gvpe/doc/vpe.texi

\input texinfo   @c -*-texinfo-*-
@c %**start of header
@setfilename vpe.info
@settitle vpe Manual
@setchapternewpage odd
@c %**end of header

@ifinfo
@dircategory Networking tools
@direntry
* vpe: (vpe).              The vpe Manual.
@end direntry

This is the info manual for vpe, a Virtual Private Ethernet daemon.

Copyright @copyright{} 2003 Marc Lehmann <vpe@@plan9.de>. The original
document that was used as a base for this manual is copyright
@copyright{} 1998-2002 Ivo Timmermans <ivo@@o2w.nl>, Guus Sliepen
<guus@@sliepen.eu.org> and Wessel Dankers <wsl@@nl.linux.org>.

Permission is granted to make and distribute verbatim copies of this
manual provided the copyright notice and this permission notice are
preserved on all copies.

Permission is granted to copy and distribute modified versions of this
manual under the conditions for verbatim copying, provided that the
entire resulting derived work is distributed under the terms of a
permission notice identical to this one.

@end ifinfo

@titlepage
@title vpe Manual
@subtitle Setting up a Virtual Private Ethernet
@author Marc Lehmann, Ivo Timmermans and Guus Sliepen

@page
@vskip 0pt plus 1filll
@cindex copyright
Copyright @copyright{} 2003 Marc Lehmann <vpe@@plan9.de>, 1998-2002 Ivo
Timmermans <ivo@@o2w.nl>, Guus Sliepen <guus@@sliepen.eu.org> and Wessel
Dankers <wsl@@nl.linux.org>.

Permission is granted to make and distribute verbatim copies of this
manual provided the copyright notice and this permission notice are
preserved on all copies.

Permission is granted to copy and distribute modified versions of this
manual under the conditions for verbatim copying, provided that the
entire resulting derived work is distributed under the terms of a
permission notice identical to this one.

@end titlepage

@c ==================================================================
@node Top, Introduction, (dir), (dir)

This is the documentation for the Virtual Private Ethernet suite.

As you will quickly find out, this manual was adopted from the tincd
manual, and many things, like the technical infromation or config
directives, aren't yet documented here.

@menu
* Introduction::                Introduction
* Preparations::
* Installation::
* Configuration::
* Running vpe::
* Technical information::
* About us::
* Concept Index::               All used terms explained
@end menu


@contents

@c ==================================================================
@node    Introduction, Preparations, Top, Top
@chapter Introduction

@cindex vpe
vpe is a Virtual Private Network (VPN) suite that uses tunneling and
encryption to create a secure (and optionally untrusted) private network
between hosts on the Internet.

Because the tunnel appears to the IP level network code as a normal
network device, there is no need to adapt any existing software.
The encrypted tunnels allows VPN sites to share information with each other
over the Internet without exposing any information to others.

This document is the manual for vpe.  Included are chapters on how to
configure your computer to use vpe, as well as the configuration process
of vpe itself.

@menu
* VPNs::                        Virtual Private Networks in general
* vpe::                         about vpe
* Supported platforms::
@end menu

@c ==================================================================
@node    VPNs, vpe, Introduction, Introduction
@section Virtual Private Networks

@cindex VPN

A Virtual Private Network or VPN is a network that can only be accessed by
a few elected computers that participate. This goal is achievable in more
than just one way.

@cindex private

Private networks can consist of a single stand-alone Ethernet LAN.  Or
even two computers hooked up using a null-modem cable.  In these cases,
it is obvious that the network is @emph{private}, no one can access it
from the outside.  But if your computers are linked to the Internet,
the network is not private anymore, unless one uses firewalls to block
all private traffic.  But then, there is no way to send private data to
trusted computers on the other end of the Internet.

@cindex virtual

This problem can be solved by using @emph{virtual} networks.  Virtual
networks can live on top of other networks, but they use encapsulation to
keep using their private address space so they do not interfere with the
Internet.  Mostly, virtual networks appear like a singe LAN, even though
they can span the entire world.  But virtual networks can't be secured
by using firewalls, because the traffic that flows through it has to go
through the Internet, where other people can look at it.

As is the case with either type of VPN, anybody could eavesdrop. Or worse,
alter data.  Hence it's probably advisable to encrypt and sign the data
that flows over the network.

When one introduces encryption, we can form a true VPN. Other people may
see encrypted traffic, but if they don't know how to decipher it (they
need to know the key for that), they cannot read the information that
flows through the VPN. This is what vpe was made for.


@c ==================================================================
@node    vpe, Supported platforms, VPNs, Introduction
@section vpe

@cindex vped

vpe supports encryption, it consists of a daemon (vped) for both the
receiving and sending end and a control program (vpectrl) that generates
keys, checks the config file and can control the daemon.

@cindex Traditional VPNs
@cindex scalability

vpe also allows more than two sites to connect to eachother and form a
single VPN.  Traditionally VPNs are created by making tunnels, which only
have two endpoints.  Larger VPNs with more sites are created by adding
more tunnels. vpe takes another approach: only endpoints are specified,
the software itself will take care of creating the tunnels.  This allows
for easier configuration and improved scalability.

