Release bumped to "gts4".

[tac_plus.git] / users_guide

                        TAC_PLUS Developer's Kit vF4.0.2.alpha
                        --------------------------------------

Author: Lol Grant

Note: this is a DEVELOPER'S KIT.  You probably shouldn't be using this
if you don't need source code. Instead, consider using CiscoSecure,
Cisco's supported, commercial Tacacs+ daemon.

   Copyright (c) 1995-1998 by Cisco systems, Inc.

   Permission to use, copy, modify, and distribute this software for
   any purpose and without fee is hereby granted, provided that this
   copyright and permission notice appear on all copies of the
   software and supporting documentation, the name of Cisco Systems,
   Inc. not be used in advertising or publicity pertaining to
   distribution of the program without specific prior permission, and
   notice be given in supporting documentation that modification,
   copying and distribution is by permission of Cisco Systems, Inc.

   Cisco Systems, Inc. makes no representations about the suitability
   of this software for any purpose.  THIS SOFTWARE IS PROVIDED ``AS
   IS'' AND WITHOUT ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING,
   WITHOUT LIMITATION, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND
   FITNESS FOR A PARTICULAR PURPOSE.

Please NOTE: None of the TACACS code available here comes with any
warranty or support, however, comments or questions may be addressed
to Cisco systems via email at the address:

                customer-service@cisco.com  (or more simply)
                      cs@cisco.com

and we will do our best to handle them, though we cannot guarantee a
timely response, as this code is UNSUPPORTED. Be sure you've read this
user's guide, including the frequently asked questions include in it,
before posting.

Cisco systems also maintains an extensive World Wide Web site at

                http://www.cisco.com/

In addition, there are two mailing lists which may be of interest to
users of Tacacs+.

The first is a mailing list run by spot.Colorado.EDU which discusses
many things pertaining to Cisco products. It is not run by Cisco
Systems, Inc. and is not part of Cisco's formal service request
channels, however, many knowledgeable people, including staff members
of Cisco Systems, Inc. voluntarily read and respond on the list.

Requests to be added to or deleted from the list at spot.Colorado.EDU,
along with other administrative issues concerning it can be sent to:

        cisco-request@spot.Colorado.EDU

There is also a relatively new list called TACPLUS-L, run by
disaster.com, created for the purpose of information exchange between
TACACS+ Users. It is intended as a supplement to the list at
spot.Colorado.EDU, aiding TACACS+ users and prospective users in many
issues including but not limited to technical support, bug reports and
workarounds, configuration information, recommendations for future
versions of TACACS+, and general talk about TACACS+ development,
implementation, administration, etc.

Please note that neither of these lists is in fact connected with
Cisco Systems, Inc. or any of its subsidiaries. Standard etiquette
rules apply.

To subscribe to the TACPLUS-L list, send a message to

        tacplus-l-request@disaster.com

In the body of the letter, enter

        SUBSCRIBE TACPLUS-L your Name

to be automatically added, or visit their web page at
http://www.disaster.com/tacplus/.

Also, Robert Kiessling maintains a TACACS+ FAQ at
http://www.easynet.de/tacacs-faq.

Lastly, I am always interested in seeing contributed patches, so
consider mailing any modifications you make, as context diffs (be sure
to indicate with the version your patches are based on), to
tacacs-patches@cisco.com. As always, no support is implied, nor any
assurance that patches will be made available via ftp (though that is
my intent) or incorporated into any code.

Definitions and Terms
---------------------

NAS --- A Network Access Server e.g. a Cisco box, or any other
        *client* which makes tacacs+ authentication and authorization
        requests, or generates Tacacs+ accounting packets.

Daemon -- A program which services network requests for authentication
        and authorization, verifies identities, grants or denies
        authorizations, and logs accounting records.

passwd(5) files -- files conforming to Unix password style format, as
        documented in section 5 of the Unix manuals.

AV pairs -- strings of text in the form "attribute=value" sent between a
        NAS and a tacacs+ daemon as part of the Tacacs+ protocol.

Since a NAS is sometimes referred to as a server, and a daemon is also
often referred to as a server, the term "server" has been avoided here
in favor of the less ambiguous terms "NAS" and "Daemon".

TACACS, XTACACS and TACACS+
---------------------------

Note that there are now at least 3 versions of authentication protocol
that people commonly refer to as "TACACS".

The first is ordinary tacacs, which was the first one offered on Cisco
boxes and has been in use for many years. The second is an extension
to the first, commonly called Extended Tacacs or XTACACS, introduced
in 1990.

The third one is TACACS+ (or T+ or tac_plus) which is what is documented
here. TACACS+ is NOT COMPATIBLE with any previous versions of tacacs.

In addition to the 3 versions of tacacs running on Cisco boxes, the
fact that we distribute the source code to the daemon has meant that
additional implementations of tacacs daemons have been produced by
people who have made modifications to our source code.

BUILDING TAC_PLUS
-----------------
Tac_plus is known to build and run on the following platforms:

AIXv3.2 (using -DAIX and bsdcc. See tac_plus.h for more details).
HP/UX A.09.01 using -DHPUX
i86 Solaris 2.4 (SUNOS 5.4), using SUNpro SW2.0.1 & -DSOLARIS
Sun4 Solaris 2.4 using SUNpro SC3.0.1 and -DSOLARIS
SUNOS 4.1.2 sparc-2
SUNOS 4.1.3
MIPS R3K SGI IRIX 4.05F (using -DMIPS)
BSDI BSD/386 1.1 (using -DBSDI)
FREEBSD 2.0-RELEASE (using -DFREEBSD)
LINUX 1.2.8 (using -DLINUX)

To build tac_plus, untar the tarfile distribution into a clean
directory e.g.

        tar -xvf tac_plus.tar

Edit the top of the Makefile to select the appropriate defaults
for your system. Then type

        make tac_plus

The default version can authenticate using its internal database,
s/key or passwd(5) style files. Authorization is done via the
internal database or via calls to external programs, which the
administrator configures.

To use S/KEY, you must obtain and build the s/key library (libskey.a)
yourself.  You can then compile in S/KEY support per the instructions
for S/KEY in the Makefile.  I got my S/KEY code originally from
crimelab.com but now it appears the only source is ftp.bellcore.com. I
suggest you try a web search for s/key source code.

Note: S/KEY is a trademark of Bell Communications Research (Bellcore).

Should you need them, there are routines for accessing password files
(getpwnam,setpwent,endpwent,setpwfile) in pw.c.

Lastly, you may also need to add lines to your /etc/services file to
get tacacs+ to work correctly e.g. something along the lines of:

tacacs      tcp/49

You'll need to consult your system manuals for exact details of how to
do this.

A NOTE ABOUT ARAP, MSCHAP AND DES
---------------------------------
If you have access to a DES library which implements the calls:

int des_init();
void des_setkey();
void des_endes();
void des_done();

then you can define DES in tac_plus.h and link tac_plus with your DES
library. This is recommended, as it will allow you to process ARAP and MSCHAP
requests on your daemon, which is more efficient, and also more secure
than processing them on the NAS.

If you don't have access to des (which is U.S. export controlled), you
can simply leave DES undefined in tac_plus.h. ARAP and MSCHAP
authentication will still work, but it will be slightly less
efficient, since the NAS will attempt to get the daemon to do the DES
calculation before falling back to the alternative of calculating DES
on the NAS.  It's also slightly less secure, because if someone
discovers your encryption key, they can then download ARAP and MSCHAP
secrets from your daemon.

Note that this issue arises solely because U.S. government regulations
currently make it difficult to export the source code for DES outside
the US and Canada, which is why it is not included in this
distribution.

Lastly, this limitation of MSCHAP, ARAP and DES has no bearing on the
use of des passwords for regular logins. Regular logins also use DES
but they do it via the "crypt" system call, which is usually found in
a library on the Unix host where you compile your daemon.

There are additional restrictions on doing MSCHAP (see the FAQ later
in this document).

CONFIGURING TAC_PLUS
---------------------

Tac_plus is configured via a single configuration file. You can create
a configuration file from scratch or, if you have passwd(5) and
supplementary files from earlier versions of tacacs, you can convert
these to configuration file format by running the supplied perl script
convert.pl.

CONVERTING EXISTING PASSWD(5) FILES
-----------------------------------

To convert an existing passwd(5) file e.g. one used with an older
version of tacacs, use the convert.pl perl script as follows:

convert.pl <passwd file> [-g] [<supplementary file>]

1). If you have no supplementary file, simply omit it.

2). If the groupid field of your passwd file does NOT represent a
valid acl number (e.g if it's really a unix passwd file this field is
a group id, not an acl number), just omit the -g flag.

The rest of this document assumes that you are configuring tac_plus
from scratch.

CONFIGURING TAC_PLUS FROM SCRATCH
---------------------------------

A configuration file consists of some top-level directives for setting
defaults and for setting up the encryption key, followed by a
declaration for each user and group you want to configure.  Within
each user or group declaration, there are declarations for
authenticating and authorizing that user.

1). Configuring the encryption key

If you want tac_plus to encrypt its packets (and you almost certainly
*DO* want this, as there can be usernames and passwords contained in
these packets), then you must specify an encryption key in the
configuration file. The identical key must also be configured on any
NAS which communicates with tac_plus.

This is done using the statement

key = "your key here"

NOTE: You only need double quotes on the daemon if your key contains
spaces.

Confusingly, even if your key does contain spaces, you should NEVER
use double quotes when you configure the matching key on the NAS.

During debugging, it may be convenient to temporarily switch off
encryption by not specifying any key.  Be careful to remember to
switch encryption back on again after you've finished debugging.

The current code does not support host-specific keys (left as an
exercise to the reader).

On the NAS, you also need to configure the *same* key. Do this by
issuing:

    aaa new-model
    tacacs-server key <your key here>

COMMENTS IN CONFIGURATION FILES
-------------------------------
Comments can appear anywhere in the configuration file, starting with
the # character and extending to the end of the current line. Should
you need to disable this special meaning of the # character, e.g. if
you have a password containing a # character, simply enclose the string
containing it within double quotes.

CONFIGURING USERS AND GROUPS
----------------------------

Each user may belong to a group (but only one group).  Each group may
in turn belong to one other group and so on ad infinitum.

Users and groups are declared as follows. Here we declare two users
"fred" and "lily", and two groups, "admin" and "staff".

Fred is a member of group "admin", and group "admin" is in turn a
member of group "staff". Lily is not a member of any group.
\f
user = lily {
    # user lily is not a member of any group
    # and has nothing else configured as yet
}

user = fred {
    # fred is a member of group admin
    member = admin
}

group = admin {
    # group admin is a member of group staff
    member = staff
}

group = staff {
    # group staff is not a member of any group
}

RECURSION AND GROUPS
--------------------

In general, when the daemon looks up values e.g. passwords, it will
look first to see if the user has her own password. If not, it looks
to see if she belongs to a group and if so, whether the group has a
password defined. If not, this process continues through the hierarchy
of groups (a group can be a member of another group) until a value is
found, or there are no more groups.

This recursive process occurs for lookups of expiration dates, for
pap, arap and chap "secrets", and also for authorization parameters (see
later).

A typical configuration technique is thus to place users into groups
and specify as many groupwide characteristics in the group declaration
as possible. Then, individual user declarations can be used to
override the group settings for selected users as needed.

