__FILES__________________________CXR__________________________________

5. FILES

The files are used to define the operation and administration of the device.
They contain: the initialization of the variables (cf. chapter VARIABLES), the
operating options for the modes and protocols and the
device's configuration commands. The lines contained in the files cannot
exceed 80 characters.

The administration of the actual files is accessible using the ls command,
given that these files are stored in the configuration memory (E2 prom or Flash 
depending on the model).
The user can create certain files that are not included in the
standard shipment using the eestore command.
SEE: command eestore.
The creation of files starting with the '_' character is forbidden.
There are a number of files whose name starts with '_' that cannot
be modified and that will be used automatically when
files required for the device's minimal operation are destroyed
(e.g.: complete reset after losing the password).
If the name of file starts with "__", it is a file "by default";
a file "by default" is a file whose contents, in most cases,
do not need to be modified, but which can, exceptionally,
contain other data. To be able to replace the original data, you
will copy the contents of "__xxxx" in "xxxx" (command "cp"). As for the "xxxx" file,
it can be edited as with any file contained in the configuration
memory.
SEE: command cp.
The modification of some files can be taken into account
immediately (e.g.: file "secret"), as opposed to others which require you
to reboot the device (e.g.: admin) or use commands with
certain options (e.g.: "orion -r" for the "fip" file). The name of these
files is followed by the mention "(dynamic)" in the manual.

A file with the suffix ".gz" (e.g.: pad.prf.gz) represents a file
compressed using "gzip". To decompress its contents, simply
request the file without the ".gz" extension (e.g.: cat pad.prf); the
file system will then decompress this file automatically
for reading.
SEE: decompression copyright.


[access]

contains the authorizations or inhibitions for using different
local TCP/UDP/ICMP services offered by the device; this makes it possible
to filter the device's incoming and outgoing communications.
This file is read when the system is reset; after any modification,
you must reboot the device in order for the modification
to be taken into account.
The file's structure is as follows:
	command:
	adr_machine
	adr_machine
	command:
	!adr_machine
	!adr_machine
where
command      is a command name using one of the TCP, UDP or
   ICMP services such as telnet, telnetd, tftp, tftpd, printerd, time, route, ping. The
   ':' character that follows the name of the command is imperative.
adr_machine   is the address of a machine either in numeric format or with
   a symbolic name if this name is defined in the hosts file. This declaration
   authorizes the processing of commands by the machine with the adr_machine. If at least
   one machine is authorized for a command, the list of all the
   authorized machines must be specified.
!adr_machine  is used to forbid the processing of commands by the machine
   with the adr_machine. Here, the user must specify the list of all
   the machines forbidden for commands.

Example:
	ping:
	123.45.67.89
	telnetd:
	!98.76.54.231
	!89.67.45.123
	tftp:
	server
	!89.67.45.123

In this case:
1. Only machine 123.45.67.89 is authorized as the destination of the
   device's ping command.
2. All the machines, other than 98.76.54.231 and 89.67.45.123, can
   connect to the device via telnet, whereby the tcpcomd daemon must be active.
3. Only the server machine can be given as an argument of the vtftp command
   on the device.
Violating the authorization for an outgoing command will initiate the display
of the message "machine unknown", while an incoming connection will be unsuccessful.
Thanks to the ADDRTEST variable, this file can also be used to simply
filter routed frames.
SEE: command access, variable ADDRTEST.


[admin]

contains the system's different variables and those of the operation. This
file is read when the system is reset; it serves to associate
values with the variables defined (cf. chapter VARIABLES). In this file, the
variables are defined as follows: variable = value with no space or separation tabulation
on either side of the = sign, nor in the value field.

	INITTAB=inittab

associates the inittab value with the INITTAB variable.


[aliases]

contains the definition of the aliases used by the shell. These declarations
are used to define command synonyms possibly accompanied
by parameters. The syntax of a line in this file is:
alias command [parameters]
Example of the contents of the alias file:

	clear cls
	c connect 3615
	d disconnect

ATTENTION: these definitions are only accessible in an interactive shell
session, but not in the files (inittab, crontab, etc.).
SEE: shell commands.




[asini.i]

(dynamic)
This file contains the orders to be executed to initialize an external
modem. The 'i' placed as a suffix after the filename represents the relevant channel.
In this case, a modem is connected on the "serial" link and is used
to transfer data (see file "numip" and "Ci_Annn_script"
in the "num" part). When the device is reset, this
file is executed like a script. The character string(s)
that is/are to be sent to the modem will be sent using the "send" command and,
if waiting for an answer is necessary, the "wait" command will be used.
This file's use has been extended, in a
general manner, to commands that use the synchronous port such as asyncd,
paf, etc.
For greater flexibility, you can use the concept of a 
section inside this file.
The section used for the channel's first initialization is represented
by a line starting with the sequence "[initialization]".
The section used at the closure of the channel is represented by 
a line starting with the sequence "[closure]".
When the line is free, the section used when the CTS signal 
changes from low to high is represented by a line starting with the
sequence "[cts/]".

Everything included between the start of the file and the description
of the first section will be executed when the channel is initialized
and closed.
If the "asini.1" file (for example) does not exist, the system will search
for an "asini" file for the same purpose. This "asini" file can be used
for all the channels available in the case where there is no "asini.xx"
(where xx is the channel number) file in the configuration.

SEE: file numip, command send, command wait.


[backup]

This is a highly specific file. In fact, it is not just a file,
but a group of files: the file with this name contains
all the configuration's files. It cannot be edited since
it contains non-ASCII characters, however it can be used to back up or
restore the entire configuration in a single transfer.
Some utilities (see sources in C language available from
your distributor) can be used to extract files contained in
"backup", then reassemble them to obtain a "backup" file again.
Note that when loading this file, all the previously recorded data
in the device's configuration will be deleted; the
data contained in the "backup" file will then be installed.
To back up or reload a configuration, use the
vtftp and ftp commands or the vtftpd and ftpd daemons.
Note that all the files must be closed before retrieving
a global configuration. If a file is open, you will get an error (access
rights) when writing to the backed up memory.
NB: do not use the "backup" file of one model in the range
to quickly configure another model. Some values can be
very different and the mismatch with the model can be fatal to
proper operation.	However, this operation is possible
provided you immediately remodify the important values.

SEE: command vtftp, command ftp, command vtftpd, command ftpd.


[bootptab]

This file contains the configuration of the bootps daemon. In it you will find the parameters
contained in the replies, the static addresses and the range of
dynamic addresses.
SEE: command bootps.


[crontabs]

contains the commands to be run off-line by the cron daemon. The syntax of
each line is:
minute hour day_of_the_month month_of_the_year day_of_the_week command
This syntax is exactly the same as that of the Unix crontabs/* files.
SEE: command cron.


[fip]

contains the description of the filter used by the "orion" command and
certain operating options.
The recognized operating options are:

  r_delay size[,time]
	- where "size" indicates (in Kb, value 4 by default) after what
	amount of bytes sent, for the delayed frames (see
	below), you must ask the system to slow down the exchange.
	To delay the exchanges concerned, you usually use an
	ICMP/Source_quench frame (slowdown request) sent
	to the sender. Some machines may take a while
	to react and make this effect inoperative; in this case, instead
	of sending this ICMP information, you can remove the last frame received,
	which provokes an absence of acknowledgement and forces the sender
	- after a few moments – to resend the frame (all data
	managed by TCP). This delay makes	it possible to slow down the exchange. This mode
	is selected by assigning a negative value to "size".
	- where "time" indicates (in seconds, value 1 by default) after how
	much time the device should consider that it must
	no longer slow down the exchanges, since only frames considered as "slowed down"
	travel on the link.
	When this function is used, the size of the frames
	is accounted for.
	If the number of bytes of "slowed down" frames exceeds by "size" Kb
	the number of bytes of "normal" frames, and "delay" seconds
	has not expired, a slowdown procedure is activated.
	If only slowed down frames are travelling, no such procedure is
	activated.
see frame slowdown

  log
	indicates that you wish to create a table of the
	device's use by the machines on the LAN.
	For each machine on the network, to each
	IP frame is added the count of bytes sent in a table indexed
	by the last part of the local machine's address.
	The rejected frames are not accounted for.
	Unless they are filtered, the frames being sent
	are also accounted for.
	This table can be displayed and reinitialized using
	the options in the "orion" command.
	This table will only be created if at least one filtering line
	is present.
	The size of this table is defined by the natural mask of
	the device's address on the LAN (il0).

The filter's description is given in the form of orders. In the first order
satisfied (all the conditions described met), the appropriate action
is executed and the next frame is presented to the filter.
Each order is delimited by an opening brace (start)
and a closing brace (end). The description lines must not
exceed 80 characters; an order can contain more than one
line. If a line contains the '#' character, the portion contained between
`#' and the end of the line will be ignored (comment).
Attention: if at least one filtering line is valid, the last line
of the filter will implicitly be:
	{to=RE,S}

 The following keywords are recognized:

 "from="	followed by "lan" or "wan" to specify whether this order
		applies to a frame from the local network (lan)
		or from the remote network (wan).
		The absence of this information indicates that this order
		applies, whatever the frame's source.
 "from="	"from="

 "tos="		through abusive simplification, the whole of the field containing
		the "Type Of Service" is compared to the value provided
		after the '=' sign. This field normally contains:
		- 3bits specifying the size of the TOS
		- 4bits specifying the value of the TOS
		- 1bit which must be 0 in normal operation.
		The absence of this information indicates that this order
		applies, whatever this frame's TOS value.

