consist of ANS, FWD, and CLS packets only. If an OPN is received, and
there are no queued input packets, a regular byte-stream connection is opened.
Any OPN’s from other hosts elicit a LOS reply as usual, as do any ANS’s, CLS’s,
etc. received at this point.

Operating systems should not retransmit BRD packets, but should leave this
up to the user program, since only it knows when it has received enough
answers (or a satisfactory answer).

BRD packets can be delivered to a host in multiple copies when there are
multiple paths through the network between the sender and that host. The
bit map only serves to cut down looping more than the forwarding-count
would, and to allow the sender to broadcast selectively to portions of the
net, but cannot eliminate multiple copies. The usual mechanisms for
discarding duplicated RFC’s will also cause most duplicated BRD’s to be
discarded.

BRD packets put a noticeable load on every host on the network, so they
should be used judiciously. “Beacons” that send a BRD every 30 seconds all
day long should not be used.

4.6 Low-level ¶

Opcode: 13 MNT Maintenance ¶

MNT is a special packet type reserved for the use of network
maintenance programs. Normal NCPs should simply discard any MNT
packets they receive. MNT packets are an escape mechanism to allow
special programs to send packets that are guaranteed not to get
confused with normal packets. MNT packets are forwarded by bridges
although usually one would not be depending on this.

Opcode: 10 RUT Routing Information ¶

RUT is a special packet type broadcast by bridges to inform other
nodes of the bridge’s ability to forward packets between subnets.
The source address is the network address of the bridge on the
subnet on which the RUT was broadcast. The destination address is zero.
The byte count is a multiple of 4, and the data field contains a series
of pairs of 16-bit bytes: a subnet number and the “cost” of getting to
that subnet via this bridge. The packet number and acknowledgement
fields are not used and should contain zero. See Routing for the details.

4.7 Connection States ¶

A user process gets to Chaosnet by means of a capability or channel
(dependent on the host operating system) which corresponds to one
end of a connection. Associated with this channel are a number of
buffers containing controlled packets output by the user and not
yet receipted, and data packets received from the network but
not yet read by the user; some of these incoming packets are
in-order by packet number and hence may be read by the user, while
others are out of order and cannot be read until packets earlier
in the stream have been received. Certain control packets are also
given to the user as if they were data packets. These are RFC,
ANS, CLS, LOS, EOF, and UNC. EOF is the only type that can ever be
out-of-order.

Also associated with the channel is a state, usually called the
connection state. Full understanding of these states depends
on the descriptions of packet-types above. The state can be one of:

Open

The connection exists and data may be transferred.

Closed

The channel does not have an associated connection. Either it
never had one or it has received or transmitted a CLS packet,
which destroyed the connection.

Listening

The channel does not have an associated connection, but it has a contact
name (usually contained in a LSN packet) for which it is listening.

RFC Received

A Listening channel enters this state when an RFC arrives.
It can become Open if the user process accepts the request.

RFC Sent

The user has transmitted an RFC. The state will change to Open or
Closed when the reply to the RFC comes back.

Broadcast Sent

The user has transmitted a BRD. In this state, the user process
is allowed to transmit additional BRD packets. All incoming packets
other than OPN’s are made available for the user process to read,
until the allowed buffering capacity is exceeded; further incoming packets
are then simply discarded. These incoming packets would normally be expected
to consist of ANS, FWD, and CLS packets only. If an OPN is received, and
there are no queued input packets, a regular byte-stream connection is opened
(the connection enters the Open state).
Any OPN’s from other hosts elicit a LOS reply as usual, as do any ANS’s, CLS’s,
etc. received at this point.

Lost

The connection has been broken by receiving a LOS packet.

Incomplete Transmission

The connection has been broken because the other end has ceased
to transmit and to respond to SNS. Either the network or the foreign
host is down. (This can also happen if the local host goes down for
a while and then is revived, if its clock runs in the meantime.)

Foreign

The channel is talking some foreign protocol, whose packets are encapsulated
in UNC packets. As far as Chaosnet is concerned there is no connection.
See Using Foreign Protocols in Chaosnet for the details.

5.1 Status ¶

All network nodes, even bridges, are required to answer RFC’s with contact
name STATUS, returning an ANS packet in a simple transaction. This
protocol is primarily used for network maintenance. The answer to a
STATUS request should be generated by the Network Control Program,
rather than by starting up a server process, in order to provide rapid
response.