MEMBERSHIP IN MULTIPLE GROUPS
-----------------------------

Besides the descibed single-membership of user to some group, you may also find
useful if user (or host) belongs to multiple groups at once.  You can naturally
specify multiple "member = group_X" commands for such user:

user = fred {
    # fred is a member of both groups: admins_company_A, admins_company_B
    member = admins_company_A
    member = admins_company_B
}

group = admins_company_A {
    # group admins_company_A is not a member of any group
    member = admins_company_A_privilege_X
    member = admins_company_A_privilege_Y
}
group = admins_company_A_privilege_X {
}
group = admins_company_A_privilege_Y {
}

group = admins_company_B {
    # group admins_company_B is not a member of any group
}

Here it is important to respect the ordering of "member" commands: Any
searching for attributes/values is done by Depth-First Search - so Daemon would
first try to look all members of admins_company_A and THEN (after it would
failed to find any) it would start to searching through admins_company_B.  The
searching through the proposed example would be done in the following order:

        fred
        admins_company_A
        admins_company_A_privilege_X
        admins_company_A_privilege_Y
        admins_company_B

Sometimes you would want to only list some members (users, hosts or groups) but
you don't want to specify any attributes for them. You would be able to do it:

group = city_X_NASes {
}
host  first.NAS.X.city {
        member = city_X_NASes
}
host second.NAS.X.city {
        member = city_X_NASes
}

But you will probably find more comfortable to use "enlist" keyword. It has
the same functionality but it goes from the other way:

    member: current entity is connected as CHILD to the specified PARENT entity
    enlist: specified entity is connected as CHILD to the current one as PARENT

The example would be re-written using "enlist" keyword as:

group = city_X_NASes {
        enlist = host  first.NAS.X.city
        enlist = host second.NAS.X.city
}

As you can see, "enlist" doesn't require the existence of the given entity as
it would loose its primary purpose. If the entity doesn't exist it will be
automatically created (as empty one) - this doesn't apply to "member"! Any
argument to "member" MUST already exist. "enlist" is provided to save you from
writing a lot of empty definition lines like:

host =  first.NAS.X.city {
}

All forward references are not a problem, you can still make membership or
enlistment on the top of the file with the entity which will be defined on the
end of the configuration file.

CONFIGURING USER AUTHENTICATION
-------------------------------

User Authentication can be specified separately for PAP, ARAP, CHAP,
and normal logins.  In addition, a user global authentication method
can be given that will be used if a per-protocol method is not
specified.

PAP, ARAP, CHAP, and global user authentication must be given in clear
text.

The following assigns the user mary five different passwords for ARAP,
inbound and outbound CHAP, inbound PAP, outbound PAP, and normal login
respectively:

    user = mary {
        arap = cleartext "arap password"
        chap = cleartext "chap password"
        pap  = cleartext "inbound pap password"
        opap = cleartext "outbound pap password"
        login = des XQj4892fjk
    }


The following assigns the user agnes a single password for all the
above types of login (except outbound PAP):

    user = agnes {
        global = cleartext "Agnes global password"
    }

NOTE: you cannot use a global user password for outbound PAP. This is
because outbound PAP is implemented by sending the password from the
daemon to the NAS. This is a security issue if the TACACS+ key is ever
compromised.

There are 3 ways to authenticate a user for login.

1). You can include a DES (or cleartext) password for a user or for a
group that s/he is a member of, viz:

    user = joe {
        member = admin
        # this is lily's encrypted DES password. It overrides the admin
        # group's password
        login = des XQkR21zMB0TDU
    }

    user = fred {
        # fred is a member of group admin. He inherits the group's password
        # as he does not have his one of his own.
        member = admin
    }

    group = admin {
        # group admin has a cleartext password which all members share
        # unless they have their own password defined
        login = cleartext foobar
    }

If no password is needed for this user, this can be accomplished with
the 'nopassword' option:

    user = foo {
       login = nopassword
    }

NOTE: The C program built from generate_passwd.c may be used to
hand-generate encrypted passwords, or they may be taken from a Unix
passwd (or shadow) file.

2). Authentication using passwd(5) files.

For selected users, you can perform DES authentication using existing
passwd(5) files instead of entering the password into the
configuration file directly (though using passwd(5) files is
noticeably less efficient for large files).

You can specify this behavior per-user, by naming a passwd(5) file in
the password declaration (instead of giving a DES password), as
follows:

    user = fred {
       # look in file /etc/tac_plus_passwords to authenticate this user
       login = file /etc/tac_plus_passwords
    }

3). Authentication using s/key.

If you have successfully built and linked in a suitable s/key library
and compiled tac_plus to use s/key, you can then specify that a user
be authenticated via s/key, as follows:

    user = fred {
      login = skey
    }

RECURSIVE PASSWORD LOOKUPS
---------------------------

As stated earlier, authentication passwords are looked up recursively:
The daemon looks first to see if the user has her own password. If
not, it looks to see if she belongs to a group which has a
password. This process recurses through the hierarchy of groups (a
group can be a member of another group) until a password is found, or
there are no more groups.

CONFIGURING DEFAULT AUTHENTICATION
-----------------------------------
By default, an unrecognized user will be denied authentication (NOTE:
there is no way to authenticate someone with no username).

At the top level of the configuration file, you can set the default
authentication to use a passwd(5) file, viz:

    default authentication = file /etc/passwd

The effect of this statement is that if a user does not appear in the
configuration file, the daemon will attempt to authenticate the user
using passwords from this file i.e. /etc/passwd in this example.

If you have passwd(5) files from previous versions of tacacs daemons,
this facility allows you to authenticate using the passwd(5) from
older versions of tacacs, while you migrate to using the new
configuration file.

CONFIGURING EXPIRY DATES
------------------------
An entry of the form:

user = lol {
    expires = "MMM DD YYYY"
    password = cleartext "bite me"
}

will cause the user's passwords to become invalid, starting on the
expiry date. The only valid date format is e.g. "Jan 1 1980". Case is
NOT significant.

A expiry warning message is sent to the user when she logs in,
starting at 14 days before the expiration date.

On expiry, the administrator must re-set the expiry date in the
configuration file in order to grant continued access. Expiry applies
to all password types except "file" passwords.

If passwd(5) files are being used for authentication, the "expires"
field in the configuration file is not consulted. Instead, the daemon
looks at the the "shell" field of the password file entry for a valid
expiry date.

If Solaris shadow password files are used for authentication, the
"expires" field in the configuration file is not consulted. The expiry
field from the shadow password file (if it exists) is used as the
expiration date.

CONFIGURING AUTHENTICATION ON THE NAS
-------------------------------------

On the NAS, to configure login authentication on all lines (including
vty and console lines)

    aaa new-model
    aaa authentication login default tacacs+

NOTE: As soon as you issue this command, you will no longer be able to
create new logins to your NAS without a functioning tacacs+ daemon
appropriately configured with usernames and password, so make sure you
have this ready.

As a safety measure while setting up, we suggest you configure an
enable secret and make it the last resort authentication method, so
if your tacacs+ daemon fails to respond you will be able to use the
NAS enable password to login. To do this, configure:

    enable secret foo
    aaa authentication login default tacacs+ enable

If all else fails, and you find yourself locked out of the NAS due to
a configuration problem, the section on "recovering from lost
passwords" on Cisco's CCO web page will help you dig your way out.

CONFIGURING ENABLE PASSWORDS
----------------------------

The default privilege level for an ordinary user on the NAS is usually
1. When a user enables, she can reset this level to a value between 0
and 15 by using the NAS "enable" command. If she doesn't specify a
level, the default level she enables to is 15.

You can enable via tacacs+ e.g. by configuring on the NAS:

        aaa authentication enable default tacacs+

then whenever you attempt to enable, an authentication request is sent
with the special username $enab<n>$ where <n> is the privilege level
you are attempting to enable to.

(Note: in order to be compatible with earlier versions of tacacs, when
the requested enable level is 15, the daemon will also try the
username $enable$ before trying username $enab15$).

For example, with the above declaration, in order to enable on the
NAS, you need a user declaration like this one, on the daemon:

user = $enab15$ {
    login = cleartext "the enable password for level 15"
}

Note: Be aware that this does have the side effect that you now have a
user named $enab15$ who can then login to your NAS if she knows the
enable password.

Here is a similar declaration allowing users to enable to level 4:

user = $enab4$ {
    login = des bsoF4OivQCY8Q
}
\f
CONFIGURING AUTHORIZATION
-------------------------

Authorization must be configured on both the NAS and the daemon to
operate correctly. By default, the NAS will allow everything until you
configure it to make authorization requests to the daemon.

On the daemon, the opposite is true: The daemon will, by default, deny
authorization of anything that isn't explicitly permitted.

Authorization allows the daemon to deny commands and services
outright, or to modify commands and services on a per-user
basis. Authorization on the daemon is divided into two separate parts:
commands and services.

AUTHORIZING COMMANDS
--------------------

Exec commands are those commands which are typed at a Cisco exec
prompt. When authorization is requested by the NAS, the entire command
is sent to the tac_plus daemon for authorization.

Command authorization is configured by specifying a list of
egrep-style regular expressions to match command arguments (see the
supplied man page, regexp.3, for a full description of regular
expressions) and an action which is "deny" or "permit".

The following configuration example permits user Fred to run the
following commands:

    telnet 131.108.13.<any number> and
    telnet 128.<any number>.12.3 and
    show <anything>

All other commands are denied (by default).

user=fred {

    cmd = telnet {
        # permit specified telnets
        permit 131\.108\.13\.[0-9]+
        permit 128\.[0-9]+\.12\.3
    }
    cmd = show {
        # permit show commands
        permit .*
    }
}

NOTE: If an argument list you specify contains spaces or tabs, you
must enclose it in double quotes.

The command and arguments which the user types gets matched to the
regular expressions you specify in the configuration file (in order of
appearance).  The first successful match performs the associated
action (permit or deny). If there is no match, the command is denied
by default.

Conversely, the following configuration example can be used to deny
the command:

    telnet 131.108.13.<any number>

and permit all other arguments, since the last line will match any
argument list. All other commands and services are permitted due to
the "default service = permit" clause.

Note: the default statement must be the first in the user clause

user=fred {
    default service = permit
    cmd = telnet {
        # allow all fred's telnet commands except to 131.108.13.*
        deny 131\.108\.13\.[0-9]+
        permit .*
    }
}

Note: Matches are not anchored, so "deny 131.108.13.[0-9]+" matches
anywhere in the command. To anchor the match, use ^ at the beginning
of the regular expression.

Note: When a command has multiple arguments, users may enter them in
many different permutations. It can be cumbersome to create regular
expressions which will reliably authorize commands under these
conditions, so administrators may wish to consider other methods of
performing authorization e.g. by configuring NAS-local privileged
enable levels on the NAS itself.

COMMAND EXPANSION
-----------------

For command authorization, the Cisco NAS expands all commands to their
full names e.g. when you type "config t" on the NAS, this will be
expanded to "configuration terminal" before being sent to the daemon
so that you don't need to list all the possible contractions of a
command.

CONFIGURING DEFAULT AUTHORIZATION
---------------------------------

There are 3 places where defaults for authorization may be
configured. Unless specified to the contrary, the default is always to
deny authorization.

1). To override the default denial of authorization for users who are
not explicitly listed in the configuration file, the ersatz user
DEFAULT, if defined, can be used for authorizing such users, viz:

default authentication = file /etc/passwd

user = DEFAULT {
    service = ppp protocol = ip {
        addr-pool=foobar
    }
}

In this example, users who do not appear elsewhere will be
authenticated via the /etc/passwd file, and authorized by the contents
of the user = DEFAULT entry.

Note: For backward compatibility, the directive,

        default authorization = permit

may still be specifed at the top level of the configuration file. This
overrides the default denial of authorization for users who are not
explicitly listed in the configuration file, permitting all
authorization requests for such users.

2). At the user level i.e. inside the braces of a user declaration,
the default for a user who doesn't have a service or command
explicitly authorized is to deny that service or command.  The
following directive will permit the service or command by default
instead:

user = lol {
    default service = permit
}

NOTE: This directive must appear first inside the user declaration.

3). At the service authorization level i.e. inside the braces of a
service declaration, arguments in an authorization request are
processed according to the algorithm described later. Some actions
when authorizing services (e.g. when matching attributes are not
found) depend on how the default is configured. The following
declaration changes the default from deny to permit for this user and
service.

user = lol {
    service = exec {
        default attribute = permit
    }
}

NOTE: This directive must appear before any others inside the service
declaration.

NOTE: for command authorization (as opposed to service authorization
being discussed here), you specify deny .* or permit .* as the last
line of the regular expression matches to create default behavior.