 "@src"	represents the frame's source address.
		It may befollowed by the '|' character which indicates that
		you must perform a "logical or" between the source address
		and the value that follows the '|' character.
		It may be followed by the `&' character which indicates that
		you must perform a "logical and" between the source address
		and the value that follows the `&' character.
		If a "logical or" and a "logical and" are
		requested at the same time, the "logical or" will be performed first, then
		the "logical and" will be performed on the result of
		the previous operation.
		Mathematically, the operation applied will be as follows:
		(@src | value_or) & value_and.
		If neither the "logical or", nor the "logical and" are requested,
		the operation will be:
		(@src | 0) & 0xFFFFFFFF, which is the same as @src.
		If only the "logical and" is not requested, the operation will be:
		(@src | value_or) & 0xFFFFFFFF, which is the same as
		a @src | value_or.
		If only the "logical or" is not requested, the operation will be:
		(@src | 0) & value_and, which is the same as @src & value_and.
		It must be followed by one of the character strings
		below and a value:
		"=="	indicating strictly equal to between the expression
			calculated and the value given,
		">"	indicating strictly greater than between the expression
			calculated and the value given,
		">="	indicating greater than or equal to between
			the expression calculated and the value given,
		"
<"	indicating strictly less than between the expression
			calculated and the value given,
		"<="	indicating less than or equal to between
			the expression calculated and the value given.
		Outside the strictly equal to condition, you can
		describe both a less than condition and a
		greater than condition.
		The value_or, the value_and and the resulting value can be written
		in "dot notation" (e.g.: 255.255.255.0), in
		hexadecimal (preceded by the "0x" string, e.g.: 0xFFFFFF00)
		or using the following keywords:
		- @iladdr (IP address of the Ethernet interface),
		- @ilmask (IP mask of the Ethernet interface),
		- @ilnet (network address of the Ethernet interface).
		The absence of this information indicates that this order
		applies, whatever the frame's source address.

 "@dst"	idem "@src", except that this information concerns the frame's
		target address.

 "prot="	followed by a protocol to specify that this order
		applies to a protocol frame strictly equal to the
		value given.
 "protm="	followed by a protocol to specify that this order
		applies to a protocol frame greater than or equal
		to the value given (minimum protocol).
 "protM="	followed by a protocol to specify that this order
		applies to a protocol frame less than or equal
		to the value given (maximum protocol).

		Note that the use of the "prot=" keyword forbids the use of the
		2 keywords "protm=" and "protM=", which will now have no object.
		If only "protm=" is defined, the implicit value of
		"protM=" will be 0xFFFF.
		If only "protM=" is defined, the implicit value of
		"protm=" will be 0.
		The absence of this information indicates that this order
		applies, whatever the frame's protocol.

 "psrc="	followed by a port (logically linked to the protocol defined) to
		specify that this order applies to a frame with
		a source port strictly equal to the value given.
 "psrcm="	followed by a port to specify that this order
		applies to a frame with a source port greater than or
		equal to the value given (minimum source port).
 "psrcM="	followed by a port to specify that this order
		applies to a frame with a source port less than or
		equal to the value given (maximum source port).

		Note that the use of the "psrc=" keyword forbids the use of the
		2 keywords "psrcm=" and "psrcM=", which will now have no object.
		If only "psrcm=" is defined, the implicit value of
		"psrcM=" will be 0xFFFF.
		If only "psrcM=" is defined, the implicit value of
		"psrcm=" will be 0.
		The absence of this information indicates that this order
		applies, whatever the frame's source port.

 "pdst="	idem "psrc=", except that this information concerns the frame's
		target port.
 "pdstm="	idem "psrcm=", except that this information concerns the frame's
		target port.
 "pdstM="	idem "psrcM=", except that this information concerns the frame's
		target port.

 "bits&"	followed by a value (hexadecimal if starting with the
		sequence "0x", octal if starting with "0" and decimal
		otherwise) which validates the entry if, for
		a logical AND with the contents of the FIPBITS variable,
		you obtain a result other than 0. This makes it possible, without
		reloading the filter and according to the outside conditions
		(e.g.: time), to validate or invalidate the entry.
see variable FIPBITS

 "to="	this keyword is not optional. It defines how
		to route the frame. For this purpose, it must be followed
		by one of the character strings below:
		"CB"	for a routing via a B channel,
		"CD"	for a routing via a virtual circuit (VC)
			on the D channel,
		"LL"	for a routing via the synchronous serial line.
			Depending on how the serial line is configured,
			the frame will either be routed in a VC if it is
			an X25 link, or it will be encapsulated in the protocol
			defined on the link (e.g.: PPP),
		"RE"	to reject this packet,
		"lan"	for a routing to the local network,
		"auto"	for a routing according to the routing table, with
			no constraints.

		This keyword can be followed by different details given
		in the form of character strings.
		- the ",S" string to indicate that, if this packet cannot be
		routed, it will be silently rejected, i.e., without sending
		any ICMP frames indicating the rejection to the sender.
		- the ",L" string to indicate that this packet must be
		presented to syslog to inform the administrator of its
		passage. Note that you are strongly advised against using
		this option for all the packets, since the workload of both the device
		and of the log machine would become needlessly high…
		- the ",R" string to slow down the frames described, which
		gives them a lesser priority.
see r_delay
If at least one order is valid, the last line will always implicitly be: "to=RE,S", that is to say that the non-described frames will be rejected. If the last order line entered by the user is "to=auto", the implicit order "to=RE,S" will have no effect. "stat=" this keyword defines a condition on the routing link indicated by "to=". The following conditions are authorized: "C" to only send the frame described if the routing link is already connected, "X" to send the frame described whatever the status of the routing link. The default value is "X", that is to say that the frames will be sent whatever the link's status. This information will have no object in the case where the value of "to=" is "RE". Example 1 of file "fip": #order 1 {from=lan prot=icmp to=CD} #order 2 {from=lan prot=tcp psrc=telnet to=CD} #order 3 {from=lan prot=tcp pdst=telnet to=CD} #order 4 {from=lan to=CB} #order 5 {from=wan @src & 0xFFFF0000 == 172.21.0.0 to=RE} #order 6 {to=auto} - Order 1 directs the ICMP frames from the LAN to ChannelD if the final destination can be reached by this support. If the destination cannot be reached by this support, an ICMP "machine unreachable" frame will be returned to the sender. - Orders 2 and 3 impose the passage of the TCP/telnet frames from the LAN by ChannelD, as with order 1, if the destination cannot be reached, the sender will be informed. - Order 4 imposes the previously non-described frames from the LAN to use ChannelB. - Order 5 rejects the frames from a remote network if this frame's source IP address starts with 172.21 (wwxxyyzz & 0xFFFF0000 = wwxx0000). - Order 6 indicates that the other frames must be routed according to the routing table, with no particular constraints. Example 2 of file "fip": #order 1 {from=lan @src & @ilmask == @ilnet to=auto} #order 2 {from=wan to=auto} - Order 1 authorizes the frames from the LAN and belonging to the network of the router's Ethernet interface to travel normally. - Order 2 authorizes the frames from the WAN to travel normally. - Implicitly, the other frames will be rejected silently. This filter is useful, for example, when the same physical network supports several logical networks (different network IP addresses) and when you do not wish to route those frames from the local network that do not belong to the local network of the device's Ethernet interface. Example 3 of file "fip": #order 1 {to=auto} #order 2 log - Order 1 authorizes the transit of frames with no constraints. - Order 2 is used to maintain a table that can be consulted using the "orion -L" command, containing the use the machines on the local network make of the router by accounting for the number of bytes sent by the various machines. This filter is an "all-pass" filter designed merely to enable the recording of the use the machines on the network make of the device.
SEE: command orion, variable FIPBITS.



[hosts]

(dynamic)
contains the correspondence between the symbolic names of the machines and their
IP address in the form of lines with 2 or 3 fields:

adr_mach name_mach [alias_mach]

adr_mach      address of the machine (in Internet 4-byte decimal format
		separated by the character '.').
name_mach      official name of the name.domain machine in the form of
		uninterrupted character strings separated by '.'.
		If this name takes the form "host.*", any request where "host"
		corresponds will be served with the specified address, regardless
		of whether or not the request contains the domain name.
alias_mach    alias of the official name of the machine (in the form of an uninterrupted
		character string).

For "name_mach" and "alias_mach", the following characters are authorized:
- letters (upper/lowercase)
- digits
- the characters '_' and '-'
The '.' character represents the separator between the name's different
parts.
Some names are reserved:
- localhost	is always 127.0.0.1
- mx.domain	see server "named"

The separation between the fields is either a space or the
tabulation character. The syntax is compatible with the file /etc/hosts on
Unix systems.
	127.0.0.1 localhost
	193.56.122.1 mymach.myplace.fr

Commands using this file: telnet, ping, netstat, etc.
To limit the number of entries present in this file, you can
use a Domain Name Server (DNS).
SEE: variable DNS, daemon named.


[codes]

This file contains the telephone codes used for channel
overflows (see variable OVERFLOW). Indeed, the overflow negotiation procedure
makes it possible to obtain the number to be called to establish the 
overflow circuit. This number is provided with no code by the remote site; if a
code is required, it will be extracted from the number of the
primary channel by comparison with the codes provided in the
codes file.
Each line in this file specifies a code with a maximum seven digits.
Example: The codes file contains
	0041
	000949
If one of the B channels is connected to the number 000949113615 and if, for the
overflow attempt, the remote site provides the secondary channel number
113616, an overflow connection will be attempted with the
number 000949113616 on the other free B channel.
SEE: variable OVERFLOW.


[inittab]

contains the different command lines to be executed when
using the system. You can give another file name
with the INITTAB variable.
SEE: command script.


[ipconf]

contains options for configuring the IP module.
You are strongly advised, except for specific applications,
to use this possibility; the default configuration is perfectly
adapted to normal device use.
The following parameters are modifiable:

keyword		Unit	Default	Comment
tcp_keep_period	second	  45	period of TCP Keep Alive
tcp_keep_retry		times	   8	#Keep Alives with no answer accepted
tcp_keep_mode		mode	   1	TCP keep alive model **

**
mode 1:	ack and seq are at -1 in relation to the last exchange
mode 2: only seq is at -1 in relation to the last exchange


[ipx.conf]

contains the configuration options of the IPX routing module when
the device works in IPX router mode (variable IPXNET(0). Each line only
contains one option followed by its parameters. These options are split
into several categories:

1/ option [num]

num is an integer defining a period in seconds. The default value of
num is 0.

saplan num
num represents the life duration of SAP information memorized in
the device. With no SAP announcement of a certain service for num seconds, the
service is announced to be unreachable to the remote devices connected, then is no longer
announced. The recommended value is CyberLS2 (2 minutes).
SAP service example: print service, remote console, etc.

sapwan num
like saplan but concerns the remote services. The recommended values are
0 for a switched connection and 240 for a permanent connection (leased line).

sapann num
where num is an integer defining the message period of the SAP services: to
the remote networks for the services detected locally, to the local network
for services announced by the remote router. The recommended value is
60.

riplan num
like saplan, but for RIP information (announces routings to the
internal and external networks). The recommended value is CyberLS2.

ripwan num
like sapwan but for RIP information. The recommended values are 0
for a switched connection and 240 for a permanent connection.

ripann num
like sapann but for RIP information. The recommended value is 60.

secann num
where num defines the period for sending IPX Serialization frames for the
remote services that are not on-line (remote device not connected). These frames
enable the servers to check that they are using software with a
unique licence number. The recommended value is 0 (no message). In
all cases, when the communication is established with a remote site, these
frames are sent.

triggered num
where num is an integer defining the message time for
modifications of SAP services or RIP routings to update remote sites
that are not connected. This mode, called triggered mode, is active if num is non-null.
The message for the modification of an SAP service or a local RIP routing is
deferred by num seconds before establishing the connection with the remote devices
to announce this modification to them. In the interests of operating costs,
we advise against using this function; the recommended value
is therefore 0.