@c ==================================================================
@node    Supported platforms,  , vpe, Introduction
@section Supported platforms

@cindex platforms

So far, vpe only owrks under Linux. Although the project it comes from,
tincd, supports FreeBSD, OpenBSD and Solaris, the lowlevel capturing
interface had to be rewritten, and due to lack of access to these
platforms, support for them has been dropped, but could be added back
relatively easily.

@ignore
@cindex release
For an up to date list of supported platforms, please check the list on
our website:
@uref{http://tinc.nl.linux.org/platforms.html}.
@end ignore


@c ==================================================================
@subsection Linux

@cindex Linux vpe was first written for Linux running on an intel x86
processor, so this is the best supported platform. The protocol however,
and actually anything about vpe, has been rewritten to support random byte
ordering and arbitrary word length (with the exception of alignment issues
within packets, so test this first ;). So in theory it should run on other
processors that Linux runs on.

@ignore
@c ==================================================================
@subsection FreeBSD

@cindex FreeBSD
vpe on FreeBSD relies on the universal tun/tap driver for its data
acquisition from the kernel.  Therefore, vpe will work on the same platforms
as this driver.  These are: FreeBSD 3.x, 4.x, 5.x.


@c ==================================================================
@subsection OpenBSD

@cindex OpenBSD
vpe on OpenBSD relies on the tun driver for its data
acquisition from the kernel. It has been verified to work under at least OpenBSD 2.9.

Tunneling IPv6 packets may not work on OpenBSD.


@c ==================================================================
@subsection Solaris

@c ==================================================================
@subsection NetBSD

@cindex NetBSD
vpe on NetBSD relies on the tun driver for its data
acquisition from the kernel. It has been verified to work under at least NetBSD 1.5.2.

Tunneling IPv6 does not work on OpenBSD.


@c ==================================================================
@subsection Solaris

@cindex Solaris
vpe on Solaris relies on the universal tun/tap driver for its data
acquisition from the kernel.  Therefore, vpe will work on the same platforms
as this driver.  These are: Solaris 8 (SunOS 5.8).

IPv6 packets cannot be tunneled on Solaris.

@c ==================================================================
@subsection Darwin (MacOS/X)

@cindex Darwin
@cindex MacOS/X
vpe on Darwin relies on the tunnel driver for its data
acquisition from the kernel. This driver is not part of Darwin but can be
downloaded from @uref{http://chrisp.de/en/projects/tunnel.html}.

IPv6 packets cannot be tunneled on Darwin.

@end ignore


@c
@c
@c
@c
@c
@c
@c       Preparing your system
@c
@c
@c
@c
@c

@c ==================================================================
@node    Preparations, Installation, Introduction, Top
@chapter Preparations

This chapter contains information on how to prepare your system to
support vpe.

@menu
* Configuring the kernel::
* Libraries::
@end menu


@c ==================================================================
@node    Configuring the kernel, Libraries, Preparations, Preparations
@section Configuring the kernel

@cindex RedHat
@cindex Debian
@cindex netlink_dev
@cindex tun
@cindex ethertap
If you are running Linux, chances are good that your kernel already supports
all the devices that vpe needs for proper operation.  For example, the
standard kernel from Redhat Linux already has support for ethertap and netlink
compiled in.  Debian users can use the modconf utility to select the modules.
If your Linux distribution supports this method of selecting devices, look out
for something called `ethertap', and `netlink_dev' if it is using a kernel
version prior to 2.4.0. In that case you will need both these devices.  If you
are using kernel 2.4.0 or later, you need to select `tun'.

@cindex Kernel-HOWTO
If you can install these devices in a similar manner, you may skip this section.
Otherwise, you will have to recompile the kernel in order to turn on the required features.
If you are unfamiliar with the process of configuring and compiling a new kernel,
you should read the @uref{http://howto.linuxberg.com/LDP/HOWTO/Kernel-HOWTO.html, Kernel HOWTO} first.

@menu
* Configuration of Linux kernels 2.4.0 and higher::
* Configuration of FreeBSD kernels::
* Configuration of OpenBSD kernels::
* Configuration of NetBSD kernels::
* Configuration of Solaris kernels::
* Configuration of Darwin (MacOS/X) kernels::
@end menu


@c ==================================================================
@node       Configuration of Linux kernels 2.4.0 and higher, Configuration of FreeBSD kernels, Configuring the kernel, Configuring the kernel
@subsection Configuration of Linux kernels 2.4.0 and higher

Here are the options you have to turn on when configuring a new kernel:

@example
Code maturity level options
[*] Prompt for development and/or incomplete code/drivers
Network device support
<M> Universal tun/tap device driver support
@end example

It's not necessary to compile this driver as a module, even if you are going to
run more than one instance of vpe.

If you have an early 2.4 kernel, you can choose both the tun/tap driver and the
`Ethertap network tap' device.  This latter is marked obsolete, and chances are
that it won't even function correctly anymore.  Make sure you select the
universal tun/tap driver.