AUTHORIZING EXEC STARTUP
-------------------------

If you authorize some exec commands, you implicitly agree to allow
that user to start an exec (it doesn't make sense to permit exec
commands if an exec can't be started to run those commands)

In addition to agreeing to allow an exec to start, you can supply some
parameters whenever an exec starts e.g. an autocommand, a dialback
string or a connection access list (acl).

In the example below, when an exec is started on the NAS, an acl of 4
will be returned to the NAS:

user=fred {

    # this following line permits an exec to start and permits
    # all commands and services by default

    default service = permit

    service = exec {
        # When an exec is started, its connection access list will be 4.
        # It also has an autocmd.
        acl = 4
        autocmd = "telnet foobar"
    }

    cmd = telnet {
        # allow all fred's telnet commands except telnet to 131.108.13.*
        deny 131\.108\.13\.[0-9]+
        permit .*
    }
}

NOTE: specifying an autocommand, or any other exec services, is part
of EXEC AUTHORIZATION. For it to work, you must also configure exec
authorization on your NAS e.g.

        aaa authorization exec tacacs+


AUTHORIZING EXEC, SLIP, PPP and ARAP SERVICES
----------------------------------------------

Authorizing exec, slip, PPP and arap services is done quite
differently from command authorization.

When authorizing these services, the NAS sends a request containing a
number of attribute-value (AV) pairs, each having the form

        attribute=value

(Note: during debugging, you may see AV pairs whose separator
character is a "*" instead of a "=" sign. This is to signify that the
value in a pair is optional. An "=" sign indicates a mandatory
value. A "*" denotes an optional value).

e.g. a user starting ppp/ip using an address of 131.108.12.44 would
generate a request with the following AV pairs:

        service=ppp
        protocol=ip
        addr*131.108.12.44

You can use the NAS debugging command

        debug aaa authorization

to see what authorization AV pairs are being used by the NAS. Note: If
you are not on the router console, you will also need to issue a
'terminal monitor' command to see debug output.

THE AUTHORIZATION PROCESS
-------------------------

Authorizing a single session can result in multiple requests being
sent to the daemon.  For example, in order to authorize a dialin ppp
user for IP, the following authorization requests will be made from
the NAS:

1). An initial authorization request to startup ppp from the exec,
using the AV pairs service=ppp, protocol=ip, will be made (Note: this
initial request will be omitted if you are autoselecting ppp, since
you won't know the username yet).

This request is really done to find the address for dumb PPP (or SLIP)
clients who can't do address negotiation. Instead, they expect you to
tell them what address to use before PPP starts up, via a text message
e.g. "Entering PPP. Your address is 1.2.3.4". They rely on parsing
this address from the message to know their address.

2). Next, an authorization request is made from the PPP subsystem to
see if ppp's LCP layer is authorized. LCP parameters can be set at
this time (e.g. callback). This request contains the AV pairs
service=ppp, protocol=lcp.

3). Next an authorization request to startup ppp's IPCP layer is made
using the AV pairs service=ppp, protocol=ipcp. Any parameters returned
by the daemon are cached.

4). Next, during PPP's address negotiation phase, each time the remote
peer requests a specific address, if that address isn't in the cache
obtained in step 3, a new authorization request is made to see if the
peers requested address is allowable.  This step can be repeated
multiple times until both sides agree on the remote peer's address or
until the NAS (or client) decide they're never going to agree and they
shut down PPP instead.

As you can see from the above, a program which plans to handle
authorization must be able to handle a variety of requests and respond
appropriately.

AUTHORIZATION RELIES ON AUTHENTICATION
--------------------------------------

Since we pretty much rely on having a username in authorization
requests to decide which addresses etc. to hand out, it is important
to know where the username for a PPP user comes from. There are
generally 2 possible sources

1). You force the user to authenticate by making her login to the exec
and you use that login name in authorization requests. This username
isn't propagated to PPP by default. To have this happen, you generally
need to configure the "if-needed" method, e.g.

aaa authentication login default tacacs+
aaa authentication ppp default if-needed


2). Alternatively, you can run an authentication protocol, PAP or CHAP
(CHAP is much preferred), to identify the user. You don't need an
explicit login step if you do this (so it's the only possibility if
you are using autoselect). This authentication gets done before you
see the first LCP authorization request of course.  Typically you
configure this by doing:

aaa authentication ppp default tacacs+
int async 1
ppp authentication chap

If you omit either of these authentication schemes, you will start to
see authorization requests in which the username is missing.

CONFIGURING SERVICE AUTHORIZATION
---------------------------------

A list of AV pairs is placed in the daemon's configuration file in
order to authorize services.  The daemon compares each NAS AV pair to
its configured AV pairs and either allows or denies the service. If
the service is allowed, the daemon may add, change or delete AV pairs
before returning them to the NAS, thereby restricting what the user is
permitted to do.

The complete algorithm by which the daemon processes its configured
AV pairs against the list the NAS sends, is given below.

The Authorization Algorithm
---------------------------

Find the user (or group) entry for this service (and protocol), then
for each AV pair sent from the NAS:

    If the AV pair from the NAS is mandatory:

        a). look for an exact attribute,value match in the user's
        mandatory list. If found, add the AV pair to the output.

        b). If an exact match doesn't exist, look in the user's
        optional list for the first attribute match. If found, add the
        NAS AV pair to the output.

        c). If no attribute match exists, deny the command if the
        default is to deny, or,

        d). If the default is permit, add the NAS AV pair to the
        output.

    If the AV pair from the NAS is optional:

        e). look for an exact attribute,value match in the user's
        mandatory list. If found, add DAEMON's AV pair to output.

        f). If not found, look for the first attribute match in the
        user's mandatory list. If found, add DAEMONS's AV pair to output.

        g). If no mandatory match exists, look for an exact
        attribute,value pair match among the daemon's optional AV
        pairs. If found add the DAEMON's matching AV pair to the
        output.

        h). If no exact match exists, locate the first attribute match
        among the daemon's optional AV pairs. If found add the
        DAEMON's matching AV pair to the output.

        i). If no match is found, delete the AV pair if the default is
        deny, or

        j). If the default is permit add the NAS AV pair to the output.

    k). After all AV pairs have been processed, for each mandatory
    DAEMON AV pair, if there is no attribute match already in the
    output list, add the AV pair (but add only ONE AV pair for each
    mandatory attribute).

RECURSIVE AUTHORIZATION
-----------------------

Remember that authorization is also recursive over groups, in the same
way that password lookups are recursive. Thus, if you place a user in
a group, the daemon will look in the group for authorization
parameters if it cannot find them in the user declaration.

EXAMPLES
--------

key = "your key here"

user=fred {
    login = des mEX027bHtzTlQ
    name = "Fred Flintstone"
    member = administrators
    expires = "May 23 2005"
    arap = cleartext "Fred's arap secret"
    chap = cleartext "Fred's chap secret"

    service = exec {
        # When Fred starts an exec, his connection access list is 5
        acl = 5

        # We require this autocmd to be done at startup
        autocmd = "telnet foo"
    }

   # All commands except show system are denied for Fred
    cmd = show {

        # Fred can run the following show command

        permit system
        deny .*
    }

    service = ppp protocol = ip {
        # Fred can run ip over ppp only if he uses one
        # of the following mandatory addresses. If he supplies no
        # address, the first one here will be mandated

        addr=131.108.12.11
        addr=131.108.12.12
        addr=131.108.12.13
        addr=131.108.12.14

        # Fred's mandatory input access list number is 101
        inacl=101

        # We will suggest an output access list of 102, but the NAS may
        # choose to ignore or override it

        optional outacl=102
    }

    service = slip {

        # Fred can run slip. When he does, he will have to use
        # these mandatory access lists

        inacl=101
        outacl=102
    }
}

user = wilma {

    # Wilma has no password of her own, but she's a group member so
    # she'll use the group password if there is one. Same for her
    # password expiry date

    member = admin
}

group = admin {

    # group members who don't have their own login password will be looked
    # up in /etc/passwd

    login = file /etc/passwd

    # group members who have no expiry date set will use this one

    expires = "Jan 1 1997"
}


CONFIGURATION RESPECTING NAS HOST OF THE USER
---------------------------------------------

Sometimes you would want to modify the configuration file according to the
source NAS where the user is being authenticated/authorized. For example if you
are big ISP you want to permit administrator of company X to be able to monitor
status of the links on NAS in company X but, of course, she should be able to
monitor any other links on any other NAS of the same ISP. As all the NASes are
authorized from the same Daemon, it seems as a problem. (You can workaround it
by using custom authorization program - see "USING PROGRAMS TO DO
AUTHORIZATION" section below - but it is not nice solution.)

For this purposes there exists another entity 'host':

                user wilma
                group admin
                host 198.133.219.25
                host nas.cisco.com

As you can see you may use either IP address or DNS name. Anyway, we strongly
recommend to always use only IP addresses - DNS subsystem may fail or it may be
forged by the enemy.

You have two methods of utilizing the differences between NASes:

1) Current user is always automatically enlisted (=given membership to) to its
   current NAS host. This looks weird as only groups can have members but this
   is the only exception to this rule, current NAS host can really have a
   member:

        ( user "current_user" |  user DEFAULT )
         |
         +-- host "current_NAS_IP_address"
         |    |
         |    +- group DEFAULT
         |
         +-- host "current_NAS_hostname"
         |    |
         |    +-- group DEFAULT
         |
         +-- group DEFAULT

   Each link only exists in the case it there exist both its peers, of course.
   user/group DEFAULT is written here only for the completeness of the chart.
   DEFAULT is discussed elsewhere in this documentation.

   According to the shown ordering, attributes/values in the host of current
   NAS identified by its IP address has _higher_ precence over the attributes
   in the current NAS identified by its (reverse-resolved) hostname.

   According to this auto-connections, you can for example permit some command
   to ALL the users on such NAS:

user = fred {
    login = cleartext LLLL
}
host = machine.A.company {
    cmd = write  {
        permit terminal
    }
}

   In this configuration file ALL the valid users can do "write terminal" when
   logged in on NAS "machine.A.company".

2) Sometimes you need to do the authorization specific only to some users on
   some NASes. For example to permit "write terminal" ONLY to user fred
   connected to NAS "machine.A.company". That means that you want to forbid it
   to user fred on any other NAS and yuo also want to forbid it to all the
   other users on NAS "machine.A.company". (Line "authorization = recursive" is
   required but read the following section "FULL RECURSIVITY" to know all its
   consequences.)

authorization = recursive
user = fred {
    login = cleartext LLLL
}
host = machine.A.company {
    when = user fred {
        cmd = write  {
            permit terminal
        }
    }
}

    This file has the same effect as:

authorization = recursive
user = fred {
   login = cleartext LLLL
   when = host machine.A.company {
        cmd = write  {
            permit terminal
        }
    }
}

   You can see the (nested) command "when" can limit the scope of existence of
   its contents. Definition of "host machine.A.company" with empty block (no
   attributes) isn't needed as "when" line will automatically create hosts not
   defined elsewhere, in the same style as "enlist" keyword creates them. Any
   "user"s or "group"s referenced by "when" MUST be defined, such entities are
   never created automatically!
   
   Unfortunately you cannot use "when" to limit any items, just a few of them
   are possible:

member
enlist
cmd
cmd arguments    (to limit specific "permit"/"deny" lines)
service          (or incl. "protocol" specification)
service AV pairs (to limit specific "attr=value" lines)
when             (enabling pure nesting of "when" blocks)

   Full flexibility to limit any contents may be done in future (needs complete
   cfgfile.c rewrite) but currently it is not supported. Fortunately you can
   get the same behaviour by appropriate usage of "member" keyword - all the
   attributes to be conditioned are put into separate group and you limit only
   the "member" keyword to such group.