2/ option [num]

num is an integer defining a wait in 1/50s of a second (20ms).
srwait num
where num is an integer defining the delay before the device answers
an SAP or RIP request. This mechanism can be used when a server's
answer delay is very long and where this server did not
announce itself beforehand. The recommended value is 0 (no declaration); this
mechanism is rarely used.

3/ option

defines an operating mode (no parameters).

track
validation of a "track" function similar to that of IPX servers which
display the receipt and transmission of the frames from services (SAP and RIP) via 
syslog.

keepipx
validation of the local reply to the IPX watchdogs when the connection is not
open to the site concerned. You are advised to use it.

keepspx
validation of the local reply to the SPX watchdogs when the connection is not
open to the site concerned.

802.2
802.3
ethII
snap
selection of the type of IPX frame on the LAN. By default, the device
listens and sends frames in 802.3 raw (keyword 802.3). If the network
is configured in another mode, you will have to specifiy which one,
by adding the keyword in the configuration file. Note that
only one mode can be selected on the LAN. This mode plays no role
in the transmission on remote networks; it is perfectly
possible for two sites to communicate where one is in 802.3 raw and
the other in Ethernet II (keyword ethII). The keywords have the following
correspondence:
	802.2	corresponds to IEEE 802.3 + 802.2
	802.3	corresponds to IEEE 802.3 raw
	ethII	corresponds to Ethernet II
	snap	corresponds to IEEE 802.3 + 802.2 + snap

master
validation of the master mode when using a number of same-type devices
on the LAN. Master mode is essential when the
site has several such devices. Once of the devices
is in master mode, while the others are in slave mode. The master
device is the one visible by the local IPX servers. It broadcasts
the routings for all the remote sites. The master controls the slave
devices via the LAN by assuming the role of local router
for the slave devices. The slaves do not announce their routes
in broadcast mode, only to the master.

slave
validation of the slave mode when using a number of same-type devices
on the LAN. All the frames exchanged on the LAN
will be from or to the master device.

nobroad
validation of the transmission of frames broadcast on the LAN. This
function is useful to forbid the broadcast of SAP and
RIP messages. These messages are announced to the servers on the restricted list
defined.

hitch
used to set the source network address in the reply frame to a
Get Nearest Server SAP. This function is required for
interoperability with certain ISDN PC cards whose external address
cannot be set. The source network address will be that which corresponds to the number
in communication on this interface.
SEE: file numipx.

4/ option adr_ipx

* followed by an IPX address, used for compatibility with other equipment.

wanaddr adr
defines the source address of the service frames sent to the remote
networks. This declaration defines an IPX WAN required by certain
LAN-WAN routers of third party origin. If this parameter is not defined, the
IPX address of the WAN interface will be the IPX address of the LAN. If only part
of the address is provided (network or node), the rest will be extracted
from the address on the LAN.
* followed by an IPX address, used to filter the service frames
  only.

gateway adr
declares the machine with the adr address as the target for the
SAP and RIP service messages from remote sites. There will be a gateway entry for each
target. This declaration can be used with the nobroad declaration
and the slave devices that ignore the messages in broadcast mode.

ipxlisten adr intf xxx
listens to the machine with the adr address for the service requests
sent to the device itself. By default, the device accepts the
service requests sent to it by all the machines.
If at least one ipxlisten declaration is used, the service requests sent
to it are all ignored except for those from machines declared
as ipxlisten.
"intf" is a separation keyword and "xxx" is "lan", "wan" or "all" depending
on whether you want to apply this restriction to the local network,
to the remote network, or to both.

ipxdonotlisten adr intf xxx
listens to all the machines whose address is other than adr for
service requests to the device itself.
"intf" is a separation keyword and "xxx" is "lan", "wan" or "all" depending
on whether you want to apply this restriction to the local network,
to the remote network, or to both.

ipxsendto adr intf xxx
only sends service requests to the machine with the adr
address. These service frames are those whose device is the source and not
coming from remote machines, i.e., resent.
"intf" is a separation keyword and "xxx" is "lan", "wan" or "all" depending
on whether you want to apply this restriction to the local network,
to the remote network, or to both.

ipxdonotsendto adr intf xxx
as opposed to ipxsendto, is used to exclude the adr target
("intf" and "xxx" have the same meaning as for "ipxlisten").

ripannounce adr intf xxx
only announces the network with the adr address, if known
("intf" and "xxx" have the same meaning as for "ipxlisten").

ripdonotannounce adr intf xxx
announces all the known networks except that with the adr address
("intf" and "xxx" have the same meaning as for "ipxlisten").

riprecord adr intf xxx
only records the network with the adr address in the table of reachable networks
("intf" and "xxx" have the same meaning as for "ipxlisten").

ripdonotrecord adr intf xxx
records all the networks except that with the adr address
("intf" and "xxx" have the same meaning as for "ipxlisten").

ripage adr intf xxx
only "ages" the network with the adr address (see riplan and ripwan), once
it is recorded
("intf" and "xxx" have the same meaning as for "ipxlisten").

ripdonotage adr intf xxx
"ages" all the recorded networks except that with the adr address, if it is
 recorded (see riplan and ripwan)
("intf" and "xxx" have the same meaning as for "ipxlisten").

5/ option adr_ipx intf xxx [svc]

followed by an IPX address, the "intf" keyword, the location of the
information (xxx=lan or xxx=wan or xxx=all, as in ipxlisten) and
possibly a service number,
used to filter on the service frames only.

sapannounce adr intf xxx [svc]
the same as ripannounce, but for the service announcement with the possibility of
limiting it to one service.

sapdonotannounce adr intf xxx [svc]
the same as ripdonotannounce, but for the service announcement with the possibility of
limiting it to one service.

saprecord adr intf xxx [svc]
the same as riprecord, but for the service announcement with the possibility of
limiting it to one service.

sapdonotrecord adr intf xxx [svc]
the same as ripdonotrecord, but for the service announcement with the possibility of
limiting it to one service.

sapage adr intf xxx [svc]
the same as ripage, but for the service announcement with the possibility of
limiting it to one service (see also saplan and sapwan).

sapdonotage adr intf xxx [svc]
the same as ripdonotage, but for the service announcement with the possibility of
limiting it to one service (see also saplan and sapwan).

6/ option [num]
num is an integer whose contents depend on the parameter used.

pobro num
where num is a port number to which packets, whose "node" part
is 0xFFFFFFFFFFFF and whose "net" part has the value of a reachable network,
are authorized to transit (normally, these packets, whose broadcast is limited,
are rejected). You can enter this order as many times as necessary, one
per line. The notation used is decimal if the number starts with a
figure between 1 and 9, hexadecimal if the number starts with the
"0x" string, and octal if the number starts with 0. Only certain highly-specific applications
request the passage of this type of frame  (which, despite everything, complies
with the novell specification). Consequently, this order will rarely be used.
If num is replaced by the '*' character, all the frames whose 
"net" address will be specified, and whose "node" address will be broadcast, will be
sent. This function may be necessary for using
Prolog above IPX. In this case, "num" is generally equal to 0x4900.
SEE: variable IPXNET.


[ipx.hosts]

contains the correspondence between the symbolic name of the IPX servers, their
address, their service number and their socket. These different fields are
described as follows:
address_IPX/service/socket symbolic_name
address_IPX      address of the machine (in IPX format) xx.yy or xx (on 32
bits) and yy (on 48 bits) are two numbers in hexadecimal format.
service          service number (in hexadecimal).
socket           socket number (in hexadecimal).
symbolic_name   symbolic name (character string limited to 48 characters).

This file is used for static IPX routings defined in the
numipx file, read by opera. The information extracted from this file will be
used in the SAP messages as long as the network described will not have been
reached. To fill in this file correctly, you are advised to establish
a connection with the remote site first, and to note the machines and services
announced using the following commands:
netstat -R
netstat -Rn
then to fill in the ipx.hosts file using this information (respecting
upper/lowercase).
Attention: Announcing invalid services can seriously disturb
IPX server and client operation.
SEE: command netstat.


[isdn.cfg]

contains the description of your ISDN subscription used by the device. The
file contains a single section consisting of a title (ISDN connector number)
and a list of parameters in the format:
parameter_name=parameter_value (one parameter per line). The parameters are as follows:

descriptor: name of the file describing the ISDN network used by the device.
	A number of files are available. The descriptor files always have
	the extension .ipo.
	Examples: vn3.ipo for France Telecom VN3, ni1.ipo for Baby Bell
	Companies, etc.
	Default value: none.
spid_1: service number 1 profile identifier. This parameter is only to be used
	on NI-1 (Baby Bell Companies). It contains a number
	of values corresponding to the SPID to be used.
	Default value: empty.
spid_2: service number 2 profile identifier. This parameter is only to be used
	if two different SPIDs have to be used.
	Default value: empty.
nac_connector: support connector number. Assign the same value as
	that in the section's title, i.e., 0.
	Default value: none.
nac_protocol: (physical protocol expected by the connector) 0x0414 is the only
	authorized value.
	Default value: none.
phy_power: 1 for yes, 0 for no. Energy loss detection
	(cable disconnected, electrical failure in the network). The
	value 0 is to be used if the connector is hooked up to a PBX
	that does not supply any power on the junction.
	Default value: 1.

Precautions must be taken when modifying the contents of the isdn.cfg file. The number
and names of sections, together with the connector and protocol parameters, must
never be modified.
To save on space, parameters taking their default value
(specified above) can be omitted. Comment lines can be
inserted in the file. In this case, the line's first non-blank character must
be the semi-colon.
For the 200 series (114, 118, 211):
- the section must be [1] and only the "descriptor" and 
  "phy_power" parameters are managed to date,
- for "descriptor", the following values are recognized (default "vn6.ipo"):
  "vn3.ipo" (33/3), "etsi.ipo" (1000/3), "tr6.ipo" (49/4), "vn6.ipo" (33/3).
For the 100 series (CyberConnect, CyberPAD, 101, 111):
- for descriptor, the values that are acceptable are associated with the
  description files present in the configuration (ls -a | grep .ipo).