If you decide to build the tun/tap driver as a kernel module, add these lines
to @file{/etc/modules.conf}:

@example
alias char-major-10-200 tun
@end example

@c ==================================================================
@node       Configuration of FreeBSD kernels, Configuration of OpenBSD kernels, Configuration of Linux kernels 2.4.0 and higher, Configuring the kernel
@subsection Configuration of FreeBSD kernels

This section will contain information on how to configure your FreeBSD
kernel to support the universal tun/tap device.  For 4.1 and higher
versions, this is included in the default kernel configuration, for earlier
systems (4.0 and earlier), you need to install the universal tun/tap driver
yourself.

Unfortunately somebody still has to write the text.


@c ==================================================================
@node       Configuration of OpenBSD kernels, Configuration of NetBSD kernels, Configuration of FreeBSD kernels, Configuring the kernel
@subsection Configuration of OpenBSD kernels

This section will contain information on how to configure your OpenBSD
kernel to support the tun device.  For 2.9 and 3.0 systems,
this is included in the default kernel configuration.

Unfortunately somebody still has to write the text.


@c ==================================================================
@node       Configuration of NetBSD kernels, Configuration of Solaris kernels, Configuration of OpenBSD kernels, Configuring the kernel
@subsection Configuration of NetBSD kernels

This section will contain information on how to configure your NetBSD
kernel to support the tun device.  For 1.5.2 systems,
this is included in the default kernel configuration.

Unfortunately somebody still has to write the text.


@c ==================================================================
@node       Configuration of Solaris kernels, Configuration of Darwin (MacOS/X) kernels, Configuration of NetBSD kernels, Configuring the kernel
@subsection Configuration of Solaris kernels

This section will contain information on how to configure your Solaris
kernel to support the universal tun/tap device.  For Solaris 8 (SunOS 5.8),
this is included in the default kernel configuration.

Unfortunately somebody still has to write the text.


@c ==================================================================
@node       Configuration of Darwin (MacOS/X) kernels, , Configuration of Solaris kernels, Configuring the kernel
@subsection Configuration of Darwin (MacOS/X) kernels

Darwin does not come with a tunnel driver. You must download it at
@uref{http://chrisp.de/en/projects/tunnel.html}. If compiling the source fails,
try the binary module. The tunnel driver must be loaded before starting vpe
with the following command:

@example
kmodload tunnel
@end example

Once loaded, the tunnel driver will automatically create @file{/dev/tun0}..@file{/dev/tun3}
and the corresponding network interfaces.


@c ==================================================================
@node    Libraries,  , Configuring the kernel, Preparations
@section Libraries

@cindex requirements
@cindex libraries

Before you can configure or build vpe, you need to have the OpenSSL
library installed on your system.  If you try to configure vpe without
having it installed, configure will give you an error message, and stop.

@menu
* OpenSSL::
@end menu


@c ==================================================================
@node       OpenSSL, , Libraries, Libraries
@subsection OpenSSL

@cindex OpenSSL
For all cryptography-related functions, vpe uses the functions provided
by the OpenSSL library.

If this library is not installed, you wil get an error when configuring
vpe for build.  Support for running vpe without having OpenSSL
installed @emph{may} be added in the future.

You can use your operating system's package manager to install this if
available.  Make sure you install the development AND runtime versions
of this package.

If you have to install OpenSSL manually, you can get the source code
from @url{http://www.openssl.org/}.  Instructions on how to configure,
build and install this package are included within the package.  Please
make sure you build development and runtime libraries (which is the
default).

If you installed the OpenSSL libraries from source, it may be necessary
to let configure know where they are, by passing configure one of the
--with-openssl-* parameters.

@example
--with-openssl=DIR      OpenSSL library and headers prefix
--with-openssl-include=DIR OpenSSL headers directory
                        (Default is OPENSSL_DIR/include)
--with-openssl-lib=DIR  OpenSSL library directory
                        (Default is OPENSSL_DIR/lib)
@end example

@ignore
@c
@c
@c WARNING: this is commented out because I didn't add this exception
@c to my (vpe) code.
@c
@c
@subsubheading License

@cindex license
Since the license under which OpenSSL is distributed is not directly
compatible with the terms of the GNU GPL
@uref{http://www.openssl.org/support/faq.html#LEGAL2}, therefore we
include an addition to the GPL (see also the file COPYING.README):

@quotation
This program is released under the GPL with the additional exemption
that compiling, linking, and/or using OpenSSL is allowed.  You may
provide binary packages linked to the OpenSSL libraries, provided that
all other requirements of the GPL are met.
@end quotation
@end ignore


@c
@c
@c
@c      Installing vpe
@c
@c
@c
@c

@c ==================================================================
@node    Installation, Configuration, Preparations, Top
@chapter Installation

@ignore
If you use Debian, you may want to install one of the
precompiled packages for your system.  These packages are equipped with
system startup scripts and sample configurations.