The STATUS protocol is used to determine whether a host is up,
to determine whether an operable path through the network exists between
two hosts, to monitor network error statistics, and to debug new Network
Control Programs and new Chaosnet hardware. The hostat function
on the Lisp machine, and the Hostat command of the CHATST program
on ITS are user ends for this protocol.

The first 32 bytes of the ANS contain the name of the node, padded on
the right with zero bytes. The rest of the packet contains blocks of
information expressed in 16-bit and 32-bit words, low byte first
(pdp11/Lisp machine style). The low-order half of a 32-bit word comes
first. Since ANS packets contain 8-bit data (not 16-bit), machines such
as pdp10s which store numbers high byte first will have to shuffle the
bytes when using this protocol. The first 16-bit word in a block is its
identification. The second 16-bit word is the number of 16-bit words to
follow. The remaining words in the block depend on the identification.

This is the only block type currently defined. All items are optional,
according to the count field, and extra items not defined here may be
present and should be ignored. Note that items after the first two
are 32-bit words.

word 0

A number between 400 and 777 octal. This is 400 plus a subnet number.
This block contains information on this host’s direct connection to
that subnet.

word 1

The number of 16-bit words to follow, usually 16.

words 2-3

The number of packets received from this subnet.

words 4-5

The number of packets transmitted to this subnet.

words 6-7

The number of transmissions to this subnet aborted by collisions or
because the receiver was busy.

words 8-9

The number of incoming packets from this subnet lost because the
host had not yet read a previous packet out of the interface and consequently
the interface could not capture the packet.

words 10-11

The number of incoming packets from this subnet with CRC errors.
These were either transmitted wrong or damaged in transmission.

words 12-13

The number of incoming packets from this subnet which had no CRC
error when received, but did have an error after being read out of
the packet buffer. This error indicates either a hardware problem with the
packet buffer or an incorrect packet length.

words 14-15

The number of incoming packets from this subnet which were rejected
due to incorrect length (typically not a multiple of 16 bits).

words 16-17

The number of incoming packets from this subnet rejected for
other reasons (e.g. too short to contain a header, garbage byte-count,
forwarded too many times.)

If the identification is a number between 0 and 377 octal, this is an obsolete
format of block. The identification is a subnet number and the counts are as
above except that they are only 16 bits instead of 32, and consequently may overflow.
This format should no longer be sent by any hosts.

Identification numbers of 1000 octal and up are reserved for future use.

5.2 Pulsar ¶

For network maintenance purposes, certain network nodes support a simple
transaction with contact name PULSAR, which controls a “pulsar”
feature. This feature periodically transmits a short packet which can be
used to test and adjust cable transceivers. The packet consists of the
three header words, a zero word, and a word of alternating ones and zeros.
It is addressed to host 177777 which is guaranteed not to exist.

The returned ANS contains a single character, which is a digit. A 0
means that the pulsar is turned off. Any other digit indicates the number
of sixtieths of a second between pulses. Sending an RFC with a digit
as an argument sets the state of the pulsar to that digit, and returns
an ANS containing the new state. Pulsars should be off by default, and
should only be turned on when debugging the network. The waste of cable
bandwidth and machine resources is negligible except in extremely large
networks, since pulsar packets are so short, but when debugging or making
measurements on cables using pulsar packets it is important to know where
the packets are coming from.

Bridge nodes which implement the PULSAR protocol and possess more
than one network interface should should have a single pulsar which
transmits on all network interfaces, rather than bothering to provide a
more complex protocol by which pulsars on the individual interfaces could
be turned on and off.

5.3 Telnet and Supdup ¶

The Telnet and Supdup protocols of the Arpanet [TELNET] [SUPDUP] exist in
identical form in Chaosnet. These protocols allow access to a
computer system as an interactive terminal from another network node.

The contact names are TELNET and SUPDUP. The direct borrowing of
the Telnet and Supdup protocols was eased by their use of 8-bit byte
streams and only a single connection. Note that these protocols define
their own character sets, which differ from each other and from the
Chaosnet standard character set.