[isdn.prf]

contains the description of the various ISDN profiles that can be used
to establish B channels. The file contains a number of sections, each of which
contains a title (profile number between brackets), and a list
of parameters in the format:
parameter_name=parameter_value (one parameter per line).
The parameters are as follows:

default_profile: 1 for yes, 0 for no. This parameter indicates whether this profile is
	to be used by default on the physical link (see file plm.cfg)
	specified by parameter ph_link_id. The default profile is used
	for incoming calls, and for outgoing calls if no other
	profile is specified in the call request.
	Default value: 0.
ph_link_id: this parameter indicates the physical link to which
	this profile is assigned by default. This parameter is ignored if
	default_profile=0. The only significant value possible is 0.
	Default value: 65535.
SEE: file plm.cfg.
spid_number: 1, 2 or 0. Service Profile Identification number (SPID) to
	be used. Put 0 if the SPID concept is not used (networks
	other than NI-1).
	The different SPIDs are declared in the isdn.cfg file.
	Default value: 0.
rem_number_type: (remote number type)
	- 0 for unknown (recommended)
	- 1 for international number
	-2 for national number
	- 3 for network-specific number
	- 4 for subscriber number
	- 6 for abbreviated number
	Default value: 0.
rem_number_plan: (remote number numbering plan)
	- 0 for unknown (recommended)
	- 1 for telephone/ISDN numbering plan
	- 8 for national numbering plan
	-9 for private numbering plan
	Default value: 0.
rem_number_chk: 1 for yes, 0 for no. Verification of the remote subscriber's number
	(type and plan only) for an incoming call.
	Default value: 0.
rem_subaddr_type: (remote subaddress type)
	- 0 for NSAP (X213 / ISO8348)
	- 2 for user-specific
	Default value: 0.
rem_subaddr_chk: 1 for yes, 0 for no. Verification of the remote subaddress
	(type only) for an incoming call.
	Default value: 0.
loc_number_type: (local number type)
	- 0 for unknown (to be used in most cases)
	- 1 for international number
	-2 for national number
	- 3 for network-specific number
	- 4 for subscriber number
	- 6 for abbreviated number
	- 255 for unsupplied local number
	Default value: 0.
loc_number_plan: (local number numbering plan)
	- 0 for unknown (recommended)
	- 1 for telephone/ISDN numbering plan
	- 8 for national numbering plan
	-9 for private numbering plan
	Default value: 0.
loc_number_pres: (presentation of the local number to the remote subscriber)
	- 0 for presentation authorized
	-1 for presentation unauthorized
	- 255 for information not used (recommended)
	Default value: 255.
loc_number_chk: 1 for yes, 0 for no. Verification of the local subscriber's number
	(number's digits only) for an incoming call.
	Default value: 0.
loc_subaddr_type: (local subaddress type)
	- 0 for NSAP (X213 / ISO8348)
	- 2 for user-specific
	Default value: 0.
loc_subaddr_chk: 1 for yes, 0 for no. Verification of the remote subaddress
	local (type and value) for an incoming call
	Default value: 0.
user_info_prot: (type of protocol for the user data in the
	ISDN packets)
	- 0 for user-specific protocol
	- 1 for OSI higher layer protocol
	- 4 for AI characters no5
	- 8 for protocol Q.931
	- 255 for unused user data
	Default value: 4.
user_info_chk: 1 for yes, 0 for no. Verification of the remote user's data
	(type only) for an incoming call.
	Default value: 0.
high_layer_prot: (type of high layer protocol)
	- 56 for X400 messaging
	- 65 for X200 application
	- 255 for information not used (recommended)
	Default value: 255.
high_layer_chk: 1 for yes, 0 for no. Verification of the high layer protocol
	for an incoming call.
	Default value: 0.
low_layer_prot_1: (type of layer 1 protocol)
	- 1 for adaptation of the standardized throughput as of 56Kbits/s
	   (calls to or from NI-1 subscribers)
	- 9 for X31-compliant throughput adaptation
	- 255 for information not used (recommended)
	Default value: 255.
low_layer_prot_2: (type of layer 2 protocol)
	- 6 for X25 link layer
	- 255 for information not used (recommended)
	Default value: 255.
low_layer_prot_3: (type of layer 3 protocol)
	- 6 for X25 packet layer
	- 255 for information not used (recommended)
	Default value: 255.
low_layer_chk: 1 for yes, 0 for no. Verification of the low level protocol
	(layers 1, 2 and 3) for an incoming call.
	Default value: 0.
bearer_transfer: (bearer's operating mode) 8 for numeric information
	with no restrictions in circuit mode at 64 kbits/s (value
	unique and compulsory).
	Default value: none.
bearer_capa_chk: 1 for yes (recommended), 0 for no. Verification of the bearer's
	operating mode for an incoming call.
	Default value: 1.

To save space, you are advised to omit parameters
taking their default value (specified above). Comment lines
can be inserted in the file. In this case, the line's first non-blank
character must be a semi-colon.
For the 200 series (114, 118, 211), the section must be [1] and only
parameters "loc_number_chk" and "loc_subaddr_chk" are managed to date.



[issue]

(dynamic)
This file contains the message displayed each time the session is closed.
SEE: file motd.


[lapb.prf]

contains the description of the various LAPB profiles (X25 link layer)
that can be used to establish X25 virtual circuits. The file contains
a number of sections, each with a title (profile number
between brackets) and a list of parameters in the format:
parameter_name=parameter_value (one parameter per line).
The parameters are as follows:

default_profile: 1 for yes, 0 for no. This parameter indicates whether this profile is
	to be used by default on the physical link (see file plm.cfg)
	specified by parameter ph_link_id. The default profile is used
	if no other profile is specified when the
	physical link is established.
	Default value: 0.
ph_link_id: this parameter indicates the physical link to which
	this profile is assigned by default.
	Default value: 65535.
See file plm.cfg for the possible values.
auto_role: 1 for yes (recommended), 0 for no. Automatic determining of the
	LAPB layer's role.
	Default value: 1.
dce_role: 1 for yes, 0 for no. The LAPB layer plays the role of DCE. This
	parameter is meaningless if auto_role=1.
	Default value: 0.
set_mode_delayed: 1 for yes, 0 for no (recommended). The establishment of
	level 2 is delayed by T1 (see below)
	Default value: 0.
generate_clock: 1 for yes, 0 for no (recommended). The LAPB layer
	generates the clock on the line.
	Default value: 0.
extended_mode: 1 for yes, 0 for no. Use of extended numbering
	(modulo 128).
	Default value: 0.
line_speed: (to be supplied even if the clock is not generated
	by the local LAPB layer).
	Default value: 9600.
t1:	timeout for frame retransmission (in tenths of a second).
	Default value: 10.
t2:	timeout for data frame acknowledgement (in tenths of a
	second).
	Default value: 1.
t3:	timeout for inactive channel detection (in tenths of a second).
	Default value: 50.
n1:	maximum number of bytes that a frame can contain, header and CRC
	included.
	The most usual case is: max(o_packet_size, i_packet_size) + 7,
	in other words for a subscription with 128-byte frames (which implies that
	o_packet_size is equal to i_packet_size which is equal to 128); you
	will set this value to 135.
	Default value: none.
n2:	number of frame repetitions.
	Default value: 10.
k:	size of window.
	Default value: none.

Note: for "t1", "t2", "t3", the rules to apply are as follows:
	t2 <= t1/2	et	t3 >= t1*5
To save space, you are advised to omit parameters
taking their default value (specified above). Comment lines
can be inserted in the file. In this case, the line's first non-blank
character must be a semi-colon.



[lcd.conf]

(dynamic, 114/118/211 only)
contains a script which will be displayed on an LCD
display. This script will be run in a loop, thus in the event
of an error or at the end of a file, the script will be rerun from the start. Since the loop
is permanent as of the initialization, you must set
the PAUSELCD variable to 1 and wait for the end of the current execution.
Subsequent to the modification, resetting PAUSELCD to 0 or removing this variable
will authorize the loop to restart.
SEE: variable PAUSELCD, command script.


[loceth]

contains the description of the presence frames, according to the same format as that
of the simeth file. This file is updated using the catch command.
SEE: command catch.



[login]

(dynamic)
if this file exists, the commands contained will be executed for each
connection to the device's desk (i.e., after ^G or password
on the console port or by telnet).
The syntax of the commands is the same as that used by the
"script" command. The contents of the "motd" file (if it exists) will be displayed
before the "login" file is run. This file is a relative equivalent
of the ".profile" or ".login" files on Unix machines.
Example, if this file contains:
	date '+Hello, it is %T on the %D%n'
For each connection to the device's desk, the message:
	Hello, it is HH:MM:SS on the DD/MM/YY
will be displayed (where HH, MM, SS, DD, MM, YY are, respectively, hour, minute,
second, day, month, year).
SEE: command script.



[motd]

(dynamic)
This file contains the message displayed for each connection to the desk (Message
Of The Day).
SEE: file issue.



[networks]

(dynamic)
contains the correspondence between the symbolic names of the networks and their
address. These symbolic names can be used in place of the IP addresses
in numerical format. The lines in the networks file contain 2 or 3
fields:
name adr alias
name    the network's official name (in the form of an uninterrupted character
       string).
adr    the network's address in IP format: 1 to 4 bytes in decimal, separated by
       the '.' character depending on the network's class.
alias  the alias of the network's official name (in the form of an uninterrupted character
       string). The field separator is either a
       space or the tabulation character. The syntax is usually
       compatible with file /etc/networks on Unix systems.


[null]

this file does not exist, or rather, it does not contain anything and
cannot be filled or read. It can be used to test the
transfer by vtftp, ftp or hload. Whatever is transferred is forgotten
on receipt.
This makes it possible to transfer data for test purposes without changing
the configuration memory.
SEE: command vtftp, command ftp, command hload.


[numeth]

contains the associations used to establish MAC remote communications
(Ethernet address). The details of the different sites to
be reached are defined in the following format:
adr_ether=num:cdg[:type][:class][,num:cdg[:type][:class]].....
 adr_ether    is given using 6 bytes in hexadecimal; it
 represents the address of the remote machine to be reached.
Fields num, cdg, type and class are the same as the fields in the
numip file.
SEE: command opera, file numip.