If you cannot use one of the precompiled packages, or you want to compile vpe
for yourself, you can use the source.  The source is distributed under
the GNU General Public License (GPL).  Download the source from the
@uref{http://vpe.nl.linux.org/download.html, download page}, which has
the checksums of these files listed; you may wish to check these with
md5sum before continuing.
@end ignore

vpe comes in a convenient autoconf/automake package, which you can just
treat the same as any other package.  Which is just untar it, type
`./configure' and then `make'.

More detailed instructions are in the file @file{INSTALL}, which is
included in the source distribution.

@menu
* Building and installing vpe::
* System files::
@end menu


@c ==================================================================
@node    Building and installing vpe, System files, Installation, Installation
@section Building and installing vpe

Detailed instructions on configuring the source, building vpe and
installing vpe can be found in the file called @file{INSTALL}.

@c ==================================================================
@node    System files,  , Building and installing vpe, Installation
@section System files

Before you can run vpe, you must make sure you have all the needed
files on your system.

@menu
* Device files::
* Other files::
@end menu


@c ==================================================================
@node       Device files, Other files, System files, System files
@subsection Device files

@cindex device files
First, you'll need the special device file(s) that form the interface
between the kernel and the daemon.

The permissions for these files have to be such that only the super user
may read/write to this file.  You'd want this, because otherwise
eavesdropping would become a bit too easy.  This does, however, imply
that you'd have to run vped as root.

If you use Linux and have a kernel version prior to 2.4.0, you have to make the
ethertap devices:

@example
mknod -m 600 /dev/tap0 c 36 16
mknod -m 600 /dev/tap1 c 36 17
...
mknod -m 600 /dev/tap@emph{N} c 36 @emph{N+16}
@end example

There is a maximum of 16 ethertap devices.

If you use the universal tun/tap driver, you have to create the
following device file (unless it already exist):

@example
mknod -m 600 /dev/tun c 10 200
@end example

If you use Linux, and you run the new 2.4 kernel using the devfs filesystem,
then the tun/tap device will probably be automatically generated as
@file{/dev/net/tun}.

Unlike the ethertap device, you do not need multiple device files if
you are planning to run multiple vpe daemons.


@c ==================================================================
@node       Other files,  , Device files, System files
@subsection Other files

@subsubheading @file{/etc/networks}

You may add a line to @file{/etc/networks} so that your VPN will get a
symbolic name.  For example:

@example
myvpn 10.0.0.0
@end example

@ignore
@subsubheading @file{/etc/services}

@cindex port numbers
You may add this line to @file{/etc/services}.  The effect is that you
may supply a @samp{vpe} as a valid port number to some programs.  The
number 655 is registered with the IANA.

@example
vpe            655/tcp    TINC
vpe            655/udp    TINC
@end example
@end ignore


@c
@c
@c
@c
@c         Configuring vpe
@c
@c
@c
@c


@c ==================================================================
@node    Configuration, Running vpe, Installation, Top
@chapter Configuration

@menu
* Configuration introduction::
* Multiple networks::
* How connections work::
* Configuration files::
* Generating keypairs::
* Network interfaces::
* Example configuration::
@end menu

@c ==================================================================
@node    Configuration introduction, Multiple networks, Configuration, Configuration
@section Configuration introduction

@cindex Network Administrators Guide
Before actually starting to configure vpe and editing files,
make sure you have read this entire section so you know what to expect.
Then, make it clear to yourself how you want to organize your VPN:
What are the nodes (computers running vpe)?
What IP addresses/subnets do they have?
What is the network mask of the entire VPN?
Do you need special firewall rules?
Do you have to set up masquerading or forwarding rules?
These questions can only be answered by yourself,
you will not find the answers in this documentation.
Make sure you have an adequate understanding of networks in general.
A good resource on networking is the
@uref{http://www.linuxdoc.org/LDP/nag2/, Linux Network Administrators Guide}.

If you have everything clearly pictured in your mind,
proceed in the following order:
First, generate the configuration files (@file{vpe.conf}, your host configuration file, @file{vpe-up} and perhaps @file{vpe-down}).
Then generate the keypairs.
Finally, distribute the host configuration files.
These steps are described in the subsections below.


@c ==================================================================
@node    Multiple networks, How connections work, Configuration introduction, Configuration
@section Multiple networks

@cindex multiple networks
@cindex netname
In order to allow you to run more than one vpe daemon on one computer,
for instance if your computer is part of more than one VPN,
you can assign a ``netname'' to your VPN.
It is not required if you only run one vpe daemon,
it doesn't even have to be the same on all the sites of your VPN,
but it is recommended that you choose one anyway.

We will asume you use a netname throughout this document.
This means that you call vped with the -n argument,
which will assign a netname to this daemon.

The effect of this is that the daemon will set its configuration
``root'' to /etc/vpe/netname/, where netname is your argument to the -n
option.  You'll notice that it appears in syslog as ``vpe.netname''.

However, it is not strictly necessary that you call vpe with the -n
option.  In this case, the network name would just be empty, and it will
be used as such.  vpe now looks for files in /etc/vpe/, instead of
/etc/vpe/netname/; the configuration file should be /etc/vpe/vpe.conf,
and the host configuration files are now expected to be in /etc/vpe/hosts/.