Chaosnet contains no counterpart of the INR/INS attention-getting feature
of the Arpanet. The Telnet protocol sends a packet with opcode 201 in
place of the INS signal. This is a controlled packet and hence does not
provide the “out of band” feature of the Arpanet INS, however it is
satisfactory for the Telnet “interrupt process” and “discard output”
operations on the kinds of hosts attached to Chaosnet.

5.4 File Access ¶

The FILE protocol is primarily used by Lisp machines to access files
on network file servers. ITS and TOPS-20 are equipped to act as
file servers. A user end for the file protocol also exists for TOPS-20 and
is used for general-purpose file transfer. For complete documentation
on the file protocol, see [FILE]. The Arpanet file transfer protocols
have not been implemented on the Chaosnet (except through the Arpanet gateway
described below).

5.5 Mail ¶

The MAIL protocol is used to transmit inter-user messages through the Chaosnet.
The Arpanet mail protocol was not used because of its complexity and poor
state of documentation. This simple protocol is by no means the last word
in mail protocols; however, it is adequate for the mail systems we presently
possess.

The sender of mail connects to contact name MAIL and establishes a stream connection.
It then sends the names of all the recipients to which the mail is to be sent
at (or via) the server host. The names are sent one to a line and terminated
by a blank line (two carriage returns in a row). The Lisp Machine character
set is used. A reply (see below) is immediately returned for each recipient.
A recipient is typically just the name of a user, but it can be a user-atsign-host
sequence or anything else acceptable to the mail system on the server machine.
After sending the recipients, the sender sends the text of the message, terminated
by an EOF. After the mail has been successfully swallowed, a reply is sent.
After the sender of mail has read the reply, both sides close the connection.

In the MAIL protocol, a reply is a signal from the server to the user
(or sender) indicating success or failure. The first character of a reply
is a plus sign for success, a minus sign for permanent failure (e.g. no such
user exists), or a percent sign for temporary failure (e.g. unable to receive message
because disk is full). The rest of a reply is a human-readable character string
explaining the situation, followed by a carriage return.

The message text transmitted through the mail protocol normally contains a header
formatted in the Arpanet standard fashion. [RFC733]

5.6 Send ¶

The SEND protocol is used to transmit an interactive message (requiring
immediate attention) between users. The sender connects to contact name
SEND at the machine to which the recipient is logged in. The remainder of
the RFC packet contains the name of the person being sent to. A stream
connection is opened and the message is transmitted, followed by an EOF.
Both sides close after following the end-of-data protocol described in End-of-Data.
The fact that the RFC was responded to affirmatively indicates that the
recipient is in fact present and accepting messages. The message text should begin
with a suitable header, naming the user that sent the message. The standard
for such headers, not currently adhered to by all hosts, is one line formatted
as in the following example:

Moon@MIT-MC 6/15/81 02:20:17

Automatic reply to the sender can be implemented by searching for the first
“@” and using the SEND protocol to the host following the “@” with the
argument preceding it.

5.7 Name ¶

The Name/Finger protocol of the Arpanet [FINGER] exists in identical form
on the Chaosnet. Both Lisp machines and timesharing machines support this
protocol and provide a display of the user(s) currently logged in to them.

The contact name is NAME which can be followed by a space and a string of
arguments like the “command line” of the Arpanet protocol. A stream
connection is established and the “finger” display is output in Lisp
Machine character set, followed by an EOF.

Lisp Machines also support the FINGER protocol, a simple-transaction version
of the NAME protocol. An RFC with contact name FINGER is transmitted and
the response is an ANS containing the following items of information separated
by carriage returns: the logged-in user ID, the location of the terminal, the
idle time in minutes or hours-colon-minutes, the user’s full name, and
the user’s group affiliation.

5.8 Time ¶

The Time protocol of the Arpanet [TIME] exists on Chaosnet as a simple
transaction. An RFC to contact name TIME evokes an ANS containing
the number of seconds since midnight Greenwich Mean Time, Jan 1, 1900
as a 32-bit number in four 8-bit bytes, least-significant byte first.
Some computers–Lisp machines, for example–which don’t have hardware
calendar-clocks use this protocol to find out the date and time when
they first come up.

5.9 Arpanet Gateway ¶

This protocol allows a Chaosnet host to access almost any service on
the Arpanet. The gateway server runs on each ITS host that is connected
to both networks. It creates an Arpanet connection and a Chaosnet connection
and forwards data bytes from one to the other. It also provides for a one-way
auxiliary connection, used for the data connection of the Arpanet File Transfer Protocol.