[numipx]

contains the associations used to establish remote communications
to the IPX networks. The details of the various sites and/or machines
to be reached are defined in the following format:
address_IPX= num:cdg[:type][:class][,num:cdg[:type][:class]].....
address_IPX is given using 2 hexadecimal fields separated by the
'.' characters; it represents the address of the remote machine or network to be
reached.

If the second field is equal to 0, this entry will be considered as describing
a network.
Note that in order for a machine described in this file to be announced, you
must describe it in the ipx.hosts file. You must also
declare the network.
The other fields are the same as those in the numip file.
SEE: command opera, file numip.


[pad.cfg]

This file is used to customize the unit's PAD service in terms of
the commands (characters) that can be sent to it and the
signals (characters) that it can generate. This file is described in the
PAD user's manual.


[pad.prf]

This file contains a set of PAD profiles supplied as standard. It
contains a number of sections, each with a title (profile number
between brackets) and a list of parameters in the format:
parameter_name=parameter_value (one parameter per line).
The parameters are arranged in order of their number (parameter 1 first,
parameter 22 last). The parameters with the PAD profile are described in
the PAD user's manual.


[pad.prf.usr]

This file contains a set of PAD profiles defined by the user. Its
contents are similar to those of the pad.prf file. However, with the
pad.usr.prf file, the contents are modifiable.


[paf.cnf]

This file contains the information required for the paf daemon. Its contents
are described after the command.
SEE: command paf.



[passwd]

(dynamic)
This file contains the passwords used for the optional authentication
(PAP) of remote sites when establishing PPP sessions.
It is also used to verify the password associated with the
user ID for the connection to the desk (according to the value of the PASSWD variable).

Each line in this file corresponds to a remote site and proposes the following fields
separated by ':':
1. field1: (max 16 char.) user name
2. field2: (max 16 char.) user's encrypted password. To enter a
   new remote access (name/password), you must use the "defuser" command
   then the "passwd" command.
   CAUTION: never enter a password in this field using the
   editor; the password is encrypted by the "passwd" command and
   manually entering a character string here may
   make you lose your access to the device.
The password encryption algorithm is incompatible
with that on Unix systems.
SEE: command passwd, command defuser, variable PASSWD.


[plm.cfg]

This file contains the description of the physical links available for the
ISDN and X25 protocols. The file contains a number of sections, each with
a title (link number between brackets), and a list
of parameters in the format:
parameter_name=parameter_value (one parameter per line).

The parameters are as follows:

ccm_module: (link type – management module identifier)
	- 1 for a leased line
	- 2 for D channel
	- 130 for B channels
	Default value: none.
connector: support connector number:
	- 0 for the S0 connector
	- 1 for the serial channel
	Default value: none.
protocol: physical protocol expected by the connector)
	- 0x03fe for a leased line
	- 0x0414 for D channel
	- 0x007e for B channels
	Default value: none.
direction: (direction of the link) its must always take the value 2 except for
	the type 1 link (ccm_module=1) if the X25 layers must play
	the role of DCE. In this case, the value to use is 1.
	Default value: 2.
connection_id: this parameter must be set to 65535 except for the type
	2 link (ccm_module=2). Its value is thus the number of the TEI to be used
	to transfer X25 packets in the D channel.
	Default value: 65535.
establish_mode: 1 for yes, 2 for no. This parameter indicates whether
	the establishment of the physical link is to be monitored by a
	timeout.
	Default value: 1.
establish_timeout: this parameter indicates (in seconds) the duration of the
	timeout monitoring the establishment of the physical
	link. Parameter ignored if establish_mode=2.
	Default value: 60.
release_mode: this parameter indicates when the physical link must be released
	once no longer used.
	- 0 for immediate release
	- 1 for deferred release (see parameter release_timeout)
	- 2 for inhibited release (appropriate for
	    type 1 or 2 links)
	Default value: 2.
release_timeout: this parameter indicates (in seconds) the duration of the
	timeout before the release of the physical link. Parameter
	ignored if establish_mode=0 or 2.
	Default value: 30.
local_number: not used (except for certain use of the pad).
remote_number: not used.

Precautions must be taken when modifying the contents of the plm.cfg file. The number and
names of sections, as well as parameters ccm_module, connector and
protocol must never be modified.


[plp.prf]

contains the description of the various X25 profiles (packet layer)
that can be used to establish X25 virtual circuits. The file contains a certain
number of sections, each with a title (profile number
between brackets) and a list of parameters in the format:
parameter_name=parameter_value (one parameter per line).
The parameters are as follows:

default_profile: 1 for yes, 0 for no. This parameter indicates whether this profile is
	to be used by default on the physical link (see file plm.cfg)
	specified by parameter ph_link_id. The default profile is used
	if no other profile is specified when the
	physical link is established.
	Default value: 0.
ph_link_id: this parameter indicates the physical link to which
	this profile is assigned by default.
	Default value: 65535.
See file plm.cfg for the possible values.
auto_role: 1 for yes (recommended), 0 for no. Automatic determining of the
	X25 packet layer's role.
	Default value: 1.
dce_role: The X25 packet layer plays the role of DCE: 1 for yes, 0 for no.
	This parameter is meaningless if auto_role=1.
	Default value: 0.
dce_diag_packet: the packet layer playing the role of DCE can send
	DIAGNOSTIC packets: 1 for yes, 0 for no.
	Default value: 1.
dte_restart_packet: 1 for yes, 0 for no. The packet layer playing the role of
	DTE can send RESTART packets.
	Default value: 0.
dte_restart_delayed: 1 for yes, 0 for no. The packet layer playing the role of
	DTE sends the first RESTART packet after a period of T20 (see
	below).
	Default value: 1.
dte_cause_field: 1 for yes, 0 for no. The packet layer playing the role of DTE
	can send cause fields codes 1xxxxxxx.
	Default value: 0.
extended_numbering: 1 for yes, 0 for no. Use of the extended number
	(modulo 128).
	Default value: 0.
flow_negociate: 1 for yes, 0 for no. Negotiation of the flow control parameters
	(window and packet sizes).
	Default value: 0.
out_call_barred / in_call_barred: 1 for yes, 0 for no. Barring of
	outgoing/incoming X25 calls.
    Default value: 0.
cnt_perm_ch / cnt_in_ch / cnt_two_ch / cnt_out_ch: number of permanent virtual
	/ incoming switched unidirectional / switched
	bidirectional / outgoing switched unidirectional circuits.
	Default value: none.
low_perm_ch / low_in_ch / low_two_ch / low_out_ch: number of the first permanent
	virtual / incoming switched unidirectional / switched
	bidirectional / outgoing switched unidirectional circuit.
	Default value: none.
high_perm_ch / high_in_ch / high_two_ch / high_out_ch: number of the first
	permanent virtual / incoming switched unidirectional / switched
	bidirectional / outgoing switched unidirectional circuit.
	Default value: none.
o_window_size / i_window_size: size of the window for sending/receiving.
	Default value: none.
o_packet_size / i_packet_size: size of the packet for sending/receiving.
	Default value: none.
o_buffer_size: size of the anticipatory buffer for sending.
	Minimum value: 2 * o_packet_size.
	Default value: none.
i_buffer_size: size of the anticipatory buffer for receiving.
	Minimum value: i_window_size * o_packet_size.
	Default value: none.
r20_counter / r22_counter / r23_counter: number of repetitions of the following packets:
	RETART / RESET / RELEASE on the DTE side.
	Default value: 2.
t10_timer / t12_timer / t13_timer: resend timeout for the following types of packets:
	RETART / RESET / RELEASE on the DTE side (in seconds).
	Default value: 60.
t20_timer / t22_timer / t23_timer: resend timeout for the following types of packets:
	RETART / RESET / RELEASE on the DTE side (in seconds).
	Default value: 180.
t11_timer: wait timeout for a reply to a call packet sent on the
	DCE side (in seconds).
	Default value: 180.
t21_timer: wait timeout for a reply to a call packet sent on the
	DTE side (in seconds).
	Default value: 200.
trs_timer: recovery timeout in the event of a level 2 or 3 drop (in
	seconds).
	Default value: 10.

To save space, you are advised to omit parameters
taking their default value (specified above). Comment lines
can be inserted in the file. In this case, the line's first non-blank
character must be a semi-colon.


[poolip]

contains the description of the dynamically-allocated addresses for connections
in PPP/IP only. The address extracted for a connection will be supplied
to the remote machine which must use it as its address. This
function enables you, for example, to provide Internet access to a
large virtual population of users while only using
a limited number of addresses (number of addresses usually equal to the number
of possible simultaneous connections on the device).
In the "ppp.conf" file, you must specify that you wish
to assign addresses to the remote devices or networks (parameter "+sa").
The recognized addresses that belong to the LAN will automatically be
declared in the ARP as "published" (similar to "opera -proxy").
The addresses of the "pool" can be defined as a range of addresses
in the format "minimum_address – maximum_address", whereby these two addresses
are included in the range, or in the format "address". You can
put as many lines as necessary, making sure, however, that you use only
one expression per line.


[ppp.conf]

contains the options of PPP sessions that can, if necessary, be negotiated with the
remote device/access on connection. Each line in the file only contains one
options and its arguments (if necessary). A comment starts with the
'#' character (after the option's arguments or on its own line).