But it is highly recommended that you use this feature of vpe, because
it will be so much clearer whom your daemon talks to.  Hence, we will
assume that you use it.


@c ==================================================================
@node    How connections work, Configuration files, Multiple networks, Configuration
@section How connections work

When vpe starts up, it parses the command-line options and then
reads in the configuration file vpe.conf.
If it sees one or more  `ConnectTo' values pointing to other vpe daemons in that file,
it will try to connect to those other daemons.
Whether this succeeds or not and whether `ConnectTo' is specified or not,
vpe will listen for incoming connection from other deamons.
If you did specify a `ConnectTo' value and the other side is not responding,
vpe will keep retrying.
This means that once started, vpe will stay running until you tell it to stop,
and failures to connect to other vpe daemons will not stop your vpe daemon
for trying again later.
This means you don't have to intervene if there are temporary network problems.

@cindex client
@cindex server
There is no real disvpetion between a server and a client in vpe.
If you wish, you can view a vpe daemon without a `ConnectTo' value as a server,
and one which does specify such a value as a client.
It does not matter if two vpe daemons have a `ConnectTo' value pointing to each other however.


@c ==================================================================
@node    Configuration files, Generating keypairs, How connections work, Configuration
@section Configuration files

The actual configuration of the daemon is done in the file
@file{/etc/vpe/netname/vpe.conf} and at least one other file in the directory
@file{/etc/vpe/netname/hosts/}.

These file consists of comments (lines started with a #) or assignments
in the form of

@example
Variable = Value.
@end example

The variable names are case insensitive, and any spaces, tabs, newlines
and carriage returns are ignored.  Note: it is not required that you put
in the `=' sign, but doing so improves readability.  If you leave it
out, remember to replace it with at least one space character.

In this section all valid variables are listed in alphabetical order.
The default value is given between parentheses,
other comments are between square brackets and
required directives are given in @strong{bold}.

@menu
* Main configuration variables::
* Host configuration variables::
* How to configure::
@end menu


@c ==================================================================
@node    Main configuration variables, Host configuration variables, Configuration files, Configuration files
@subsection Main configuration variables

@table @asis
@cindex AddressFamily
@item AddressFamily = <ipv4|ipv6|any> (ipv4) [experimental]
This option affects the address family of listening and outgoing sockets.
If "any" is selected, then depending on the operating system
both IPv4 and IPv6 or just IPv6 listening sockets will be created.

@cindex BindToInterface
@item BindToInterface = <interface> [experimental]
If you have more than one network interface in your computer, vpe will
by default listen on all of them for incoming connections.  It is
possible to bind vpe to a single interface like eth0 or ppp0 with this
variable.

This option may not work on all platforms.

@cindex ConnectTo
@item @strong{ConnectTo = <name>}
Specifies which other vpe daemon to connect to on startup.
Multiple ConnectTo variables may be specified,
in which case outgoing connections to each specified vpe daemon are made.
The names should be known to this vpe daemon
(i.e., there should be a host configuration file for the name on the ConnectTo line).

If you don't specify a host with ConnectTo,
vpe won't try to connect to other daemons at all,
and will instead just listen for incoming connections.

@cindex Device
@item @strong{Device = <device>} (/dev/tap0 or /dev/net/tun)
The virtual network device to use.  Note that you can only use one device per
daemon.  See also @ref{Device files}.

@cindex Hostnames
@item Hostnames = <yes|no> (no)
This option selects whether IP addresses (both real and on the VPN)
should be resolved.  Since DNS lookups are blocking, it might affect
vpe's efficiency, even stopping the daemon for a few seconds everytime
it does a lookup if your DNS server is not responding.

This does not affect resolving hostnames to IP addresses from the
configuration file.

@cindex Interface
@item Interface = <interface>
Defines the name of the interface corresponding to the virtual network device.
Depending on the operating system and the type of device this may or may not actually set the name.
Currently this option only affects the Linux tun/tap device.

@cindex Mode
@item Mode = <router|switch|hub> (router)
This option selects the way packets are routed to other daemons.

@table @asis
@cindex router
@item router
In this mode Subnet
variables in the host configuration files will be used to form a routing table.
Only unicast packets of routable protocols (IPv4 and IPv6) are supported in this mode.

@cindex switch
@item switch
In this mode the MAC addresses of the packets on the VPN will be used to
dynamically create a routing table just like an Ethernet switch does.
Unicast, multicast and broadcast packets of every protocol that runs over Ethernet are supported in this mode
at the cost of frequent broadcast ARP requests and routing table updates.

@cindex hub
@item hub
This mode is almost the same as the switch mode, but instead
every packet will be broadcast to the other daemons
while no routing table is managed.
@end table

@cindex KeyExpire
@item KeyExpire = <seconds> (3600)
This option controls the time the encryption keys used to encrypt the data
are valid.  It is common practice to change keys at regular intervals to
make it even harder for crackers, even though it is thought to be nearly
impossible to crack a single key.