The RFC packet contains a contact name of ARPA, a space, the name
of the Arpanet host to be connected to, optionally followed by a space
and the contact-socket number in octal, which defaults to 1 if omitted.
The Arpanet Initial Connection Protocol is used to establish a bi-directional
8-bit connection.

If a data packet with opcode 201 (octal) is received, an Arpanet INS
signal will be transmitted. Any data bytes in this packet are transmitted normally.

If a data packet with opcode 210 (octal) is received, an auxiliary
connection on each network is opened. The first eight data bytes are
the Chaosnet contact name for the auxiliary connection; the user should
send an RFC with this name to the server. The next four data bytes are the
Arpanet socket number to be connected to, in the wrong order, most-significant
byte first. The byte-size of the auxiliary connection is 8 bits.

The normal closing of an Arpanet connection corresponds to an EOF packet.
Closing due to an error, such as Host Dead, corresponds to a CLS packet.

5.10 Host Table ¶

The HOSTAB protocol may be used to access tables of host addresses
on other networks, such as the Arpanet. Servers for this protocol
currently exist for Tenex and TOPS-20.

The user connects to contact name HOSTAB, undertakes a number of
transactions, then closes the connection. Each transaction is initiated
by the user transmitting a host name followed by a carriage return. The
server responds with information about that host, terminated with an EOF,
and is then ready for another transaction. The server’s response
consists of a number of attributes of the host. Each attribute consists
of an identifying name, a space character, the value of the attribute,
and a carriage return. Values may be strings (free of carriage returns
and not surrounded by double-quotes) or octal numbers.
Attribute names and most values are in upper case. There can be more than
one attribute with the same name; for example, a host may have more than
one name or more than one network address.

The standard attribute names defined now are as follows. Note that more
are likely to be added in the future.

ERROR

The value is an error message. The only error one might expect
to get is “no such host”.

NAME

The value is a name of the host. There may be more than one NAME
attribute; the first one is always the official name, and any additional
names are nicknames.

MACHINE-TYPE

The value is the type of machine, such as LISPM, PDP10, etc.

SYSTEM-TYPE

The value is the type of software running on the machine,
such as LISPM, ITS, etc.

ARPA

The value is an address of the host on the Arpanet, in the form
host/imp. The two numbers are decimal.

CHAOS

The value is an address of the host on Chaosnet, as an octal number.

DIAL

The value is an address of the host on Dialnet, as a telephone number.

LCS

The value is an address of the host on the LCSnet, as two octal numbers
separated by a slash.

SU

The value is an address of the host on the SUnet, in the form
net#host. The two numbers are octal.

5.11 Dover ¶

A press file may be sent to the Dover printer by connecting to contact
name DOVER at host AI-CHAOS-11. This host provides a protocol
translation service which translates from Chaosnet stream protocol to
the EFTP protocol spoken by the Dover printer. Only one file at
a time can be sent to the Dover, so an attempt to use this service may
be refused by a CLS packet containing the string "BUSY".
Once the connection has been established, the press file is transmitted
as a sequence of 8-bit bytes in data packets (opcode 200). It is necessary
to provide packets rapidly enough to keep the Dover’s program (Spruce)
from timing out; a packet every five seconds suffices. Of course, packets
are normally transmitted much more rapidly.

Once the file has been transmitted, an EOF packet must be sent.
The transmitter must wait for that EOF to be acknowledged, send a
second one, then close the connection. The two EOF’s are
necessary to provide the proper connection-closing sequence for the
EFTP protocol. Once the press file has been transmitted to the
Dover in this way and stored on the Dover’s local disk, it will be
processed and prepared for printing, and then printed.

If an error message is returned by the Dover while the press file is
being transmitted, it will be reported back through the Chaosnet as a
LOS containing the text of the error message. Such errors are
fairly common; the sender of the press file should be prepared to retry
the operation a few times.

Most programs that send press files to the Dover first wait for the
Dover to be idle, using the Foreign Protocol mechanism of Chaosnet to
check the status of the Dover. This is optional, but is courteous to
other users since it prevents printing from being held up while
additional files are sent to the Dover and queued on its local disk.