	The following options are possible:
-all		invalidates all the options, even those by default.
-ac		forbids the compression of control fields and
		addresses in PPP frames.
active		forces the active mode. By default, an incoming connection is
		handled in passive mode, an outgoing connection in active mode.
-am		invalidates the negotiation of the character map that must be
		preceded by an escape sequence (asynchronous PPP
		only).
-as val		sets the character map that must be preceded by an
		escape sequence (asynchronous PPP only).
		val is a hexadecimal integer containing, in the form of
		binary elements, the fact that the character concerned (bit 0 for
		the character 0, etc.) is "escaped". This value only
		concerns the first 32 characters (0 to 31).
		You are advised against modifying the default character map.
autoroute	for the IPCP/CI_ADDR negotiation
		- If an address is given and no remote number has been
		received, the remote number present in the log will be
		deduced from this address, provided it figures in the
		numip file.
		- If an address is given and you are working in
		IP address translation mode with a fixed address (other
		than 0), the address received will be refused, and
		the translation address proposed (if they are different).
		- If an address is given and you are working in
		IP address translation mode with a variable address (equal to 0),
		the address received will be used as the translation address
		for this interface.
		- If no address is given (value at 0.0.0.0),
		an address will be proposed: the translation address if you
		are working in IP address translation mode, or the address of the
		interface on the LAN side otherwise.
		! Note that this keyword must or must not be present
		! according to the remote device (only a test will give the
		! reply). In fact, this keyword changes the address
		! assignment strategy. You are advised to perform an
		! initial test without this keyword. The communication will
		! or will not work; the use of the strategy
		! specified by "autoroute" does not affect the
		! speed of the link.
SEE: file trans.conf, file numip.
callwait is an integer representing the time between
		the event requesting the callback and the callback
		order. This value is expressed in 1/50s of a second.
		When the callback is on the initiative of the device
		(option 'c' in numxx), it's default value is
		50 (1 second), its minimum value 25 (0.5 seconds),
		and its maximum value 500 (10 seconds). This period will
		be used in RSD mode and/or in CHAP callback mode.
		If the value given is outside this range, the default value
		will be used.
		When the callback is on the initiative of the remote device
		(option 'CB' in numxx), it's default and minimum value is
		200 (4 second) and determines the minimum delay
		between 2 calls made to the remote device.
SEE: file numip.
+chap		imposes the authentication of the remote device in accordance with the
		CHAP (Cryptographic Handshake Authentication
		Protocol) protocol. The two devices in communication can
		have the option set; in this case, the authentication procedure
		is mutual since it is ensured by each
		part. See file secret for the definition
		of passwords and their association with the remote devices
		and variables MYNAME and DOMAIN (see also this file's
		"myname" option).
SEE: file secret, variable MYNAME, variable DOMAIN.
-chap		forbids the CHAP. It may be necessary to use this keyword
		for certain devices made by other manufacturers, 
		for example, when they can handle PAP and CHAP,
		and only PAP is configured.
chaprange min max  forces the minimum and maximum limits to use the
		the "min" and "max" values for the CHAP challenge.
		These values are usually defined by the RFC, but some
		competitor devices have a problem in this regard. You
		must therefore limit the size to interoperate
		correctly with these devices. For certain versions of
		Radius, you will have to use the value 16
		for the min and max parameters.
		If the values chosen are outside the authorized limits
		(5 to 64), they will be ignored.
-cipx		forbids the compression of IPX and NCP headers in IPX frames
		when the device works as an IPX router.
-d		validates the debug option.
debug		equivalent to -d.
-mn		forbids the negotiation of the magic number.
-mru		forbids the negotiation of the maximum size of frames
		being received.
mru n		authorizes the negotiation of the MRU to the value n; the default value
		is 1500 bytes.
myname name	replaces the default chap user id ($MYNAME.$DOMAIN) with
		"name". This makes it possible, when using a number
		of profiles, for different names to be presented to the remote devices.
-p		forces the passive mode.
passive		forces the passive mode. The negotiation will always be on
		the initiative of the remote device. If the remote device is also in
		passive mode, the connection will necessarily fail.
-pap		equivalent of -ua.
+pap		equivalent of +ua.
-pc		forbids the compression of the protocol field in PPP frames.
ppp_user user	replaces the PAP user ID defined in PPP_USER by "user".
ppp_pwd passwd	replaces the PAP password defined in PPP_PWD by "passwd".
+sa		in the IPCP/CI_ADDR negotiation, proposes the IP address
		corresponding to the connection parameter (remote number)
		in the "numip" file. This option is used to
		connect isolated machines or routers in address translation
		mode, to which you must supply an IP address.
		NB: in numip, the IP address corresponding to the remote number
		must therefore not be "default" or "0.0.0.0".
+smip		in the IPCP negotiation, on some devices, the
		knowledge of the other part's address is
		necessary (?) whether you are working with or without the address
		translation. To specify our address (N.B.: disclosure
		of sensitive information), if this option
		has been validated, you send our address in the CI_ADDR parameter
		right at the start of the negotiation.

-ua		forbids PAP.
+ua		imposes the authentication of the remote device in accordance with the
		UPAP (User Password Authentication Protocol) protocol. See
		commands "passwd" and "defuser", file "passwd" and
		variables PPP_USER and PPP_PWD and options ppp_user and
		ppp_pwd in this file.
		NB: UPAP is commonly – and abusively – called PAP.
SEE: command defuser, command passwd, variable PPP_USER, variable PPP_PWD.
-vj		forbids Van Jacobson compression of TCP/IP headers.
		This keyword may be necessary for correct operation
		with other manufacturers' devices
		(ICMP or UDP transmissions are not affected (ping), however
		TCP connections only work very, very slowly
		(telnet, ftp, etc.)).
vjmode mode  where "mode" can take the following values:
	old	 for a header compression compliant with the old RFC,
	rfc1172  for a header compression compliant with this RFC,
	rfc1332  for a header compression compliant with this RFC,
	by default, mode "rfc1332" is selected.

    !! MP is available in versions from 02/99 onwards!!

mp		activation of the MP overflow module.
		This parameter is the only one required to activate the
		overflow, since the following mp parameters have default
		values designed for everyday use:
mpssnh		requests and accepts the use of short MP sequence
		numbers.
mpminnchan	followed by an integer representing the minimum number of a channels
		to use to the same destination in MP (default = 1).
mpmaxnchan	followed by an integer representing the maximum number of a channels
		to use to the same destination in MP (default = 2).
mpbasenchan	followed by an integer representing the number of a channels
		to use as soon as the connection is opened in MP (default = 1)
mpseuilhaut	followed by an integer representing the threshold from which
		an additional channel must be added in MP (default = 60%).
mpseuilHhaut	followed by an integer representing the threshold from which
		all the channels (max = mpmaxnchan) must be
		open (default = 90%).
mpseuilbas	followed by an integer representing the threshold from which
		a channel must be removed (and thus disconnected) from the list
		of channels connected to the same destination
		in MP (default = 30%).
mpseuilBbas	followed by an integer representing the threshold from which
		the maximum number of channels must become mpminnchan
		in MP (default = 15%).
mpbwsampling	followed by an integer representing the sampling period,
		expressed in seconds, of the line's throughput (default = 5).
mpbwchgdelai	followed by an integer representing the delay, expressed in seconds,
		during which – after adding/removing a channel – no
		other addition/removal of a channel will be performed on this
		group (default = 10).
mppmgt		authorizes this device to be accessed by a manager
		working above MP+ (default = no). On this
		range of equipment and with certain options, the "cu" command
		enables you to access this "virtual" console port.
(See command cu)

To be valid, the mp parameter definition must satisfy the following conditions:
	mpbasenchan  >= mpminnchan
	mpmaxnchan   >= mpbasenchan
	mpseuilbas   >= mpseuilBbas
	mpseuilhaut  >= mpseuilbas
	mpseuilHhaut >= mpseuilhaut

PPP proposes two solutions to authenticate machines connecting to a
site: CHAP and PAP.
1. PAP: the machine with the +ua option set requests the remote device
   to provide its name and password on connection, whatever the
   direction of this connection. The remote device must respectively provide its name
   and password (i.e. the contents of variables PPP_USER and
   PPP_PWD). The server then encrypts the password given and searches for the
   same pair (user,passws) in the passwd file. If this search
   is successful, the connection is accepted; if not, it is refused.
2. CHAP: SEE file secret.
Default configuration of PPP:
> compression of the control and address fields,
> passive mode if the connection is incoming, that is if the machine is
  called, otherwise active, that is for the machines on the initiative of
  the call and for the interfaces declared as leased lines,
> negotiation of the MRU,
> no authentication,
> VJ header compression,
> no debug.

You can define more than one profile. When the file is read,
as long as the '[' character is not found at the start of a line, the
configuration is stored in the default profile. This default profile
serves as the basis for all the other profiles defined. For example, if you find
the "-chap" sequence in this part, chap will be refused in all the profiles.
To define another profile, you separate the information by default (if this information
exists) by a line containing "[numpro]", where "numpro" is
the profile for which the parameters will follow ("numpro" must be
between 1 and 65534 inclusive) and the '[' and ']' characters enable
the profile number to be detected.
The options presented can all be given in a profile, but
whatever the profile selected, the "debug", "chaprange"
and "autoroute" options will have an effect on all the profiles.
These profiles can be dynamically assigned, both for
incoming and outgoing connections. To select a
specific profile in the list, in the "type" field of "numxx"
(numip, numipx, etc.), you will give the profile number preceded by the 'p' character.
If the selected profile number does not exist, the default profile
will be used.
Example of file ppp.conf:
	-vj
	[1]
	-chap
	ppp_user toto
	ppp_pwd titi
	[2]
	+chap
And in numip, you can give profile 1 for the connection to an Internet
access provider, and profile 2 (imposing CHAP) for the incoming connections.
See file numip.


[radius.conf]

contains the operating functions of authentication requests and
RADIUS (Remote Authentication Dial In User Service) records.
Radius can be used to record WAN connections with
the device (PPP and FR only) and to make
authentication requests. Note that modifications to this file are only accepted
when the device is initialized and with the
"opera -r" command.
SEE: command opera.

The recording part is grouped in the section that starts with the
mention "[accounting]".
In this section, you can find the following keywords:
  .chap     indicating that the recording of a session will only be requested
            after "ppp/chap" identification.
  
.pap     indicating that the recording of a session will only be requested
            after "ppp/pap" identification.
  .clid     indicating that the recording of a session will only be requested
            as soon as a call is received/sent. Note that the mention ".clid"
            invalidates ".chap" and ".pap", since the call must take place
            before any authentication.
  .port nnn   where "nnn" indicates the UDP port to be used (instead of 1813 by
            default, value given in the RFC).
            Note that historically, the "accounting" port took the value
            1646. You may need to use this value to obtain
            correct operation.

The authentication part is grouped in the section that starts with the
mention "[authentication]".
In this section, you can find the following keywords:
  .chap     indicating that if the remote_device_name/password pair is not
            found in the "secret" file, the authentication
            must be requested from the "radius" server.
            Note: some Radius server versions require
            a fixed-length challenge of 16. You will need
            to specify this limit in the ppp.conf file.
  .pap     indicating that if the remote_device_name/password pair is not
            found or is invalid, in the "passwd" file,
            the authentication must be requested from the "radius" server.
  .clid     indicating that if the remote device number is not found in the
            numxx file, agreement must be requested from the "radius" server.
            An entry for verification in this mode will necessarily
            have "CXR-CLID" as the password on the radius server.
  .port nnn   where "nnn" indicates the UDP port to be used (instead of 1812 by
            default, value given in the RFC).
            Note that historically, the "accounting" port took the value
            1645. You may need to use this value to obtain
            correct operation.