   "when" command has the form "when = CONDITION { BLOCK }", CONDITION can be:

user  USR       i.e. that current user is "USR")
host  HOSTNAME  i.e. that current user is on NAS "HOSTNAME")
group GRP       i.e. current user belongs to the group "GRP",
                  All possible "when" conditions to reach such belonging
                  have to be successfuly met to make this condition successful.
                  Realize that "member" can be limited by "when" keyword.
CONDITION and CONDITION and CONDITION ...
CONDITION or  CONDITION or  CONDITION ...
not CONDITION
( CONDITION )

   You can see that you CANNOT use for example "user A and user B or user C",
   such condition would have ambiguous precedence of operators, you must
   explicitly write:  ( user A and user B ) or user C
                 or:  user A and ( user B or user C )

   Both parentheses have to be written as separate words, NOT "(user A)"
   but "( user A )" instead.
   (Proper solution would also need the complete cfgfile.c rewrite.)

   "not" operator has the highest precendence, so: not user A or user B
   has the same meaning as: ( not user A ) or user B

   Sometimes the "when" condition is so-called unresolvable. Example:

group = GRP { }
user = USR {
    when = group GRP {
        member = GRP
        }
}

   It is looping, when we would be the member of GRP, we would be really the
   member of GRP but otherwise sould wouldn't be the member of GRP. Weird?
   Yes, such case is considered as configuration bug, it is reported as:

        Unable to resolve expression from line 243, some looping occured

   Generally such unresolvable conditional expression is considered as UNKNOWN
   and the "when" keyword is then evaluated as "false" (=skip it's block).
   This MAY produce unwanted privilege access (if you were conditionally
   forbidding some privileges) so always watch out all the error messages in
   logs! (You should generally do default deny and specifically only "permit"
   all the privileges so the unintentional unresolvable expressions shouldn't
   hurt you in real.)

FULL RECURSIVITY
----------------

We have written in the previous examples line:

authorization = recursive

This changes some behaviour of Daemon:

1) Looping of memberships: By default any looped memberships are reported as
   error with message: recursively defined groups: ...

   After "authorization = recursive" any looping is silently accepted as it
   would be sometimes hard to prevent it in some complex configurations.
   Searching is done in normal Depth-First Search, but traversion is
   backtracked one step when the entity was already visited. Simply it will
   work 'magically', just safely ignore the fact.

2) "default service = default" is supported only with:
                authorization = recursive
   This isn't any real change, just some formality. In fact each entity block
   has "default service = default" in effect as default (in the case you don't
   write any "default service =" assign).

   "authorization = recursive" is enforced here due to the change of Daemon
   behaviour: In non-recursive (old) case the Daemon assumes "deny .*" as the
   last line of "cmd" block. This makes proper NAS-host based permissions
   impossible. In "authorization = recursive" mode it requires appropriate
   "permit" or "deny" line to apply - otherwise the Depth-First Search through
   the entities will continue to find another applicable "cmd" block (toplevel
   "default authorization" may get into effect as the last resort).

   "default service" keyword takes then another meaning: In non-recursive mode
   it will mean "if the command wasn't found". In fully-recursive mode it will
   mean "if any cmd block didn't decide".

   Please keep in mind that "service =" (with its possible "protocol =" mate)
   has always the same behaviour, nothing is changed with "authorization =
   recursive" enabled. Its functionality is already a bit complex (with AV
   pairs get erased, which added, which copied, which are optional from NAS,
   which optional from Daemon, whether to do default permit or deny etc.).
   Fortunately we didn't found that there would be needed some extended
   'recursiveness' functionality of "service" for application of NAS-host based
   authorization.

3) "host" entities are not recursive at all (=its attributes aren't looked up
   in its parent groups) without "authorization = recursive".  This isn't any
   much change as in the previous versions of Daemon "host" entity was either
   completely unsupported or it was supported only with the only one attribute:

   "key": Attribute defines different protocol key (instead of the specified
          toplevel one) for the specified host. Its use is recommended to
          improve overall network security (by using unique key for each NAS).

USING PROGRAMS TO DO AUTHORIZATION
----------------------------------

There are some limitations to the authorization that can be done using
a configuration file. The main ones are that you're constrained by the
algorithm the daemon uses, and that the configuration is basically
static, so if you're trying to use it to allocate dynamic things (such
as addresses from a pool) that vary over time, you need another
mechanism.

One solution is to arrange for the daemon to call your own
user-supplied programs to control authorization. These "callouts"
permit almost complete control over authorization, allowing you to
read all the fields in the authorization packet sent by the NAS
including all its AV pairs, and to set authorization status and send a
new set of AV pairs to the NAS in response.

USING AV PAIRS FOR AUTHORIZATION
--------------------------------

During authorization, the NAS sends an authorization request packet
containing various fields of interest and a set of AV pairs (see the
tacacs+ protocol specification for a list of fields and pairs).

Fields from the authorization packet can be supplied to the programs
you call on their command line, by using the appropriate dollar
variables in the configuration file (see below).

AV pairs from the authorization packet are fed to the program's
standard input, one per line. The program is expected to process the
AV pairs and write them to its standard output, one per line. What
happens then is determined by the exit status of the program.

NOTE: AV pairs are text strings with the format
attribute=value. Unlike the configuration file which allows spaces
when specifying AV pairs, there should be no spaces surrounding the
"=" sign when using the programmatic interface.

CALLING SCRIPTS BEFORE AUTHORIZATION
------------------------------------

You can specify a per-user program to be called before any other
attempt to authorize is made by using a "before" clause e.g.

user = auth1 {
    before authorization "/usr/bin/pre_authorize $user $port $address"
}

The AV pairs sent from the NAS will be supplied to this program's
standard input, one pair per line.

Fields from the initiating authorization packet which the NAS sends to
the daemon can also be passed to the program by using dollar variables
in the command line. A complete list of available variables is as
follows (consult the API specification for more details).

    user    -- user name
    name    -- Nas name
    port    -- Nas port
    address -- Nac address (remote user location)
    priv    -- privilege level (a digit, 0 to 15)
    method  -- (a digit, 1 to 4)
    type    -- (a digit, 1 to 4)
    service -- (a digit, 1 to 7)
    status  -- (pass, fail, error, unknown)

Unrecognized variables will appear as the string "unknown".

If the program returns a status of 0, authorization is unconditionally
permitted. No further processing is done on this request and no AV
pairs are returned to the NAS.

If the program returns a status of 1, authorization is unconditionally
denied. No further processing is done on this request and no AV pairs
are returned to the NAS.

If the program returns a status of 2, authorization is permitted.  The
program is expected to modify the AV pairs that it receives on its
standard input (or to create entirely new ones) and to write them, one
per line, to its standard output. The new AV pairs will be sent to the
NAS with a status of AUTHOR_STATUS_PASS_REPL.  No further processing
takes place on this request.

If the program returns a status of 3, authorization is denied, but all
attributes returned by the program via stdout are returned to the
NAS. Also, whatever the program returns on stderr is placed into the
server-msg field and returned to the NAS as well.

Any other status value returned from the program will cause an error
to be returned to the NAS.

Note that a status of 2 is not acceptable when doing command
authorization.

CALLING PROGRAMS AFTER AUTHORIZATION
------------------------------------

You can specify a per-user program to be called after authorization
processing has been carried out by the daemon (but before the
authorization status and AV pairs have been transmitted to the NAS).

The program can optionally modify the AV pairs being sent back to the
NAS and change the authorization status if required.

group = auth1 {
    # call /usr/bin/post_authorize passing it the username, port
    # and current authorization status.
    after authorization "/usr/bin/post_authorize $user $port $status"
}

The AV pairs resulting from the authorization algorithm that the
daemon proposes to return to the NAS, are supplied to the program on
standard input, one AV pair per line, so they can be modified if
required.

Fields from the incoming authorization packet which the NAS sent to
the daemon can also be passed to the program on its command line by
specifying dollar variables in the command line (see previous
section).

The program is expected to process the AV pairs and write them to its
standard output, one per line. What happens then is determined by the
exit status of the program:

If the program returns a status of 0, authorization continues as if
the program had never been called.  Use this if e.g. you just want a
program to send mail when an authorization occurs, without otherwise
affecting normal authorization.

If the program returns a status of 1, authorization is unconditionally
denied. No AV pairs are returned to the NAS. No further authorization
processing occurs on this request.

If the program returns a status of 2, authorization is permitted and
any AV pairs returned from the program on its standard output are sent
to the NAS in place of any AV pairs that the daemon may have
constructed.

Any other value will cause an error to be returned the the NAS by the
daemon.

WARNINGS AND CAUTIONS
---------------------

Customers attempting to write authorization scripts will find the NAS
debugging command "debug aaa authorization" invaluable.

Pre and post authorization programs are invoked by handing the command
line to the Bourne shell. On many Unix systems, if the shell doesn't
find the specified program it returns a status of one, which denies
authorization. However, at least one Unix system (BSDI) returns a
status code of 2 under these circumstances, which will permit
authorization, and probably isn't what you intended.

Note also that if your program hangs, the authorization will time out
and return an error on the NAS, and you'll tie up a process slot on
the daemon host, eventually running out of resources. There is no
special code to detect this in the daemon.

Unless you make special arrangements, the daemon will run as root and
hence the programs it invokes will also run as root, which is a
security weakness. It is strongly recommended that you use absolute
pathnames when specifying programs to execute, and that you use the
Makefile options TAC_PLUS_USERID and TAC_PLUS_GROUPID so that the
daemon is not running as root when calling these programs,

The daemon communicates with pre and post authorization programs over
a pair of pipes. Programs using the standard i/o library will use full
buffering in these circumstances. This shouldn't be a problem for most
programs, since they'll read AV pairs till they see end of file on
input, and they'll flush all output when they exit.

Note that when avpairs containing spaces are listed in the
configuration file, you need to enclose them in double quotes so that
they are parsed correctly. Avpairs which are returned via standard
output do not need delimiters and so should not be enclosed in double
quotes.

CONFIGURING AUTHORIZATION ON THE NAS
------------------------------------

If authorization is not explicitly configured on the NAS, no
authorization takes place i.e. effectively, everything is
permitted. Note that this is the converse of what happens on the
daemon, where anything not explicitly permitted is denied by default.

To configure command authorization on the NAS, issue the following NAS
configuration commands:

    aaa authorization commands 1 tacacs+
    aaa authorization commands 15 tacacs+


This will make the NAS send tacacs+ requests for all level 1 (ordinary
user) and level 15 (privileged level) commands on all lines/interfaces.

NOTE: As soon as you configure the above on your NAS, you will only be
permitted to execute NAS commands which are permitted by your tacacs+
daemon. So make sure you have configured, on the daemon, an
authenticated user who is authorized to run commands, or you will be
unable to do much on the NAS after turning on authorization.

Alternatively, or in addition, you may also want to configure the
following:

    aaa authorization commands 1 tacacs+ if-authenticated

This will use tacacs+ authorization for level 1 (user-level commands)
but if problems arise, you can just switch off the tacacs+ server and
authorization will then be granted to anyone who is authenticated.

The following daemon configuration should be sufficient to ensure that
you can always login as username "admin" (with a suitable password)
and run any command as that user:

user = admin {
    default service = permit
    login = des kppPfHq/j6gXs
}


ACCOUNTING
-----------

There is only one configurable accounting parameter -- the accounting
file name. All accounting records are written, as text, to this
filename. The filename is configured as follows at the top-level of
the configuration file:

accounting file = <filename>

Since accounting requests occur (and are serviced) asynchronously, it
is necessary to lock the accounting file so that two writers don't
simultaneously update it.  The daemon uses the fcntl call to do this
locking, so it is recommended that the accounting file reside on a
local filesystem. Although fcntl locking over NFS is supported on some
Unix implementations, it is notoriously unreliable. Even if your
implementation is reliable, locking is likely to be extremely
inefficient over NFS.

NAS CONFIGURATION
-----------------

To get accounting records equivalent to previous versions of tacacs,
the following is sufficient. "Stop" records contain elapsed time for
connections and exec sessions.

aaa accounting network stop-only tacacs+
aaa accounting exec stop-only tacacs+


CONFIGURING CALLBACK WITH TACACS+
---------------------------------

Note: Callback is available only in IOS 11.1 and later, and can only
be controlled via Tacacs+ for ASYNC lines. ISDN callback can be
configured on the NAS but cannot be controlled via AAA.