It would be possible to send to a press file to the Dover using its
EFTP protocol through the Foreign Protocol mechanism, rather than
using the AI-CHAOS-11 gateway service. This is not usually done
because EFTP, which requires a handshake for every packet, tends to
be very slow on a timesharing system.

9.1 Opening Connections ¶

A (potential) Chaosnet connection is represented by a JFN. When the JFN
is opened, an actual Chaosnet connection is created. The GTJFN syntax
is as follows:

The device name is CHA:. The filename is the symbolic name or octal
number of the host at the other end of the connection, or a null string if
it is desired to listen for an incoming Request For Connection rather than
initiating a connection. The extension (filetype) is normally the contact
name; some special cases are described below. Use of * names and JFN
stepping is not permitted. The directory, generation (version) number, and
semicolon attributes are ignored if present.

When the JFN is opened (with OPENF), normally the system will wait
for the connection to open up; a user connection (nonblank filename) will
wait for a response to be returned to its RFC, and a server connection
(blank filename) will wait for an RFC to come in to its contact name. If
an RFC is refused or the foreign host is not up, OPENF will return an error.
If data mode 6 or 7 is used with OPENF, it will return immediately,
without waiting for the connection to open. This is useful if you want to
open several Chaosnet connections simultaneously, or if you want to determine
the reason for failure if the connection does not open; if the normal
data mode 0 OPENF
fails, the operating system will not let you read the CLS packet
nor do the .MOERR operation (described below).

There are a number of special cases in the GTJFN syntax. If the extension
is a null string, then the contact name is specified by OPENF rather than
by GTJFN; AC3 contains the number of characters in the contact name
in the left half, and the address of the contact name in the right half.
In listening mode (the filename is a null string), then if the extension is
also null, this JFN will listen to any RFC that is not otherwise
serviced. Privileges are required and only one job at a time can do this.
This mechanism is used by a system server process. If the filename is null
and the extension is a hyphen, the JFN is put in a special mode for
simple transactions; packet-level I/O may be used to transmit any number of RFCs and
receive any response packets (ANS, FWD, LOS, or CLS).

When a JFN is closed with CLOSF, if it has an open connection the
end-of-data protocol is used (an EOF packet is transmitted and its
acknowledgement is awaited), and then a CLS packet is transmitted. (This
is not completely implemented yet; currently no EOF is sent, and .MOEOF (see
below) must be used.) If the JFN is in the RFC-Received state, the RFC
is refused by sending a CLS.

9.2 Stream Input and Output ¶

The normal I/O JSYSes (BIN, BOUT, SIN, SOUT) work on
Chaosnet JFNs. When the connection was created by listening, doing I/O
to it automatically accepts it first (sending an OPN). The input and
output data are transmitted as 8-bit bytes in standard data packets (opcode
200). On input, if an EOF packet is encountered the standard end-of-file
action occurs. If a CLS or LOS is encountered, or the connection is in a
bad state, an error is signalled. The message from the CLS or LOS may be
picked up with .MOERR (see below).

On TOPS-20, but not Tenex, the SIBE, SOBE, DIBE, DOBE,
SINR, and SOUTR JSYSes may be used. The latter two treat each
packet as a separate record.

The OPENF byte size may be either 7 or 8. With a byte size of 8, the
raw Chaosnet data bytes are transmitted. With a byte size of 7, the system
converts between the ASCII code it uses normally and the Lisp Machine
character set, which is standard on Chaosnet. (This is not yet implemented;
currently a byte size of 7 will be accepted but will behave the same as 8.)

9.3 Packet Input and Output ¶

It is possible to do packet-level input and output and to deal directly with the
details of the Chaosnet protocol by using the special operations described in the
following section. Note that stream I/O and packet I/O should not be mixed
in the same connection, unless you know exactly what you are doing, since you
can get your data out of order.

9.4 Special Operations ¶

GDSTS returns device-dependent status. AC2 returns the state of
the connection, and AC3 returns the number of packet slots available in the
output window in the left half, and the number of available input packets in the
right half. The symbolic names for the connection states are as follows:

.CSCLS

The connection is closed (or was never opened).

.CSLSN

The connection is listening for an RFC.

.CSRFC

An RFC has been received by a listening connection.

.CSRFS

An RFC has been sent.

.CSOPN

The connection is open.

.CSLOS

The connection has been broken by a LOS packet.

.CSINC

The connection has been broken by Incomplete Transmission (no response from the
other end for a long time).