@cindex MACExpire
@item MACExpire = <seconds> (600)
This option controls the amount of time MAC addresses are kept before they are removed.
This only has effect when Mode is set to "switch".

@cindex Name
@item @strong{Name = <name>}
This is a symbolic name for this connection.  It can be anything

@cindex PingTimeout
@item PingTimeout = <seconds> (60)
The number of seconds of inactivity that vpe will wait before sending a
probe to the other end.  If that other end doesn't answer within that
same amount of seconds, the connection is terminated, and the others
will be notified of this.

@cindex PriorityInheritance
@item PriorityInheritance = <yes|no> (no) [experimental]
When this option is enabled the value of the TOS field of tunneled IPv4 packets
will be inherited by the UDP packets that are sent out.

@cindex PrivateKey
@item PrivateKey = <key> [obsolete]
This is the RSA private key for vpe. However, for safety reasons it is
advised to store private keys of any kind in separate files. This prevents
accidental eavesdropping if you are editting the configuration file.

@cindex PrivateKeyFile
@item @strong{PrivateKeyFile = <path>} [recommended]
This is the full path name of the RSA private key file that was
generated by ``vped --generate-keys''.  It must be a full path, not a
relative directory.

Note that there must be exactly one of PrivateKey
or PrivateKeyFile
specified in the configuration file.

@end table


@c ==================================================================
@node    Host configuration variables, How to configure, Main configuration variables, Configuration files
@subsection Host configuration variables

@table @asis
@cindex Address
@item @strong{Address = <IP address|hostname>} [recommended]
This variable is only required if you want to connect to this host.  It
must resolve to the external IP address where the host can be reached,
not the one that is internal to the VPN.

@cindex Cipher
@item Cipher = <cipher> (blowfish)
The symmetric cipher algorithm used to encrypt UDP packets.
Any cipher supported by OpenSSL is recognized.

@cindex Compression
@item Compression = <level> (0)
This option sets the level of compression used for UDP packets.
Possible values are 0 (off), 1 (fast) and any integer up to 9 (best).

@cindex Digest
@item Digest = <digest> (sha1)
The digest algorithm used to authenticate UDP packets.
Any digest supported by OpenSSL is recognized.
Furthermore, specifying "none" will turn off packet authentication.

@cindex IndirectData
@item IndirectData = <yes|no> (no)
This option specifies whether other vpe daemons besides the one you
specified with ConnectTo can make a direct connection to you.  This is
especially useful if you are behind a firewall and it is impossible to
make a connection from the outside to your vpe daemon.  Otherwise, it
is best to leave this option out or set it to no.

@cindex MACLength
@item MACLength = <length> (4)
The length of the message authentication code used to authenticate UDP packets.
Can be anything from 0
up to the length of the digest produced by the digest algorithm.

@cindex Port
@item Port = <port> (655)
This is the port this vpe daemon listens on.
You can use decimal portnumbers or symbolic names (as listed in /etc/services).

@cindex PublicKey
@item PublicKey = <key> [obsolete]
This is the RSA public key for this host.

@cindex PublicKeyFile
@item PublicKeyFile = <path> [obsolete]
This is the full path name of the RSA public key file that was generated
by ``vped --generate-keys''.  It must be a full path, not a relative
directory.

@cindex PEM format
From version 1.0pre4 on vpe will store the public key directly into the
host configuration file in PEM format, the above two options then are not
necessary. Either the PEM format is used, or exactly
@strong{one of the above two options} must be specified
in each host configuration file, if you want to be able to establish a
connection with that host.

@cindex Subnet
@item Subnet = <address[/prefixlength]>
The subnet which this vpe daemon will serve.
vpe tries to look up which other daemon it should send a packet to by searching the appropiate subnet.
If the packet matches a subnet,
it will be sent to the daemon who has this subnet in his host configuration file.
Multiple subnet lines can be specified for each daemon.

Subnets can either be single MAC, IPv4 or IPv6 addresses,
in which case a subnet consisting of only that single address is assumed,
or they can be a IPv4 or IPv6 network address with a prefixlength.
Shorthand notations are not supported.
For example, IPv4 subnets must be in a form like 192.168.1.0/24,
where 192.168.1.0 is the network address and 24 is the number of bits set in the netmask.
Note that subnets like 192.168.1.1/24 are invalid!
Read a networking HOWTO/FAQ/guide if you don't understand this.
IPv6 subnets are notated like fec0:0:0:1:0:0:0:0/64.
MAC addresses are notated like 0:1a:2b:3c:4d:5e.