Here is an example of AAA configuration (with exec and network
accounting enabled):

NAS configuration:

aaa new-model
tacacs-server host XX.XX.XX.XX
tacacs-server key fookey
aaa accounting exec wait-start tacacs+
aaa accounting network wait-start tacacs+

! Example of AAA configuration for Exec:
aaa authentication login execcheck tacacs+
aaa authorization network tacacs+
service exec-callback
:
line 4
login authentication execcheck

! Example of AAA configuration for ARAP:
aaa authentication arap arapcheck tacacs+
aaa authorization network tacacs+
arap callback
:
line 4
arap authentication arapcheck

! Example of AAA-specific configuration for PPP callback:
aaa new-model
aaa authentication ppp pppcheck tacacs+
aaa authorization network tacacs+
:
int async 6
ppp authentication chap pppcheck
ppp callback accept

Daemon configuration:

Example of remote TACACS+ server CONFIG file entry for username `foobar':

user = foobar {
   arap = cleartext AAAA
   login = cleartext LLLL
   chap = cleartext CCCC
   pap = cleartext PPPP
   opap = cleartext OOOO
   service = ppp protocol = lcp {
        callback-dialstring=123456
   }
   service = arap {
        callback-dialstring=2345678
   }
   service = exec {
        callback-dialstring=3456789
        callback-line=7
        nocallback-verify=1
   }
}




DEBUGGING CONFIGURATION FILES
-----------------------------

When creating configuration files, it is convenient to check their
syntax using the -P flag to tac_plus e.g.

    tac_plus -P -C <config file name>

will syntax check the configuration file and print any error messages
on the terminal.

DEBUGGING A RUNNING SERVER
--------------------------

There is a myriad of debugging values that can be used in conjunction
with the -d flag to produce debugging output in /var/log/tac_plus.log.

For example, starting the daemon with

        tac_plus -C CONFIG -d 16

will put authentication debugging into /var/log/tac_plus.log. You can
view this information by using the tail command.

        tail -f /var/log/tac_plus.log

See the man page for more information.

CHANGING CONFIGURATIONS
-----------------------

To change a configuration file, you must edit the configuration file
and then send the daemon a SIGUSR1. This will cause it to reinitialize
itself and re-read the configuration file.

On startup, tac_plus creates the file /var/run/tac_plus.pid , if possible,
containing its process id. If you invoke the daemon so that it listens
on a non-standard port, the file created is /var/run/tac_plus.pid.<port>
instead, where <port> is the port number the daemon is listening on.

Assuming you are listening on the default port 49, something like the
following should work:

# kill -USR1 `cat /var/run/tac_plus.pid`

It's a good idea to check that the daemon is still running after
sending it a SIGUSR1, since a syntactically incorrect configuration
file will cause the daemon to die.

NOTE: The perl script generate_passwd.pl may be used to hand-generate
encrypted passwords, or they may be taken from a Unix passwd file.

FREQUENTLY ASKED QUESTIONS
--------------------------
Q). Does T+ required a working DNS?

A). As distributed, whenever a START packet arrive, the daemon makes a
call to getpeername to find out the name of the requestor. Depending
entirely on how your Unix host is set up, this may make DNS
queries. If this is a problem, comment out the code in tac_plus.c
which does this, and just use ip addresses instead of host names.

Q). Is the "name" field used for anything

A). No. It's purely for documentation. I had once thought it might be
useful when outputting error messages, and that you might need the
information if you converted a passwd(5) style file, but no use is
currently made of the field.

Q). Why do I get PPP authorization failures because of "no username in
request" when I've already logged in and authenticated?

A). With "aaa authentication PPP default tacacs+", the ppp
authentication overrides the earlier login authentication.  If the ppp
authentication fails, the username ends up blank.  Changing the config
to "aaa authentication ppp default if-needed tacacs+" fixes this
problem.

Q). I'm sure I configured it correctly, but accounting still doesn't
work.

A). You will find that although you can configure accounting in 10.3,
and it outputs debug messages, it doesn't send any accounting
packets. This is because Accounting only works from 11.0 onwards.

Q). Does TACACS+ use a database instead of a flat (/etc/passwd like)
file to decrease search times, say if we are talking about a large
database of 40,000 users?

A). The TACACS+ authentication database is held internally as a hash
table. This makes lookup times fast and fairly linear, at the expense
of making the server use potentially large amounts of memory space.
NOTE: If you specify that the server uses passwd(5) files for
authentication, then you don't get this speed benefit, but you save
space.

If you're willing to write the code, it should be a relatively simple
matter to interface the code to a database scheme e.g. unix dbm files,
or some proprietary database package, if you wish.

Q). Is there any way to avoid having clear text versions of the
ARAP and CHAP secrets in the configuration file?

CHAP and ARAP require that the server knows the cleartext password (or
equivalently, something from which the server can generate the
cleartext password). Note that this is part of the definition of CHAP
and ARAP, not just the whim of some Cisco engineer who drank too much
coffee late one night.

If we encrypted the CHAP and ARAP passwords in the database, then we'd
need to keep a key around so that the server can decrypt them when
CHAP or ARAP needs them.  So this only ends up being a slight
obfuscation and not much more secure than the original scheme.

In extended TACACS, the CHAP and ARAP secrets were separated from the
password file because the password file may be a system password file
and hence world readable.  But with TACACS+'s native database, there
is no such requirement, so we think the best solution is to
read-protect the files.  Note that this is the same problem that a
kerberos server has.  If your security is compromised on the kerberos
server, then your database is wide open.  Kerberos does encrypt the
database, but if you want your server to automatically restart, then
you end up having to "kstash" the key in a file anyway and you're back
to the same security problem.

So storing the cleartext password on the security server is really an
absolute requirement of the CHAP and ARAP protocols, not something
imposed by TACACS+.

We could have chosen a scheme where the NAS sends the challenge
information to the TACACS+ daemon and the daemon uses the cleartext
password to generate the response and returns that, but that means
that we must include specific protocol knowledge into the protocol for
both ARAP and CHAP and we would have to update the protocol every time
a new authentication protocol is added.  Hence we decided to go with
the SENDPASS mechanism.

Note that the above doesn't apply to PAP. You can keep an inbound PAP
password des-encrypted, since all you need to do with it is verify
that the password the principal gave you is correct.

Q). How is the typical login authentication sequence done?

A).     NAS sends START packet to daemon
        Daemon send GETUSER containing login prompt to NAS
        NAS prompts user for username
        NAS sends pkt to daemon
        Daemon sends GETPASS containing password prompt to the NAS
        NAS prompts user for password
        NAS sends pkt to daemon
        Daemon sends accept, reject or error to NAS


Q). Is there a GUI for the configuration file?
A). No. Use your favourite text editor.


Q). What does "default service = permit" really do?

A). When a request comes in to authorize exec startup, or ppp (with
protocol lcp, ip, ipx), or slip, or arap or a specific command, the
daemon looks for a matching declarations for the user (or groups the
user is a member of).

For exec startup, it looks for a "service=exec" OR any command
configured.

For ppp, it looks for a "service=ppp" and "protocol=(one of lcp, ip,
ipx)".

For slip there must be a "service=slip" and for arap a "service=arap"
clause.

For specific commands, there must be a matching cmd=<cmdname>.

If these aren't found, authorization will fail, *unless* you say
"default service = permit".

Q). How do I make PAP work?

A). Avoid using PAP if possible since it's not very secure. If you
*must* use it, PAP passwords may be specified along with arap and chap
passwords for each user. Note that the details of this changed in
version 3.0 and onwards.

For outbound PAP, where you are forced to send a password to the
remote host to identify yourself, there is now a separate "opap"
directive e.g.

        opap = cleartext OOOO

You use this to set the outbound PAP password. It must be a cleartext
password.

NOTE: It is very bad practice to use an outbound PAP password that is
the same as any of your inbound passwords. For this reason, a "global"
password does not apply to outbound PAP, only to inbound PAP,
bidirectional CHAP and ARAP.

Before 3.0, PAP logins were treated like ordinary user logins, so you
needed to declare a user in the Daemon configuration file whose name
was typically the remote hostname (or user), with a login password, in
order to process the PAP request.

Q). How can I deny some one from telneting from a commserver by ip
address only. i.e. when command is 10.0.1.6 rather than telnet
10.0.1.6.

A).  The best way to restrict telnet access is by applying an outbound
access list via the access class command (or equivalently, via the
"acl" avpair). The NAS configuration command "access-class <n> out"
for example applies a pre-defined standard IP access list (where n is
a number from 1 through 99) that governs telnet access from a NAS.

E.g. the following configuration commands permit outgoing Telnet
access from line 1 on the NAS *only* to hosts on network 192.85.55.0:

  access-list 12 permit 192.85.55.0 0.0.0.255
  line 1
  access-class 12 out

Note: you must define "access-list 12 permit 192.85.55.0 0.0.0.255" on
the NAS. Only then can you use the acl avpair to apply it to a line
that a user dials in on.

Alternatively, you can try configuring "transport preferred none" on
the lines in question. This will force a user to always type "telnet
10.0.1.6" in order to telnet out from the NAS.  Then you can apply
command authorization to this command to restrict it.

Q). I have an autocommand configured in the NAS-local database and I'm
using "aaa authentication local-override". The autocommand doesn't
work, but the username/password does. Why?

A). The "local-override" only applies to the authentication portion of
the local database, so if you want an autocommand for this user, you
need to also do:

        aaa authorization exec local if-authenticated

This will use the local DB entry if one exists, allow authenticated
users otherwise, or fail.

We don't have a "aaa authorization local-override" like we do for
authentication. Unlike authentication, the local method for
authorization is sort of equivalent to a local-override.

Q). Can tacacs+ only be enabled on a global basis? I want to
selectively turn it on for, e.g. only modem-connected lines. How do I
do this?

A). You turn tacacs+ ON on a global basis, but you can then change the
behavior of individual lines to whatever you want, e.g.

        aaa authentication login default tacacs+ none
        aaa authentication login oldstyle line
        aaa authentication login none none
        line 1 16
        login authentication default
        line vty 0 4
        login authentication oldstyle
        line 0
        login authentication none

Note that unfortunately, you can't (yet) apply authorization
differently to selected lines and interfaces.

Q). I have leased lines running PPP, and AAA authorization is also
configured, so the authorization on the leased lines fails.  What should
I do?

A). Since you can't (yet) configure authorization on a per-line basis,
you have to turn on authentication on the leased lines running PPP and
configure your T+ server so that it will authorize these lines
correctly.

A more demanding alternative is to modify the TACACS+ server source code
to allow any authorizations coming in from the port "SerialXXX" to
succeed.

A third possibility is to not use PPP on those lines, e.g. use HDLC
instead. HDLC doesn't require authentication or authorization.

Q). What are the memory recommendations for TACACS+?

A). Unless you're using passwd style files, TACACS+ holds entries in
hash tables in memory. The overhead is modest e.g. each user entry
occupies 72 bytes, plus space for strings like username and password
etc. Access time should thus be pretty constant regardless of number
of users. On a sparc 2, a config file containing 2000 users requires
about 0.5M of swap.

Q). How many users will a TACACS+ server support?  What happens when
the maximum is exceeded?

There are 2 issues. The first is that each entry in the config file
occupies memory (swap space) so you could run out of swap space either
starting up the daemon or when it forks to answer incoming requests
(see above). If this happens, the daemon will drop the connection
which should appear as an error to the NAS) and the logfile will
contain appropriate error messages. The solution is usually to add
more disk space for swapping.

The second issue is speed: Using config files containing 75,000 user
entries, I'm seeing about 3 authentications per second on a sparc 2
without noticeable performance impact, though I haven't benchmarked
this formally.

So more than about 3 authentications per second on this platform will
result in users seeing delays and having to wait for prompts.  The
usual solution to this is to add more daemons to spread the load out.