.CSPRF

This is the “permanent RFC” state which is entered by GTJFN with
a null filename and an extension of just a hyphen.

MTOPR performs a variety of special operations, with the JFN in AC1,
one of the following function codes in AC2, and an argument and/or return
value in AC3.

.MOPKS

Send a packet. AC3 contains the address of the first word of the packet.
An error return is taken if the connection is in a bad state for the kind of packet
being transmitted. This will wait for space to be available in the window.

.MOPKR

Receive a packet. AC3 contains the address of a 126-word buffer in
which the packet is to be stored. This will wait until an input packet arrives.

.MOOPN

Accept a Request for Connection. Error if the connection is not in the
RFC-received state.

.MOSND

Force out any buffered stream output.

.MOEOF

Force out any buffered stream output, then send an EOF packet.

.MONOP

Force out any buffered stream output, then wait for it to be transmitted and
acknowledged. (This is not a “no op”, but .MONOP is the system standard
name for this operation.)

.MOERR

Returns the error message from a received CLS or LOS packet.
An error is signalled if no error message is available. AC3
is a string pointer to where to put the error message; it is updated
to point at the terminating null character which makes the message an
ASCIZ string.

.MOACN

Assigns PSI (interrupt) channels. The left half of AC3 is
the channel number for output interrupts, and the right half is the channel number
for input and state-change interrupts. Specifying -1 as a channel number
disables interrupts. Output interrupts are signalled when the window is
full and then an acknowledgement is received which makes some space so that
more packets may be output. Input interrupts are signalled when
the state changes, and when there are no input packets available
and then a packet is received.

.MOSWS

Sets the receive window size from AC3.

.MORWS

Returns the receive window size in the left half of AC3
and the transmit window size in the right half.

.MOAWS

Returns the available space in the transmit window in AC3.

.MOUAC

Returns the number of unacknowledged output packets in AC3.

.MOFHS

Returns the foreign host number in AC3.

.MOSIZ

Returns the maximum packet size in bytes in AC3.
This can be smaller than the Chaosnet standard (488) on machines
encumbered with an RSX20F front end.

.MOSRT

Sets the RFC timeout period in milliseconds from AC3.
The maximum is 262 seconds.

9.5 Utility Programs ¶

There are two Chaosnet utility programs, both named CHASTA.
One prints one line for each connection that exists, giving its
state, number of input and output packets, who it is connected to, etc.
The other prints the STATUS protocol information for every host
on the network, including the host name, when it was last up, and its
packet throughput and error counts. This information is maintained by
a system daemon process.

9.6 Server Programs ¶

When an RFC is received for contact-name, if no process is listening
for contact-name and the file SYSTEM:CHAOS.contact-name (or
DSK:CHAOS.contact-name on Tenex) exists, the server
program contained in that file is run. The server program should open
CHA:.contact-name. This is implemented by the CHARFC
program which runs as a daemon job and opens CHA:., the magic name
which gets a copy of all unclaimed RFCs. Normally the server program is
run in a freshly-created job, and may log in if it wishes, but if the
file is marked as ephemeral (the “;E” attribute), it is run in a subfork
of the CHARFC job. Ephemeral servers should be used for protocols
that don’t involve a long-term connection.

The TELNET and SUPDUP servers attach their Chaosnet connection
directly to an NVT, just as the corresponding Arpanet servers do.

When the system starts up, the file SYSTEM:HOSTS2.BIN
(or DSK:HOSTS2.BIN on Tenex) is read in and used to initialize
the host name table inside the system used by GTJFN. This is the
ITS/TOPS-20/WAITS standard multi-network host table.

12.2 Special Files for Creating Connections ¶

There are several special files in the file system that provide
ways of creating and accessing connections (these names are defined in ):

CHRFCDEV

Opening this file creates a connection to a remote host.
The host address should follow the file name as an additional
pathname component containing ascii digits representing the
Chaosnet address in decimal (soon to be octal). The rest of
the pathname after the host address is taken as the contact
name that will be sent in the RFC packet. Thus to open
a TELNET connection to the host at address 234 use:

#include 
#include 
char pathbuf[100];
int fd;
sprintf(pathbuf,
        "%s/%d/%s", CHRFCDEV, 234, CHAOS_TELNET);
fd = open(pathbuf, 2);

To send a message to User at host address 567 use:

sprintf(pathbuf,
        "%s/%d/%s %s", CHRFCDEV, 567, CHAOS_SEND, User);
fd = open(pathbuf, 1);

Opening CHRFCDEV returns when the response to the RFC is
received from the remote host, or a fixed timeout, whichever
happens first. Other timeouts may be implemented by the user program,
using the alarm system call. ANS packets are acceptable
responses. The data in the ANS packet will be readable, and will be followed
by end-of-file as with a full connection or a normal file.