@cindex CIDR notation
prefixlength is the number of bits set to 1 in the netmask part; for
example: netmask 255.255.255.0 would become /24, 255.255.252.0 becomes
/22. This conforms to standard CIDR notation as described in
@uref{ftp://ftp.isi.edu/in-notes/rfc1519.txt, RFC1519}

@cindex TCPonly
@item TCPonly = <yes|no> (no) [experimental]
If this variable is set to yes, then the packets are tunnelled over a
TCP connection instead of a UDP connection.  This is especially useful
for those who want to run a vpe daemon from behind a masquerading
firewall, or if UDP packet routing is disabled somehow.
Setting this options also implicitly sets IndirectData.
@end table


@c ==================================================================
@node    How to configure,  , Host configuration variables, Configuration files
@subsection How to configure

@subsubheading Step 1.  Creating the main configuration file

The main configuration file will be called @file{/etc/vpe/netname/vpe.conf}.
Adapt the following example to create a basic configuration file:

@example
Name = @emph{yourname}
Device = @emph{/dev/tap0}
PrivateKeyFile = /etc/vpe/@emph{netname}/rsa_key.priv
@end example

Then, if you know to which other vpe daemon(s) yours is going to connect,
add `ConnectTo' values.

@subsubheading Step 2.  Creating your host configuration file

If you added a line containing `Name = yourname' in the main configuarion file,
you will need to create a host configuration file @file{/etc/vpe/netname/hosts/yourname}.
Adapt the following example to create a host configuration file:

@example
Address = @emph{your.real.hostname.org}
Subnet = @emph{192.168.1.0/24}
@end example

You can also use an IP address instead of a hostname.
The `Subnet' specifies the address range that is local for @emph{your part of the VPN only}.
If you have multiple address ranges you can specify more than one `Subnet'.
You might also need to add a `Port' if you want your vpe daemon to run on a different port number than the default (655).


@c ==================================================================
@node    Generating keypairs, Network interfaces, Configuration files, Configuration
@section Generating keypairs

@cindex key generation
Now that you have already created the main configuration file and your host configuration file,
you can easily create a public/private keypair by entering the following command:

@example
vped -n @emph{netname} -K
@end example

vpe will generate a public and a private key and ask you where to put them.
Just press enter to accept the defaults.


@c ==================================================================
@node    Network interfaces, Example configuration, Generating keypairs, Configuration
@section Network interfaces

Before vpe can start transmitting data over the tunnel, it must
set up the virtual network interface.

First, decide which IP addresses you want to have associated with these
devices, and what network mask they must have.

vpe will open a virtual network device (@file{/dev/tun}, @file{/dev/tap0} or similar),
which will also create a network interface called something like `tun0', `tap0', or,
if you are using the Linux tun/tap driver, the network interface will by default have the same name as the netname.

@cindex vpe-up
You can configure the network interface by putting ordinary ifconfig, route, and other commands
to a script named @file{/etc/vpe/netname/vpe-up}. When vpe starts, this script
will be executed. When vpe exits, it will execute the script named
@file{/etc/vpe/netname/vpe-down}, but normally you don't need to create that script.

An example @file{vpe-up} script:

@example
#!/bin/sh
ifconfig $INTERFACE hw ether fe:fd:0:0:0:0
ifconfig $INTERFACE 192.168.1.1 netmask 255.255.0.0
ifconfig $INTERFACE -arp
@end example

@cindex MAC address
@cindex hardware address
The first line sets up the MAC address of the network interface.
Due to the nature of how Ethernet and vpe work, it has to be set to fe:fd:0:0:0:0
for vpe to work in it's normal mode.
If you configured vpe to work in `switch' or `hub' mode, the hardware address should instead
be set to a unique address instead of fe:fd:0:0:0:0.

You can use the environment variable $INTERFACE to get the name of the interface.
However, this might not be reliable. If in doubt, use the name of the interface explicitly.

@cindex ifconfig
The next line gives the interface an IP address and a netmask.
The kernel will also automatically add a route to this interface, so normally you don't need
to add route commands to the @file{vpe-up} script.
The kernel will also bring the interface up after this command.
@cindex netmask
The netmask is the mask of the @emph{entire} VPN network, not just your
own subnet.

@cindex arp
The last line tells the kernel not to use ARP on that interface.
Again this has to do with how Ethernet and vpe work.
Use this option only if you are running vpe under Linux and are using vpe's normal routing mode.


@c ==================================================================
@node    Example configuration,  , Network interfaces, Configuration
@section Example configuration


@cindex example
Imagine the following situation.  Branch A of our example `company' wants to connect
three branch offices in B, C and D using the Internet.  All four offices
have a 24/7 connection to the Internet.

A is going to serve as the center of the network.  B and C will connect
to A, and D will connect to C.  Each office will be assigned their own IP
network, 10.x.0.0.

@example
A: net 10.1.0.0 mask 255.255.0.0 gateway 10.1.54.1 internet IP 1.2.3.4
B: net 10.2.0.0 mask 255.255.0.0 gateway 10.2.1.12 internet IP 2.3.4.5
C: net 10.3.0.0 mask 255.255.0.0 gateway 10.3.69.254 internet IP 3.4.5.6
D: net 10.4.0.0 mask 255.255.0.0 gateway 10.4.3.32 internet IP 4.5.6.7
@end example

``gateway'' is the VPN IP address of the machine that is running the
vped.  ``internet IP'' is the IP address of the firewall, which does not
need to run vped, but it must do a port forwarding of TCP&UDP on port
655 (unless otherwise configured).