Q). How many characters may a TACACS+ Username and Password contain?

A). The short answer is 31 bytes of username, with up to 254 bytes of
password if they are cleartext (8 bytes if passwords are des
encrypted).

The long answer is that the Cisco NAS allocates a buffer of 1024
bytes, so this is the maximum you can type in, in response to a NAS
prompt.

But the protocol spec allows a username or password length field of
just one byte in an authentication packet, so only the first 255 of
these characters can be sent to the daemon.

Then, the API spec states that the username in the identity structure
on the daemon is 32 bytes long, so only the first 31 bytes of username
will be copied from the authentication packet into this structure,
which is then null-terminated.

The password, on the other hand, is copied into malloc'ed memory, so
it can still be up to 255 characters long.

Now if it's a des encrypted password, then only the first 8 bytes are
significant, per the common unix implementations of crypt.

Lastly, there is also the question of how long a username/password can
be configured in the daemon configuration file. The answer is given by
the value of MAX_INPUT_LINE_LEN, currently set to 255, which
determines the length of the longest string you can enter in the
configuration file.

Q). What is the format of accounting records?

Accounting records are text lines containing tab-separated fields. The
first 6 fields are always the same. These are:

timestamp, NAS name, username, port, address, record type.

Following these, a variable number of fields are written, depending on
the accounting record type. All are of the form attribute=value. There
will always be a task_id field.

Current attributes are:

"unknown"
"service"
"start_time"
"port"
"elapsed_time"
"status"
"priv_level"
"cmd"
"protocol"
"cmd-arg"
"bytes_in"
"bytes_out"
"paks_in"
"paks_out"
"address"
"task_id"
"callback-dialstring"
"nocallback-verify"
"callback-line"
"callback-rotary"

I expect more will be added over time.

Example records are thus:

Thu Jul 13 13:35:28 1995        cherub.cisco.com        chein   tty5    171.69.1.141    stop    task_id=12028   service=exec    port=5  elapsed_time=875
Thu Jul 13 13:37:04 1995        cherub.cisco.com        lol     tty18   171.69.1.129    stop    task_id=11613   service=exec    port=18 service=exec    port=18 elapsed_time=909
Thu Jul 13 14:09:02 1995        cherub.cisco.com        billw   tty18   171.69.1.152    start   task_id=17150   service=exec    port=18
Thu Jul 13 14:09:02 1995        cherub.cisco.com        billw   tty18   171.69.1.152    start   task_id=17150   service=exec    port=18 service=exec    port=18


Elapsed time is in seconds, and is the field most people are usually
interested in.

Q). How do I limit the number of sessions a user can have?

A). The TACACS+ daemon can enforce how many simultaneous sessions a
given user is allowed to have.  You must compile the daemon with the
MAXSESS symbol defined (see the Makefile).

Maximum sessions are configured on the daemon for a user as follows:

user = joeslip {
    login = cleartext XXX

    # only allow two sessions max for joeslip
    maxsess = 2

    name = "Joe SLIP User"
    ...
}

It can also be configured under a group:

group = slip_users {
    maxsess = 2
    ...
}
...
user = fred {
    ...
    member = slip_users
    ...
}

The daemon keeps a count of how many sessions a given user owns by
monitoring START and STOP accounting records.  Thus, exec and network
accounting must be configured for this feature to operate for exec and
ppp users.

As the restriction is enforced during the authorization phase of
login, exec and network authorization must be configured as well, viz:

aaa authentication login default tacacs+
aaa authentication ppp default tacacs+
aaa authorization exec tacacs+
aaa authorization network tacacs+
aaa accounting exec start-stop tacacs+
aaa accounting network start-stop tacacs+

Due to network outages (or other disruptions), it is possible for the
TACACS+ daemon's record of usage to become out of sync with reality,
so before denying access because it thinks a user is running too many
sessions, the TACACS+ daemon will use the finger service on the NAS to
verify how many sessions a user is running there.

If the result of finger indicates that the daemon should permit
access, access will be granted.  Note that for this check to work via
finger, "service finger" must also be configured on the NAS.

Lastly, note that because finger output truncates usernames at 10
characters, you may encounter trouble if you have users whose names
are not unique within those first 10 characters.
v
Also recall that authorization works differently on the console. So
many people locked themselves out of their boxes after configuring
authorization, that we stopped requiring authorization on the console
for authenticated users. Since there's no authorization on the
console, MAXSESS is not enforced there.

Q). How can I configure time-outs on an interface via Tacacs+?

A). Certain per-user/per-interface timeouts may be set by Tacacs+
during authorization. As of 11.0, you can set an arap session timeout,
and an exec timeout. As of 11.1 you can also set an exec idle timeout.

There are currently no settable timeouts for PPP or SLIP sessions, but
there is a workaround which applies to ASYNC PPP/SLIP idle timeouts
started via exec sessions only: This workaround is to set an EXEC
(idletime) timeout on an exec session which is later used to start up
PPP or SLIP (either via a T+ autocommand or via the user explicitly
invoking PPP or SLIP). In this case, the exec idle timeout will
correctly terminate an idle PPP or SLIP session. Note that this
workaround cannot be used for sessions which autoselect PPP or SLIP.

An idle timeout terminates a connection when the interface is idle for
a given period of time (this is equivalent to the "session-timeout"
Cisco IOS configuration directive). The other timeouts are
absolute. Of course, any timeouts set by Tacacs+ apply only to the
current connection.


user = lol {
    login = cleartext foobar
    service = exec {
        # disconnect lol if there is no traffic for 5 minutes
        idletime = 5
        # disconnect lol unconditionally after one hour
        timeout = 60
}

You also need to configure exec authorization on the NAS for the above
timeouts, e.g.

        aaa authorization exec tacacs+

Note that these timeouts only work for async lines, not for ISDN
currently.


Note also that you cannot use the authorization "if-authenticated"
option with these parameters, since that skips authorization if the
user has successfully authenticated.

Q). How do I send VPDN forwarding decisions to an authorization
server?

A). In 11.2 onwards, VPDN NASs can use T+ to allow an authorization
server to make the decision to forward users.

If VPDN forwarding is turned on, and the username is of the form
user@domain, and "domain" matches a vpdn outgoing configured domain,
then an authorization attempt is made for "domain" (see below).

When making an authorization call for VPDN, a service type of "ppp"
with a protocol type of "vpdn", with a username of "domain" will be
made e.g. when a PPP user comes up on a line with the username
foo@bar.com, if "vpdn enable" and "aaa authorization ...." is enabled
on the box, then a one-time authorization of the name "bar.com" is
attempted.

If this authorization is successful, no local authentication is
attempted on the NAS, and the connection is forwarded via VPDN
instead.

If no VPDN-specific information comes back from this authorization
call, the login proceeds as follows:

If tacacs-server directed-requests are configured (note: this is true
by default), then IOS will strip off the domain part of a name of the
form user@domain and use "domain" to try and select a T+ server. If
successful, the username portion "user", without the domain, will be
used for all subsequent authentication, authorization and accounting.

If directed requests are turned off, then the entire username
user@domain is treated as a username.

vpdn specific information includes the attributes "tunnel-id",
"source-ip" (deprecated) and "ip-addresses":

tunnel-id:
        This AV pair specifies the username that will be used to
        authenticate the tunnel over which the individual user MID
        will be projected.  This is analogous to the "NAS name" in the
        "vpdn outgoing" command.

ip-addresses:
        This is a list of possible IP addresses that can be used for
        the end-point of the tunnel. Consult the text at the end of
        this document for more details on how to configure this
        attribute.

source-ip: (This is now deprecated. It began in release 11.2(1.4),
            and was removed in 11.2(4.0.2)).
        This ip address will be used as the source of all VPDN packets
        generated as part of the VPDN tunnel (see the source-ip
        keyword in the vpdn outgoing command).

Tacacs+ syntax
--------------

The following syntax is used on the public domain Tacacs+ server.

username = domain {
    service = ppp protocol = vpdn {
      tunnel-id = <name for tunnel authentication>
      ip-addresses = <addr> [<addr> ...]
      source-ip = <ip-address>
    }
}

In addition the T+ server can be used to store the usernames for both
the NAS (the username specified by "tunnel-id" above) and the Home
Gateway.  These will be used to authenticate the tunnel.

Example:

user = foobar.cisco.com {
    service = ppp protocol = vpdn {
        tunnel-id = my_nas
        ip-addresses = "173.20.12.19 173.20.12.20"
        source-ip = 173.5.10.1
    }
}

user = my_nas {
    global = cleartext egad
}

user = my_home_gateway {
    global = cleartext wowser
}

IOS Configuration
-----------------

To enable AAA assisted VPDN forwarding on a NAS, the following minimum
configuration is required:

        vpdn enable
        aaa new-model
        aaa authorization network tacacs+ ...

In addition, if the passwords for the home gateway and NAS are stored
on the T+ daemon, the command

        aaa authentication login tacacs+ ....

should also be present.

Beginning with release 11.2(1.4), the additional configuration

        vpdn outgoing cisco.com ip NAS [ source-ip X.X.X.X ]

can be used. This changes in 11.2(4.0.2) and becomes:

        vpdn source-ip X.X.X.X
        vpdn outgoing cisco.com ip NAS


Q). Can someone expand on the use of the "optional" keyword.

A). Most attributes are mandatory i.e. if the daemon sends them to the
NAS, the NAS must obey them or deny the authorization. This is the
default. It is possible to mark attributes as optional, in which case
a NAS which cannot support the attribute is free to simply ignore it
without causing the authorization to fail.

This was intended to be useful in cutover situations where you have
multiple NASes running different versions of IOS, some of which
support more attributes than others. If you make the new attributes
optional, older NASes could ignore the optional attributes while new
NASes could apply them. Note that this weakens your security a little,
since you are no longer guaranteed that attributes are always applied
on successful authorization, so it should be used judiciously.

Q). Can I do do address pooling on the T+ daemon?

A). Not really since nothing on the daemon tracks whether an address
is already in use. The best way to do manage address pools is to
define a non-overlapping pool of addresses on each NAS and return the
name of this pool during authorization whenever a user logs in e.g.

NAS:

ip local pool foo 1.0.0.1 1.0.0.10

Daemon:

user = lol {
    service = ppp protocol = ip {
        addr-pool = foo
    }
}


Q). What about MSCHAP?

A). Version F4.X contains mschap support. Mschap is configured the
same way as chap, only using the "mschap" keyword in place of the
"chap" keyword.

Like ARAP, MSCHAP works best when you have a des library (see the note
at the beginning of this document), but this is optional, as the DES
calculation can be done by the NAS instead. It also optionally
requires a key from Microsoft which is not public domain, but this can
also be done on the NAS too, so you can live without it.

To compile the daemon with MSCHAP support, uncomment the MSCHAP line
in the Makefile. This will add the MSCHAP code to your daemon. You can
leave MSCHAP_DES undefined, in which case the MSCHAP DES calculation
will be done by NAS, and no special MSCHAP key will be required.

If you have a DES library and want to use it with MSCHAP (this is more
efficient for authentication), then you can also uncomment the
MSCHAP_DES and MSCHAP_MD4_SRC lines in the Makefile. This will ensure
that the DES calculation done in the daemon. A key will need to be
obtained from Microsoft and used to replace the definition of
MSCHAP_KEY in the file mschap.h. If you're thinking by now that this
is all too much trouble, I can't say I blame you....

Q). Can I do wtmp-style logging like xtacacd used to do?

A). Wtmp file logging is supported. The "-u <wtmpfilename>" command
line flag can be used to specify a filename which will be used for
logging wtmp style records instead of the regular T+ accounting
records.

wtmp records are generated into wtmpfilename. Since file locking is
also used when writing to wtmpfilename, the usual provisos regarding
NFS and locking (see accounting above) should be observed.