CHRFCADEV

This device provides the same functions as CHRFCDEV except that it
returns immediately after transmission of the RFC packet with the connection
in the CSRFCSENT state. This allows the user
program to have access to the contents of packets refusing the
connection (CLS, LOS). See the CHIOCSWAIT and CHIOCPREAD
ioctl’s below.

CHLISTDEV

Opening this file creates a connection in the Listening state
with the contact name given as the pathname component following
the device name. Thus to listen for a TELNET connection use:

sprintf(pathbuf, "%s/%s", CHLISTDEV, CHAOS_TELNET);
fd = open(pathbuf, 2);

Use the CHIOCSWAIT ioctl to wait for a RFC to arrive, and
CHIOCREJECT, CHIOCACCEPT, or CHIOCANSWER to respond.

CHURFCDEV

This file, the “unmatched RFC server” device, when opened and read will return
the contents of RFC packets
that have no listener. Read calls on this connection just return RFC data.
If another read on this file is done before the RFC is matched, it is
discarded. This file may only be opened by one user at a time. Normally
this file is opened by the system unmatched-RFC server process.

12.3 Stream Mode Connections ¶

The default mode when a Chaosnet device is opened is stream mode.
This mode makes the connection behave like a UNIX
file, with the exception that seek system calls are disallowed and
read calls will return whatever data is available if there
is any. (Rather than returning the full number of bytes requested.)
Thus standard I/O library routines can easily be used
to read and write on these connections. A normal UNIX end of file
indication will be returned on receiving an EOF packet,
and will continue to be returned until either the connection
is closed or more data arrives on the connection. If the
connection is closed before an EOF packet is received
(via a CLS or LOS packet arriving or a connection timeout occurring), an error
will be returned after all data and EOF packets have been read.

If the file was opened for writing (open mode 1 or 2), then
an EOF packet will be sent and its acknowledgement awaited
when the file is closed (unless the connection was already
closed).

In stream mode all non-data packets are discarded and data
packet opcodes are all treated the same. Ioctl’s can be used to
read non-data packets (RFC, CLS, LOS etc.) and perform other network
specific functions.

The contents of ANS and UNC packets are read as data in the stream,
just as if they had been data packets.

12.4 Record Mode Connections ¶

This mode, set on a connection by issuing

ioctl(fd, CHIOCSMODE, CHRECORD);

gives the user program access to packet opcodes and packet boundaries, without any
further awareness of network data structures. Read calls from the connection
return all the data in a single packet, with the first byte of the data
being the opcode in the packet. The count returned is, as normal, the
count of bytes transfered, including the opcode. Opcodes are defined in
.

RFC, ANS, CLS, LOS, EOF, UNC, FWD, and data packets will be returned to the
user. The buffer given must be large enough to fit the entire packet including the
opcode byte, or an error is returned.

Write calls must include the desired opcode as the first byte of data and
also reflect this byte in the byte count. The data to be written must not
exceed the maximum packet size.

If a record mode connection is closed in the OPEN state a CLS packet
will automatically be sent. The CHIOCREJECT ioctl should be used to send
a CLS packet containing a specific reason.

12.5 TTY Mode Connections ¶

TTY mode (via CHTTY) connections allow the connection to act exactly as a UNIX tty,
allowing, for example, remote login service with no extra process for the
NVT. Unfortunately, none of the remote protocols (TELNET, SUPDUP) can
work over a transparent connection that just acts like a terminal.
This mode is currently unused.

12.6 Foreign Protocol Mode ¶

Used for a connection to transmit and receive foreign protocols encapsulated
in UNC packets. (unimplemented)

12.7 IOCTL System Call Commands ¶

The following ioctl codes can be used on Chaosnet connections.