You should find at least one declaration in all the sections:
     name key  where "name" indicates the name of the machine (or its IP address)
             with the radius server and "key" the key shared between
             the device and the radius server. If several servers are
             proposed, the first in the list will be used; if the device
             does not receive any acknowledgement, the next one will be
             queried. When the end of the list is reached, the pointer will be repositioned at
             the start of the list.
SEE: file ppp.conf, list of RFCs.


[rap.conf]

contains the operation and restriction options of the
dynamic IP routing module based on the RIP protocol. These options are of two types:
1. restriction declarations including the following fields:
command IP_address intf {all | lan | wan}
 with:
> command can take the values below:
listen        accept the information from this address.
donotlisten   do not accept the information from this address,
sendto        send RIP frames to this address,
donotsendto   do not send any RIP frames to this address,
announce      announce this address,
donotannounce do not announce this address,
record        record the routing for this address,
donotrecord   do not record the routing for this address,
donotage      do not "age" this address.
> IP_address is the IP address concerned by this command.
> intf is a separator keyword.
> {all | lan | wan} is/are the interface(s) concerned by this command.
* lan corresponds to the local network,
* wan corresponds to the remote network(s),
* all corresponds to the LAN and WAN.

For all these options, the order take priority over the counter-order, and makes
it useless for the configuration, and changes the default value
from "yes" to "no".
For example, if "listen" is used, all the accepted addresses must
be specified and "donotlisten" will no longer have any effect on the operation.

2. Operating options.
> noripin selects the transmission of RIP frames without receiving any frames.
> noreflect forbids the rebroadcasting of the routings found on the same type of
  network (no announcement on the LAN of the routings found on the LAN
  side and no announcement on the remote network of the routings found out on the remote
  network side). As of 02/99, this option is the default option.
> reflect authorizes the rebroadcasting of the routings found out on the same type of
  network.
> debug selects the display - via syslog - of routing recordings and erasures
  (note that if a lot of routes have to be announced,
  this information represents a considerable workload).
> direct limits the transmission of RIP frames to routers specified by the
  ripgateway command.
> quiet selects the receipt of RIP frames without sending frames.
> ripgateway adr is used to define a router with the adr IP address to which a frame
  will be sent by name in addition to the transmission in broadcast mode. You
  will need one declaration of this type for each router. This declaration is
  useful in complex architectures where routers are separated by
  devices that filter the communication in broadcast mode.
> adv followed by "onconnection", "ondisconnection", "onchange", is used to announce
  the routing table for the connection of a channel, the
  disconnection of a channel and the change of status of a channel (sum of the 2
  previous orders) respectively. By default, the "onconnection" option is
  used. This announcement is performed in addition to the normal announcement.
> brec followed by the size of the receiving buffer in number of frames.
  By default (and when the value is insufficient), the size of the
  buffer is calculated automatically according to the frames received.
> itrame followed by the duration in milliseconds separating the transmission of 2
  announcement frames. Some devices may become saturated when a
  large number of routes has to be announced, and this interframe makes it
  possible to limit the transmission speed. By default, there is no
  wait and the maximum interframe accepted is 1000ms (1 second).
> version followed by 1 or 2, selects the RIP operating mode,
  bearing in mind that the main difference between versions 1 and 2 is
  the indication of the mask to be used in version 2.
Note that in version 1, only version 1 frames will be taken into account;
in version 2, the version 2 frames will be taken into account as well as
the version 1 frames, with the automatic calculation of the mask depending on the
target address. Before selecting version 2, you must check that the
machines on the network accept it.

If this file does not exist, a default configuration is provided (see
the details in the description of the "rapd" command).
SEE: command rapd, list of RFCs.


[rcmds]

contains the machines, users and commands authorized for execution in
the context of the rshd daemon. The format is given below. Any line not
corresponding to this syntax will be ignored. Comments can be inserted using
using the '#' character. The authorization lines must contain
neither spaces nor tabulations.

{*|address|name}:{*|user}:{*|cmd1[,cmd2,...,cmdn]}
or, in other words
machine:user:commands

machine       represents the remote machine which can be given as an
IP address, name (if known in the "hosts" file, or obtained
via the name server) or the '*' character, which corresponds to all
machines.
user   authorizes the user to execute commands. The
'*' character authorizes all users.
commands     represents the command(s) authorized for execution. In the
case of a series of commands, the names will be separated by the ',' character.
The '*' character authorizes all the commands to be executed.
Example 1:
# authorization of all the machines
# all the users, all the commands
*:*:*
Example 2:
# authorization of the numstat and disconnect commands
# for user "smurf" from all the machines
# on the LAN or directly connected
# (command run: rshd -lrR&)
*:smurf:numstat,disconnect
SEE: command rshd.


[reserve]

contains factory-defined system variables. These variables are ETHADDR
(device's Ethernet address) and HOSTID (device's serial
number). This file is not modifiable.
SEE: variable HOSTID, variable ETHADDR.



[secret]

(dynamic)
contains the remote identifiers and their associated passwords for the
CHAP PPP authentication procedure (see option +chap in the
ppp.conf file). The passwords are entered in clear text, but can be subsequently
encrypted using the secrypt command. The format of each line is
as follows:
	remote_user_name password

This file is useful in the case where the user wishes to authenticate
the remote users on incoming or outgoing communications. The '$' or
'#' characters must be preceded by the '\' character if you do not want them
to be interpreted respectively as a variable or a comment.
Each communicator must have – in the "secret" file – a line
corresponding to the name of the remote user associated with the same password.
The contents of this file are dynamic, i.e., any modifications
will be taken into account immediately.

remote_user_name   is the contents, in the remote part, of the MYNAME variable
	      followed, if the DOMAIN variable is initialized, by '.' then the
	      contents of the DOMAIN variable.
	      Be careful with the limits in length on the
	      DOMAIN and MYNAME variables; the latter may be automatically
	      truncated (with a message obviously).
	      Case (upper/lower) is not considered for this
	      parameter.
	      This name can be "default". This is a keyword which
	      corresponds to all the proposals that are not defined in the
	      file.
	      In the case of a communication with an access provider which
	      uses CHAP, it is usually the name of the remote router, but
	      some providers use more than one server for access to the
	      different names. In this case, you are advised to use the
	      "default" keyword.

password is the two communicators' common identifier. It is exchanged,
	      with the remote_user_name, in an encrypted packet according to the
	      MD5 Message Digest Algorithm of RSA Data Security, Inc.
Example:
*  site paris:
- file ppp.conf: +chap
- admin file: DOMAIN=idf
- admin file: MYNAME=paris
- secret file: lyon.rhone cx_paris_lyon
* site lyon:
- admin file: DOMAIN=rhone
- admin file: MYNAME=lyon
- secret file: paris.idf cx_paris_lyon

If the paris site's DOMAIN variable was not initialized, you would find
the following on the lyon site:
- secret file: paris cx_paris_lyon
SEE: file ppp.conf, variable MYNAME, variable DOMAIN.


[simeth]

contains the description of the frames to be sent using the simula command. Each
frame is defined as follows:
the '{' character indicates the start of a frame,
the '}' character indicates the end of a frame,
the ' [' and ']' characters indicate the start and end of
a repetition sequence consisting of the value of the byte to repeat, the
* character and the number of times the byte is to be repeated: this last
parameter is to be specified in decimal format. Each sequence must be
fully described on a single line.
The value of each byte must be described in hexadecimal format using
two characters ('0' to '9', 'a' to 'f' and 'A' to 'F'). A frame can be
described on several lines (with the previously-indicated restriction concerning the
[OO*N] sequences).
Example:
 {404040404040 303030303030 0800 AB CD [ff*10]}
represents a frame with the target Ethernet address 404040404040, the
source address 303030303030, type 800, followed by AB CD FF FF FF FF FF FF FF FF FF
FF.
Multicast source addresses are forbidden, as are frames whose
length exceeds 1000 bytes. The frame sent must be at least 60
bytes, even if the length of the description is less (padding with
null-value bytes).
SEE: command simula.


[slot]

contains information used and which can only be used
by the "slot" command. Do not edit this file.
SEE: command slot.


[tadi.prf]

contains the description of the various V120 profiles that can be used
to establish V120 virtual circuits. The file can contain
a number of sections, each with a title (profile number
between brackets) and a list of parameters in the format:
parameter_name=parameter_value (one parameter per line).
The parameters are as follows:

default_profile: 1 for yes, 0 for no. This parameter indicates whether this profile is
	    to be used by default on the physical link (see file plm.cfg)
	    specified by parameter ph_link_id. The default profile is used
	    if no other profile is specified when the
	    physical link is established.
ph_link_id: this parameter indicates the physical link to which
	    this profile is assigned by default.
lap_n200:   this parameter indicates the maximum number of retransmissions.
lap_k:	    this parameter indicates the number of frames in the LAP window.
lap_t200:   this parameter indicates the value of the delay before retransmission
	    (unit: 1/10 second).
lap_t203:   this parameter indicates the value of the activity timer
	    (unit: 1/10 second).
lci_n201:   this parameter indicates the frame size (in bytes) to be used.
nac_direction: 1 to play the role of the DCE, 2 to play the role of the DTE,
	    0 to automatically determine the role (DCE if the physical link
	    is incoming, DTE is it is outgoing).
	    Attention, some access providers (called therefore) need
	    to have a DCE (??) in front of them; if the connection
	    fails, you may have to modify this parameter.


[tcpad]

contains a series of lines defining how
PAD communications are opened above TCP.
Using this file, for incoming PAD connections, you can associate
a PAD connection to a given machine according to
the end of the X25 address requested (example 1).
For outgoing communications, you can associate – with a TCP port – an
X25 address that will automatically be called when a connection is
established on this dedicated port (example 2).

Example 1:
	device's X25 address:
		123456789
	contents of the "tcpad" file:
		99 duponT/telnet
		xdef duponD/telnet
		listen 4
		log

If the number called is "12345678999", a telnet connection request will be established to the
duponT machine.
For any other number, you will attempt to establish a connection
to the duponD machine. The "xdef" key indicates that for unresolved
cases, this line will be used. If this entry is not present and if
there is no corresponding address, the communication will be refused.
The PAD connection will be closed automatically if the TCP session is closed.
Similarly, if the PAD connection is closed, the TCP session will be closed.
The "listen" keyword determines the number of simultaneous connections
that are acceptable for the device in the PAD->TCP direction. If the value is
not given or is less than 2, the value 2 will be taken by default.
The "log" keyword indicates that the establishment of connections will
produce a message sent via syslog.