To generate correct wtmp log records, the NAS needs to be configured
as follows:

aaa accounting exec stop-only tacacs+
aaa accounting network stop-only tacacs+
aaa accounting system start-stop tacacs+

CANNED CONFIGURATIONS
---------------------

Here are some canned configurations for getting demos started:

1). A canned configuration for login authentication only. This allows
user fred to login with password "abcdef". If the tacacs+ server dies,
the enable secret will be accepted as a login password instead.

DAEMON:

key = <some key>

# repeat as necessary for each user
user = fred {
   login = cleartext abcdef
}

NAS:

aaa new-model
enable secret foobar
! use tacacs+. If server dies, use the enable secret password
aaa authentication login default tacacs+ enable
tacacs-server host <some host ip address>
tacacs-server key <some key>

2). A canned configuration for command authorization. This will allow
user fred to login with password abcdef and to run the privileged
(level 15) commands 'write terminal' and 'configure'. All other
privileged commands will be denied.

The "none" keyword in the NAS configuration line means that if the
tacacs+ server dies, any command will be allowed.

DAEMON:

key = <some key>

# repeat as necessary for each user
user = fred {
   login = cleartext abcdef
   cmd = write  {
        permit terminal
    }
   cmd = configure {
       permit .*
   }
}


NAS:

aaa new-model
! all level 15 (privileged commands). If server dies, allow everything
aaa authorization commands 15 tacacs+ none
tacacs-server host <some host ip address>
tacacs-server key <some key>

3). Canned configuration for network access authorization. This config
allows "fred" to login to line 1 with password abcdef (or to and to
run ppp using chap authentication. The chap password is "lab".

DAEMON:

key = <some key>

# repeat as necessary for each user
user = fred {
   login = cleartext abcdef
   chap = cleartext lab
   service = ppp protocol = ip {
        addr=1.0.0.2
    }
}

NAS:

aaa new-model
! authenticate exec logins (if not autoselecting)
aaa authentication login default tacacs+
! authorize network services via tacacs+
aaa authorization network tacacs+
! use tacacs+ for authenticating ppp users
aaa authentication ppp default tacacs+
tacacs-server host <some host ip address>
tacacs-server key <some key>
interface Async1
ip address 1.0.0.1 255.0.0.0
async default ip address 172.21.14.55
encapsulation ppp
async dynamic address
async mode interactive
! use chap to authenticate ppp users
ppp authentication chap
line 1
! need "modem inout" here and flow control if using a modem

4). Canned configuration for ARAP.

NAS:

aaa new-model
aaa authentication arap default guest tacacs+
aaa authorization network tacacs+
aaa accounting network start-stop tacacs+
!
appletalk routing
arap network <number> <name>
!
interface Ethernet0
 appletalk cable-range <range>
 appletalk zone <zonename>
!
tacacs-server host <host>
tacacs-server key <key>
!
line 1
 location a modem
 modem answer-timeout 0
 modem InOut
 autoselect arap
 autoselect during-login
 arap enable
 speed <speed>
 flowcontrol hardware

Daemon:

key = "some key"

user = lol {
    arap = cleartext <arap secret>
    service = arap { }
}

Authorization AV pairs
----------------------
The following authorization AV pairs are supported by 10.3(3) onwards
except where specifically noted.

The following AV pairs specify which service is being authorized. They
are typically accompanied by protocol AV pairs and other, additional
pairs from the lists below.

service=arap
service=shell (for exec startup, and also for command authorizations)
service=ppp
service=slip

service=system (not used).

service=raccess
        Used for managing reverse telnet connections e.g.

        user = jim {
            login = cleartext lab
            service = raccess {
                port#1 = nasname1/tty2
                port#2 = nasname2/tty5
            }
        }

        Requires IOS configuration

                aaa authorization reverse-access tacacs+

        See the IOS docs for more details.

protocol=lcp
        The lower layer of PPP, always brought up before IP, IPX, etc.
        is brought up.

protocol=ip
        Used with service=ppp and service=slip to indicate which
        protocol layer is being authorized.


protocol=ipx
        Used with service=ppp to indicate which protocol layer
        is being authorized.

protocol=atalk
        with service=ppp or service=arap

protocol=vines
        For vines over ppp.

protocol=ccp
        Authorization of CCP.  Compression Control Protocol). No other
        av-pairs associated with this.

protocol=cdp
        Authorization of CDP (Cisco Discovery Protocol). No other
        av-pairs associated with this.

protocol=multilink
        Authorization of multilink PPP. See 'max-links' and 'load-threshold'.

protocol=unknown
        For undefined/unsupported conditions. Should not occur under
        normal circumstances.

cmd (EXEC)
        If the value of cmd is NULL e.g. the AV pair is cmd=, then
        this is an authorization request for starting an exec.

        If cmd is non-null, this is a command authorization request,
        It contains the name of the command being authorized
        e.g. cmd=telnet

cmd-arg (EXEC)
        During command authorization, the name of the command is given
        by an accompanying "cmd=" AV pair, and each command argument
        is represented by a cmd-arg AV pair e.g. cmd-arg=archie.sura.net

        NOTE: 'cmd-arg' should never appear in a configuration file.
        It is used internally by the daemon to construct a string
        which is then matched against the regular expressions which appear
        in a cmd clause in the configuration file.

acl (ARAP, EXEC)
        For ARAP this contains an access-list number.  For EXEC
        authorization it contains an access-class number, e.g. acl=2.
        which is applied to the line as the output access class equivalent
        to the configuration command

                line <n>
                access-class 2 out

        An outbound access-class is the best way to restrict outgoing telnet
        connections. Note that a suitable access list (in this case,
        numbered 2) must be predefined on the NAS.

inacl (PPP/IP/IPX)
        This AV pair contains an IP or IPX input access list number
        for slip or PPP e.g. inacl=2. The access list itself must be
        pre-configured on the Cisco box. Per-user access lists do not
        work with ISDN interfaces unless you also configure a virtual
        interface. After 11.2(5.1)F, you can also use the name of a
        predefined named access list, instead of a number, for the
        value of this attribute.

        Note: For IPX, inacl is only valid after 11.2(4)F.

inacl#<n> (PPP/IP, PPP/IPX, 11.2(4)F)
        This AV pair contains the definition of an input access list
        to be installed and applied to an interface for the duration
        of the current connection, e.g.

        inacl#1="permit ip any any precedence immediate"
        inacl#2="deny igrp 0.0.1.2 255.255.0.0 any"

        Attributes are sorted numerically before they are applied.
        For IP, standard OR extended access list syntax may be used,
        but it is an error to mix the two within a given access-list.

        For IPX, only extended access list syntax may be used.

        See also:
        sho ip access-lists
        sho ip interface
        sho ipx access-lists
        sho ipx interface
        debug aaa author
        debug aaa per-user

outacl (PPP/IP, PPP/IPX)
        This AV pair contains an IP or IPX output access list number
        for SLIP. PPP/IP or PPP/IPX connections e.g. outacl=4. The
        access list itself must be pre-configured on the Cisco
        box. Per-user access lists do not work with ISDN interfaces
        unless you also configure a virtual interface.  PPP/IPX is
        supported in 11.1 onwards only. After 11.2(5.1)F, you can also
        use the name of a predefined named access list, as well as a
        number, for the value of this attribute.


outacl#<n> (PPP/IP, PPP/IPX, 11.2(4)F)
        This AV pair contains an output access list definition to be
        installed and applied to an interface for the duration of the
        current connection, e.g.

        outacl#1="permit ip any any precedence immediate"
        outacl#2="deny igrp 0.0.9.10 255.255.0.0 any"

        Attributes are sorted numerically before they are applied.
        For IP, standard OR extended access list syntax may be used,
        but it is an error to mix the two within a given access-list.

        For IPX, only extended access list syntax may be used.

        See also:
        sho ip access-lists
        sho ip interface
        sho ipx access-lists
        sho ipx interface
        debug aaa author
        debug aaa per-user

addr (SLIP, PPP/IP)
        The IP address the remote host should be assigned when a slip
        or PPP/IP connection is made e.g. addr=1.2.3.4

routing (SLIP, PPP/IP)
        Equivalent to the /routing flag in slip and ppp commands. Can
        have as its value the string "true" or "false".

timeout (11.0 onwards, ARAP, EXEC)
        Sets the time until an arap or exec session disconnects
        unconditionally (in minutes), e.g. timeout=60

autocmd (EXEC)
        During exec startup, this specifies an autocommand, like the
        autocommand option to the username configuration command,
        e.g. autocmd="telnet foo.com"

noescape (EXEC)
        During exec startup, this specifies "noescape", like the
        noescape option to the username configuration command.  Can
        have as its value the string "true" or "false",
        e.g. noescape=true

nohangup (EXEC)
        During exec startup, this specifies "nohangup", like the
        nohangup option to the username configuration command.  Can
        have as its value the string "true" or "false",
        e.g. nohangup=true

priv-lvl (EXEC)
        Specifies the current privilege level for command
        authorizations, a number from zero to 15 e.g. priv_lvl=5.

        Note: in 10.3 this attribute was priv_lvl i.e.
        it contained an underscore instead of a hyphen.

zonelist (ARAP)
        An Appletalk zonelist for arap equivalent to the line
        configuration command "arap zonelist" e.g. zonelist=5

addr-pool (11.0 onwards, PPP/IP, SLIP)
        This AV pair specifies the name of a local pool from which to
        get the IP address of the remote host.

        Note: addr-pool works in conjunction with local pooling.  It
        specifies the name of a local pool (which needs to be
        pre-configured on the NAS). Use the ip-local pool command to
        declare local pools, e.g on the NAS:

        ip address-pool local
        ip local pool foo 1.0.0.1 1.0.0.10
        ip local pool baz 2.0.0.1 2.0.0.20

        then you can use Tacacs+ to return addr-pool=foo or
        addr-pool=baz to indicate which address pool you want to get
        this remote node's address from, e.g. on the daemon:

        user = lol {
            service = ppp protocol = ip {
                addr-pool=foo
            }
        }

route (11.1 onwards, PPP/IP, SLIP).
        This AV pair specifies a route to be applied to an interface.

        During network authorization, the "route" attribute may be
        used to specify a per-user static route, to be installed via
        Tacacs+.

        The daemon side declaration is:

        service=ppp protocol=ip {
            route="<dst_addr> <mask> [ <gateway> ]"
        }

        This indicates a temporary static route that is to be
        applied. "<dst_address>, <mask> and [<gateway>]" are expected
        to be in the usual dotted-decimal notation, with meanings the
        same as for the familiar "ip route" configuration command on a
        NAS.

        If gateway is omitted, the peer's address is taken to be the gateway.

        The route is expunged once the connection terminates.

route#<n> (PPP/IP/IPX, 11.2(4)F)
        Same as the "route" attribute, except that these are valid for
        IPX as well as IP, and they are numbered, allowing multiple
        routes to be applied e.g.

        route#1="3.0.0.0 255.0.0.0 1.2.3.4"
        route#2="4.0.0.0 255.0.0.0"


        or, for IPX,

        route#1="4C000000 ff000000 30.12.3.4"
        route#2="5C000000 ff000000 30.12.3.5"

        See also:
        sho ip route
        sho ipx route
        debug aaa author
        debug aaa per-user

callback-rotary (11.1 onwards, valid for ARAP, EXEC, SLIP or PPP)
        The number of a rotary group (between 0 and 100 inclusive)
        to use for callback e.g. callback-rotary=34. Not valid for ISDN.

callback-dialstring (11.1 onwards, valid for ARAP, EXEC, SLIP or PPP)
        sets the telephone number for a callback e.g.
        callback-dialstring=408-555-1212. Not valid for ISDN.

callback-line (11.1 onwards, valid for ARAP, EXEC, SLIP or PPP)
        The number of a tty line to use for callback e.g.
        callback-line=4. Not valid for ISDN.

nocallback-verify (11.1 onwards, valid for ARAP, EXEC)
        Indicates that no callback verification is required. The only
        valid value for this parameter is the digit one i.e.
        nocallback-verify=1. Not valid for ISDN.

idletime (11.1 onwards, EXEC)
        Sets a value, in minutes, after which an IDLE session will be
        terminated. N.B. Does NOT work for PPP.

tunnel-id (11.2 onwards, PPP/VPDN)
        This AV pair specifies the username that will be used to
        authenticate the tunnel over which the individual user MID
        will be projected.  This is analogous to the "NAS name" in the
        "vpdn outgoing" command.

ip-addresses (11.2 onwards, PPP/VPDN)
        This is a space separated list of possible IP addresses that
        can be used for the end-point of the tunnel.