CHIOCSMODE

Set the connection mode. Argument is
CHSTREAM, CHRECORD, CHTTY, or CHFOREIGN.

CHIOCSWAIT

Wait until the connection state changes from
the given state (in the third argument). Typically used
for listeners waiting for a RFC:

ioctl(fd, CHIOCSWAIT, CSLISTEN);

or for a user end waiting for a respone to a RFC:

ioctl(fd, CHIOCSWAIT, CSRFCSENT);

CHIOCFLUSH

In stream mode, send out any data waiting for a full packet.
This is done every 1/2 second at clock level anyway.

CHIOCOWAIT

Wait for all transmitted data (after doing a CHIOCFLUSH) to
be acknowledged by the other end of the connection. If the
argument is non-zero, an EOF packet is sent first, and then it
must be acknowledged also.

CHIOCGSTAT

Get the status of the connection. The argument is the address
where the status structure (struct chst in ) will be returned.
This is frequently used to ascertain the state of the
connection after a CHIOCSWAIT call, or to find out the Chaosnet address
of the other end.

CHIOCANSWER

When a connection is in the CSRFCRCVD state, this causes
data writes on the connection to be sent using an ANS packet.
In stream mode, the packet is filled incrementally. In record
mode the first packet sent is made into an ANS. In every case
only one packet is sent and the connection is made closed.

CHIOCACCEPT

When a connection is in the CSRFCRCVD state, this causes an
OPEN packet to be sent and the connection opened.

CHIOCREJECT

When the connection is in either the CSRFCRCVD or CSOPEN state
this causes a CLS packet to be sent, closing the connection.
The argument is the address of a null-terminated string which
is copied into the close packet.

CHIOCPREAD

Read a packet from the received packet queue. Used in stream mode
to read control packets that are otherwise ignored. Typically
used to read RFC or CLS packets.

CHIOCRSKIP

Skip over the unmatched RFC at the head of the unmatched RFC
queue and mark it to only be matched against a listen, not
queued as an unmatched RFC. This is used by the unmatched
RFC server to ignore RFC’s it knows someone else might want.

FIONREAD

A normal UNIX ioctl, this returns an integer at the address
specified by the argument, containing the number of byes available to
be read.

12.8 Signals ¶

All read, write, open, close, and ioctl’s are interruptable except when
waiting for buffer allocation. Read and write calls are automatically
restarted (on VAX UNIX). All others currently return EINTR errors.

12.9 Software Installation ¶

Global header files are placed in a “chaos” subdirectory of the system header
file directory (usually /usr/include) so that #include works.

The kernel code is found in two subdirectories of the kernel source
directory, parallel to sys, dev, and conf:

chncp

This directory contains the parts of the NCP which do not depend
on the operating system. It also contains the actual Chaosnet
interface drivers – which do have some operating-system-dependent
code. These drivers interface only to the Chaosnet code (except interrupt
vectors) and thus are not usable as UNIX device drivers (this may change).

chunix

This directory contains the top level device driver interface from the
system call level (through cdevsw) to the NCP and some system
dependent utilities (buffer allocation etc.)

The NCP needs two entries in the character device switch and one other
small change in conf.c.

In the UNIX kernel proper several small changes are required:

nami.c

A small (2 line) change is required to nami to allow Chaosnet special
files to have additional pathname components after the one that matches the
special file in the file system.

clock.c

A call to the Chaosnet clock process routine must be placed in the
clock routine called every clock tick (this should be done with timeouts
but isn’t).

fio.c

Several bugs in closef which never were encountered by
other drivers need fixing.

All these changes are conditioned on #ifdef CHAOS. In the normal
(Berkeley) VAX configuration scheme, normal entries made in the
configuration file suffice to cause all the right files to be included
in the system if the line

is specified and the CHAOS option is included in the “options” line.
The “files” file gets a few more lines.

For pdp11 UNIX, since the configuration system is much more primitive, some
handwork is usually required to make the kernel correctly.

All user program sources can be put in /usr/src/cmd/chaos,
/usr/local/src/cmd/chaos, /usr/src/local/cmd/chaos.
The make file contains
variables for destinations of all programs. The default destination for user
programs is /usr/local.
Server programs are placed in /usr/local/lib/chaos.
The unmatched RFC server (“chserver“) is placed in /etc and should be
started in the /etc/rc file at boot time. It may be killed and restarted at any
time.