Example 2:
	contents of the "tcpad" file:
		o -p51 -l -a112
		o -p51 -l -a102 4201

If a connection on the TCP/4200 port is opened, a PAD connection will
automatically be attempted to the X25 112 address, using the
PAD 51 profile and saving the call request (see command padd).
If a connection on the TCP/4201 port is opened, a PAD connection will
automatically be attempted to the X25 102 address, using the
PAD 51 profile and saving the call request (see command padd).
For the first entry, you do not need to specify the port,
since port 4200 is assigned by default to command padd.
The PAD connection will be closed automatically if the TCP session is closed.
Similarly, if the PAD connection is closed, the TCP session will be closed.
SEE: command padd, service syslog.


[trans.conf]

contains optional and necessary information for using
IP address translation (see also file "numip").
- Necessary information:
  The only necessary information in this file is the address that will be
  presented on the remote networks. It must be written as
  follows:
    external IP_address
  where "external" is the password and "IP_address" is the address that will replace
  the LAN's source address when frames are sent
  to the remote networks selected (file "numip", option ":t").
  This address can be provided in the different formats used
  in IP. If the value is 0 or "auto", it must be provided
  by PPP when the connection is made with the remote site.

- Optional information:
  Each item of information below has a default value; you do not
  therefore have to enter any of them.
  For the 1st series of parameters, the main purpose of the setting
  appears when clearly-identified malfunctions occur, in other
  words, you are advised against assigning values other than those
  by default. The passwords used to modify
  the default value of the different parameters assigned dynamically by
  the device on translation are given below:
    minudp value   minimum value of the UDP port (default  8000)
    maxudp value   maximum value of the UDP port (default  32767)
    mintcp value   minimum value of the TCP port (default  8000)
    maxtcp value   maximum value of the TCP port (default  32767)

  The second series of parameters enables you to access the device when
  logical connections are established on external initiative. Validating these options
  may represent a danger, but allows for easier management
  by the operator. For all these parameters, "value" can be 0 (default)
  to forbid, or 1 to authorize:
    transicmp value  ICMP access -> device by the translation address
    lanicmp   value  ICMP access -> device by its LAN address
    transudp  value UDP access -> device by the translation address
    lanudp    value  UDP access -> device by its LAN address
    transtcp value  TCP access -> device by the translation address
    lantcp   value  TCP access -> device by its LAN address
  NB: the LAN's address is that assigned to the "il0" interface.

  The third series of parameters is used to select special functions
  for the address translation.
    mcast	authorizes the receipt of multicast frames from the WAN and their
		transmission on the LAN without being modified.
    vdo		authorizes the handling of VDOLive(tm) connections without having
		to use this software with static ports.
    sustain sec specifies the period during which a non-used
		translation entry remains valid (value
		expressed in seconds, 5 by default).
		On expiry, the entry will be destroyed and a frame from
		outside will not reach its recipient.
    flush	requests the immediate destruction of the entries concerned
		by the address translation of the channel, which closes.
    autoflush	requests the immediate destruction of the entries concerned
		by the address translation of the channel, which closes if
		the translation address is acquired dynamically.
    dropold	forbids the establishment of a WAN connection to send
		a reset of a TCP end of connection (often unjustified,
		due to offhand server behavior).
    tcpsum	normally the TCP checksum is not verified
		when datagrams are received, since the links
		are reputed to be reliable. However, it may turn out that the quality
		of the transmission or of certain hardware requires
		an additional verification on this level.
		When this keyword is present, the TCP checksum
		is verified. Frames whose checksum
		is false are eliminated.
    debug val	sets the debug level to the value "val".
		Enables you to obtain, via syslog, the information required
		for searches on the eliminated frames. It generally
		concerns connection attempts from
		outside on ports that are not redirected internally.

  The fourth series of parameters is used to access machines on the
  LAN when logical connections are established on the initiative of a remote device.
  There is no doubt that this represents a breach in the protection function,
  however this feature allows for standard operation of
  certain address translation applications that were not originally planned for.
  In principle, you define the address of a machine present on the LAN
  for certain TCP or UDP ports. If a logical connection
  is opened to this port from outside,
  the target address will be that defined in this port's
  configuration.
    tcport IPAdronLan/TcpPort
    udport IPAdronLan/UdpPort
  For values less than 1024, the port must be specified by its
  name. Beyond this value, the number will be accepted.
  E.g.:
    tcport $MYADDR/auth
  An authentication frame, sometimes used more or less needlessly by
  certain servers (the RFC931 indicates that this information is not
  necessarily sure) intended for the translation address (the
  only one visible on the WAN) will be directed to the device itself (provided
  that $MYADDR does indeed represent the device's LAN address).
  Since the device does not have any authd service, an ICMP frame will
  be immediately returned, giving the information to the caller server
  which, more often than not, will continue its job without waiting any longer for
  a hypothetical reply. Caution: even if this configuration sometimes
  proves necessary, it represents a breach in security, since an outside
  connection attempt generates traffic in return.

  If you wish to redirect a range of ports to the same machine, you
  can either give as many tcport/udport lines as necessary,
  or specify the limits by noting TcpPort/UdpPort as
  minimal_port:maximal_port (the ':' character is the separator).
  The minimal_port and maximal_port values are included in the range.
  With this additional information, you obtain the following syntax:
    tcport IPAdronLan/MinTcpPort:MaxTcpPort
    udport IPAdronLan/MinUdpPort:MaxUdpPort

  When the translation is carried out to a number of addresses
  (special case to be examined with your access provider), you can
  select the address seen on the remote network side by giving the offset in
  relation to the basic address (defined externally). The syntax becomes:
    tcport IPAdronLan/MinTcpPort[:MaxTcpPort] @offset
    udport IPAdronLan/MinUdpPort[:MaxUdpPort] @offset
  where "@offset" can be a value in the format "@+1" or "@-1", a value
  which will give the address on the remote network side (default value: 0).
  E.g.: with "external auto" and "192.168.1.1" assigned by the access provider
    tcport 192.168.2.1/telnet @+1
  will direct all the frames intended for 192.168.1.2/port 23 to
  the machine with the address 192.168.2.1.

  If you want to redirect a specific port on the WAN to a
  specific address with a different port on the LAN, you will use
  the following syntax:
    tcport IPAdronLan/LanTcpPort @0/WanTcpPort
    udport IPAdronLan/LanUdpPort @0/WanUdpPort
  E.g.: with "external auto" and "192.168.1.1" assigned by the access provider
    tcport 192.168.2.1/telnet @0/2200
  will make it possible to reach the "telnet" port of the machine on the LAN with
  the address "192.168.2.1" by requesting a connection from the
  remote network on the address "192.168.1.1" (translation address), port "2200".
  In this example, the offset is not used, but the configuration
  authorizes it ("@0" can be, for example, "@-1").

  You can also define a range of ports using the
  syntax defined above:
  ("LanTcpPort" becomes "MinLanTcpPort:MaxLanTcpPort" and
   "LanUdpPort" becomes "MinLanUdpPort:MaxLanUdpPort").
  The port no. on the WAN ("2200" in the example) is sufficient to limit the range
  of ports on the WAN, whereby the link is 1 port for 1 port.
  E.g.: with "external auto" and "192.168.1.1" assigned by the access provider
    tcport 192.168.2.1/telnet @0/2200
  wan 192.168.1.1/2200 is directed to lan 192.168.2.1/23(telnet)
  wan 192.168.1.1/2201 is directed to lan 192.168.2.1/24
  wan 192.168.1.1/2202 is directed to lan 192.168.2.1/25(smtp)

  NB: 1) The use of the offset is worthwhile and has the desired effect
  only if the access provider handles it (routing of several addresses).
      2) Redirecting a port to another
  requires excellent knowledge of the network's topology and is
  only used for highly specific needs.

  You can also direct all the frames for a specific
  WAN address to a specific LAN address.
  To indicate this operation, use the following syntax:
    static IPAdronLan @offset
  E.g.: with "external auto" and "192.168.1.1" assigned by the access provider
    static 192.168.2.1 @+1
  will direct all the frames intended for 192.168.1.2 to 192.168.2.1
  Symmetrically, all the frames with the source address 192.168.2.will appear,
  after translation, as 192.168.1.2.
  The port numbers will not be modified.
  NB: The use of the offset is worthwhile and has the desired effect
  only if the access provider handles it (routing of several addresses).

Note, the redirection of incoming ports (keyword "tcport", "udport") and
especially the static translation (keyword "static") reduce the
protection provided by the address translation.

See: list of port names.

  One last series of parameters enables you to assign – if necessary
  and if supported by your access provider – several
  translation addresses.
  To define the address used by a range of addresses, you
  will use the following syntax:
    segment lan_address_min [lan_address_max] @offset
  "lan_address_min" represents the first address in the address range
    on the LAN side to be considered.
  "lan_address_max" is an optional parameter used to end
    the range to be considered (if this parameter is not specified,
    "lan_address_min" will be used).
  "@offset" represents the offset in relation to the address obtained
  dynamically ("external auto") or statically. The value of the offset
  can be positive or negative and must be preceded by the '@' character.
  A range of addresses that is not described will use the basic translation address.
  The values of "lan_address_min" and "lan_address_max" are included in
  the segment defined.
  The "segment" parameter is ineffective for incoming IP connections,
  as opposed to the "static" parameter. It can thus be accompanied
  by the necessary "tcport" and "udport" keywords.

In operation, you can consult information on the address translation
using the "numstat –t" command.
SEE: file numip, file ppp.conf, command numstat.


[update]

This file is actually totally virtual. This name is used to update
the device's software version. When this file is loaded,
the end of its loading proposes an execution. This is when
the contents of the file will be placed in the flash memory (program memory).
Since this procedure must be performed by an authorized personnel member, we
will not give further details on the intrinsic operation.
NB: the update cannot be performed from one device to another
(code downloaded different from that present in the memory).


[watch.conf]

contains the configuration at the initialization of the program for displaying the
Ethernet frames received. You are advised against modifying the data contained
in this file other than via the watch command.
SEE: command watch.

 __________________________________CXR____________________________rev-__
[Back to index]