        In 11.2(5.4)F, this attribute was extended as follows:

        1) comma (',') is also consider as a delimiter
        For example the avpair can now be written as

        ip-addresses = 172.21.9.26,172.21.9.15,172.21.9.4

        2) '/' is considered a priority delimiter. When you have a
        number of Home Gateway routers, it is desirable to consider some
        as the primary routers and some as backup routers.

        The '/' allow you to config the routers into priority groups,
        so that the NAS will try to forward the users to the high
        priority routers, before forwarding to the low priority one.

        For example in the following avpair:

        ip-addresses = "172.21.9.26 / 172.21.9.15 / 172.21.9.4"

        172.21.9.26 is considered to be priority 1
        172.21.9.15 is considered to be priority 2
        172.21.9.4  is considered to be priority 3

        The NAS will try to forward the users to 172.21.9.26, before
        trying 172.21.9.15.  If the NAS can't forward users to
        172.21.9.26, it will try 172.21.9.15 next. If it fails with
        172.21.9.15, it will then try forwarding to 172.21.9.4.

source-ip (PPP/VPDN, now deprecated, only existed in releases 11.2(1.4)
        thru 11.2(4.0.2)). This specifies a single ip address will be
        used as the source of all VPDN packets generated as part of
        the VPDN tunnel (see the equivalent source-ip keyword in the
        IOS vpdn outgoing command).

nas-password (PPP/VPDN, 11.2(3.4)F, 11.2(4.0.2)F)
        During L2F tunnel authentication, nas-password specifies the password
        for the NAS.

gw-password (PPP/VPDN, 11.2(3.4)F, 11.2(4.0.2)F)
        During L2F tunnel authentication, gw-password specifies the password
        for the home gateway.

rte-ftr-in#<n> (PPP -- IP/IPX, 11.2(4)F)
        This AV pair specifies an input access list definition to be
        installed and applied to routing updates on the current
        interface, for the duration of the current connection.

        For IP, both standard and extended Cisco access list syntax is
        recognised, but it is an error to mix the two within a given
        access-list.

        For IPX, only Cisco extended access list syntax is legal.

        Attributes are sorted numerically before being applied.  For
        IP, the first attribute must contain the name of a routing
        process and its identifier (except for rip, where no
        identifier is needed), e.g.

        rte-fltr-in#0="router igrp 60"
        rte-fltr-in#1="permit 0.0.3.4 255.255.0.0"
        rte-fltr-in#2="deny any"

        For IPX, no routing process is needed, e.g.

        rte-fltr-in#1="deny 3C01.0000.0000.0001"
        rte-fltr-in#2="deny 4C01.0000.0000.0002"

        See also:

        show ip access-lists
        show ip protocols
        sho ipx access-lists
        sho ipx interface
        debug aaa author
        debug aaa per-user

        Related IOS commands:
        IP:
                router <routing process identifier>
                [no] distribute-list <list-name> in <interface>

        IPX:
                ipx input-network-filter <access-list-number>


rte-ftr-out#<n> (PPP/IP, 11.2(4)F)
        This AV pair specifies an input access list definition to be
        installed and applied to routing updates on the current
        interface, for the duration of the current connection.

        For IP, both standard and extended Cisco access list syntax is
        recognised, but it is an error to mix the two within a given
        access-list.

        Attributes are sorted numerically before being applied.  The
        first attribute must contain the name of a routing process and
        its identifier (except for rip, where no identifier is
        needed), e.g.

        rte-fltr-out#0="router igrp 60"
        rte-fltr-out#3="permit 0.0.5.6 255.255.0.0"
        rte-fltr-out#4="permit any"

        For IPX, no routing process is specified, e.g.

        rte-fltr-out#1="deny 3C01.0000.0000.0001"
        rte-fltr-out#2="deny 4C01.0000.0000.0002"

        See also:

        sho ipx access-lists
        sho ipx interface
        show ip access-lists
        show ip protocols
        debug aaa author
        debug aaa per-user

        Related IOS commands:
        IP:
                router <routing process identifier>
                [no] distribute-list <list-name> in <interface>

        IPX:
                ipx output-network-filter <access-list-number>


sap#<n> (11.2(4)F PPP/IPX)
        This AV pair specifies static saps to be installed for the duration
        of a connection e.g.

        sap#1="4 CE1-LAB 1234.0000.0000.0001 451 4"
        sap#2="5 CE3-LAB 2345.0000.0000.0001 452 5"

        The syntax of static saps is the same as that used by the IOS
        "ipx sap" command.

        See also:
        sho ipx servers
        debug aaa author
        debug aaa per-user

        Related IOS commands:
        [no] ipx sap ....


route#<n> (PPP/IP/IPX, 11.2(4)F)

        Same as the "route" attribute, except that these are valid for
        IPX as well as IP, and they are numbered, allowing multiple
        routes to be applied e.g.

        route#1="3.0.0.0 255.0.0.0 1.2.3.4"
        route#2="4.0.0.0 255.0.0.0"


        or, for IPX,

        route#1="4C000000 ff000000 30.12.3.4"
        route#2="5C000000 ff000000 30.12.3.5"

        See also:
        sho ip route
        sho ipx route
        debug aaa author
        debug aaa per-user


sap-fltr-in#<n> (PPP/IPX, 11.2(4)F)
        This AV pair specifies an input sap filter access list
        definition to be installed and applied on the current
        interface, for the duration of the current connection.

        Only Cisco extended access list syntax is legal, e.g

        sap-fltr-in#1="deny 6C01.0000.0000.0001"
        sap-fltr-in#2="permit -1"

        Attributes are sorted numerically before being applied.

        sho ipx access-lists
        sho ipx interface
        debug aaa author
        debug aaa per-user

        [no] ipx input-sap-filter <number>


sap-fltr-out#<n> (PPP/IPX 11.2(4)F)
        This AV pair specifies an output sap filter access list
        definition to be installed and applied on the current
        interface, for the duration of the current connection.

        Only Cisco extended access list syntax is legal, e.g

        sap-fltr-out#1="deny 6C01.0000.0000.0001"
        sap-fltr-out#2="permit -1"

        Attributes are sorted numerically before being applied.

        sho ipx access-lists
        sho ipx interface
        debug aaa author
        debug aaa per-user

        [no] ipx output-sap-filter <number>



pool-def#<n> (PPP/IP, 11.2(4)F)
        This attribute is used to define ip address pools on the NAS.

        During IPCP address negotiation, if an ip pool name is
        specified for a user (see the addr-pool attribute), a check is
        made to see if the named pool is defined on the NAS. If it is,
        the pool is consulted for an ip address.

        If the required pool is not present on the NAS (either in the
        local config, or as a result of a previous download
        operation), then an authorization call to obtain it is
        attempted, using the special username:

            <nas-name>-pools

        where <nas-name> is the configured name of the NAS.

        Note: This username can be changed using the IOS configuration
        directive e.g.

            aaa configuration config-name nas1-pools-definition.cisco.us

        The pool-def attribute is used to define ip address pools for
        the above authorization call e.g.

        user = foo {
            login = cleartext lab
            service = ppp protocol = ip {
                addr-pool=bbb
            }
        }

        user = nas1-pools {
             service = ppp protocol = ip {
                pool-def#1 = "aaa 1.0.0.1 1.0.0.3"
                pool-def#2 = "bbb 2.0.0.1 2.0.0.10"
                pool-def#3 = "ccc 3.0.0.1 3.0.0.20"
                pool-timeout=60
             }
        }


        In the example above is a configuration file fragment for defining 3
        pools named "aaa", "bbb" and "ccc" on the NAS named "nas1".

        When the user "foo" refers to the pool named "bbb", if the
        pool "bbb" isn't defined, the NAS will attempt to download the
        definition contained in the "nas1-pools" entry.

        The other pools will also be defined at the same time (or they
        will be ignored if they are already defined).

        Since this procedure is only activated when an undefined pool
        is referenced, one way to redefine a pool once it has been
        downloaded is to manually delete the definition e.g. by
        logging into the NAS, enabling, and configuring:

        config t
        no ip local pool bbb
        ^Z

        When a pool is deleted, there is no interruption in service
        for any user who is currently using a pool address. If a pool
        is deleted and then subsequently redefined to include a pool
        address that was previously allocated, the new pool will pick
        up the allocated address and track it as expected.

        Since downloaded pools do not appear in the NAS configuration,
        any downloaded pool definitions automatically disappear
        whenever a NAS reboots. These pools are marked as "dynamic"
        when they appear in the output of the "show ip local pools"
        NAS command.

        Since it is desirable not to have to manually delete pools to
        redefine them, the AV pair pool-timeout=<n> can be used to
        timeout any downloaded pool definitions. The timeout <n> is in
        minutes.

        The effect of the pool-timeout attribute is to start a timer
        when the pool definitions are downloaded.  When the timer
        expires, the pools are deleted. The next reference to a
        deleted pool via will cause a re-fetch of the pool
        definitions.  This allows pool changes to be made on the
        daemon and propagated to the NAS in a timely manner.

        See also:

        sho ip local pool
        sho ip local pool <pool-name>

        IOS commands:

        ip local pool <name> <start address> <end address>


old-prompts (PPP/SLIP)
        This attribute is integrated into the following IOS images:

        11.2(7.4)P 11.2(7.4) 11.1(13.1) 11.(16.2) 11.1(13.1)AA
        11.1(13.1)CA 11.1(13.1)IA 11.2(8.0.1)F 11.0(16.2)BT)

        This attribute allows providers to make the prompts in T+
        appear identical to those of earlier systems (tacacs and
        xtacacs). This will allow administrators to upgrade from
        tacacs/xtacacs to T+ transparently to users.

        The difference between the prompts is as follows:

        In xtacacs, when the user types "slip" or "ppp" the system
        prompts for an address followed by a password, whereas T+
        prompts only for an address.

        In xtacacs, if the user types "slip host" or "ppp host", the
        system prompts for a password. In T+, there is no prompt.

        Using this attribute, T+ can be made to mimic the prompting
        behaviour of xtacacs, by configuring network authorization on
        IOS, and using the "old-prompts=true" attribute value pair for
        slip and ppp/ip, viz:

        user = joe {
            global = cleartext foo

            service = exec {
            }
            service = slip {
                default attribute = permit
                old-prompts=true
            }
            service = ppp protocol = ip {
                default attribute = permit
                old-prompts=true
            }
        }

        i.e. the prompts are controlled by the addition of the
        "old-prompts=true" attribute.


max-links (PPP/multilink - Multilink parameter; 11.3)
        This AV pair restricts the number of multilink bundle links
        that a user can have.

        The daemon side declaration is:

        service=ppp protocol=multilink {
            max-links=<n>
        }

        The range of <n> is [1-255].

        Related NAS commands:
        int <foo>
          [no] ppp multilink
        int virtual-template X
          multilink max-links <n>

        show ppp multilink
        debug multilink

load-threshold (PPP/multilink - Multilink parameter; 11.3)
        This AV pair sets the load threshold at which an additional
        multilink link is added to the bundle (if load goes above) or
        deleted (if load goes below).

        The daemon side declaration is:

        service=ppp protocol=multilink {
            load-threshold=<n>
        }

        The range of <n> is [1-255].

        Related NAS commands:
        int <foo>
          [no] ppp multilink
        int virtual-template X
          multilink load-threshold <n>

        show ppp multilink
        debug multilink


Reserved for future use:

ppp-vj-slot-compression
link-compression
asyncmap
x25-addresses (PPP/VPDN)
frame-relay (PPP/VPDN)