In this example, it is assumed that eth0 is the interface that points to
the inner (physical) LAN of the office, although this could also be the
same as the interface that leads to the Internet.  The configuration of
the real interface is also shown as a comment, to give you an idea of
how these example host is set up. All branches use the netname `company'
for this particular VPN.

@subsubheading For Branch A

@emph{BranchA} would be configured like this:

In @file{/etc/vpe/company/vpe-up}:

@example
# Real interface of internal network:
# ifconfig eth0 10.1.54.1 netmask 255.255.0.0 broadcast 10.1.255.255

ifconfig tap0 hw ether fe:fd:0:0:0:0
ifconfig tap0 10.1.54.1 netmask 255.0.0.0
ifconfig tap0 -arp
@end example

and in @file{/etc/vpe/company/vpe.conf}:

@example
Name = BranchA
PrivateKey = /etc/vpe/company/rsa_key.priv
Device = /dev/tap0
@end example

On all hosts, /etc/vpe/company/hosts/BranchA contains:

@example
Subnet = 10.1.0.0/16
Address = 1.2.3.4

Note that the IP addresses of eth0 and tap0 are the same.
This is quite possible, if you make sure that the netmasks of the interfaces are different.
It is in fact recommended to give give both real internal network interfaces and tap interfaces the same IP address,
since that will make things a lot easier to remember and set up.

-----BEGIN RSA PUBLIC KEY-----
...
-----END RSA PUBLIC KEY-----
@end example


@subsubheading For Branch B

In @file{/etc/vpe/company/vpe-up}:

@example
# Real interface of internal network:
# ifconfig eth0 10.2.43.8 netmask 255.255.0.0 broadcast 10.2.255.255

ifconfig tap0 hw ether fe:fd:0:0:0:0
ifconfig tap0 10.2.1.12 netmask 255.0.0.0
ifconfig tap0 -arp
@end example

and in @file{/etc/vpe/company/vpe.conf}:

@example
Name = BranchB
ConnectTo = BranchA
PrivateKey = /etc/vpe/company/rsa_key.priv
@end example

Note here that the internal address (on eth0) doesn't have to be the
same as on the tap0 device.  Also, ConnectTo is given so that no-one can
connect to this node.

On all hosts, in @file{/etc/vpe/company/hosts/BranchB}:

@example
Subnet = 10.2.0.0/16
Address = 2.3.4.5

-----BEGIN RSA PUBLIC KEY-----
...
-----END RSA PUBLIC KEY-----
@end example


@subsubheading For Branch C

In @file{/etc/vpe/company/vpe-up}:

@example
# Real interface of internal network:
# ifconfig eth0 10.3.69.254 netmask 255.255.0.0 broadcast 10.3.255.255

ifconfig tap1 hw ether fe:fd:0:0:0:0
ifconfig tap1 10.3.69.254 netmask 255.0.0.0
ifconfig tap1 -arp
@end example

and in @file{/etc/vpe/company/vpe.conf}:

@example
Name = BranchC
ConnectTo = BranchA
Device = /dev/tap1
@end example

C already has another daemon that runs on port 655, so they have to
reserve another port for vpe. It knows the portnumber it has to listen on
from it's own host configuration file.

On all hosts, in @file{/etc/vpe/company/hosts/BranchC}:

@example
Address = 3.4.5.6
Subnet = 10.3.0.0/16
Port = 2000

-----BEGIN RSA PUBLIC KEY-----
...
-----END RSA PUBLIC KEY-----
@end example


@subsubheading For Branch D

In @file{/etc/vpe/company/vpe-up}:

@example
# Real interface of internal network:
# ifconfig eth0 10.4.3.32 netmask 255.255.0.0 broadcast 10.4.255.255

ifconfig company hw ether fe:fd:0:0:0:0
ifconfig company 10.4.3.32 netmask 255.0.0.0
ifconfig company -arp
@end example

and in @file{/etc/vpe/company/vpe.conf}:

@example
Name = BranchD
ConnectTo = BranchC
Device = /dev/net/tun
PrivateKeyFile = /etc/vpe/company/rsa_key.priv
@end example

D will be connecting to C, which has a vped running for this network on
port 2000. It knows the port number from the host configuration file.
Also note that since D uses the tun/tap driver, the network interface
will not be called `tun' or `tap0' or something like that, but will
have the same name as netname.

On all hosts, in @file{/etc/vpe/company/hosts/BranchD}:

@example
Subnet = 10.4.0.0/16
Address = 4.5.6.7

-----BEGIN RSA PUBLIC KEY-----
...
-----END RSA PUBLIC KEY-----
@end example

@subsubheading Key files

A, B, C and D all have generated a public/private keypair with the following command:

@example
vped -n company -K
@end example

The private key is stored in @file{/etc/vpe/company/rsa_key.priv},
the public key is put into the host configuration file in the @file{/etc/vpe/company/hosts/} directory.
During key generation, vpe automatically guesses the right filenames based on the -n option and
the Name directive in the @file{vpe.conf} file (if it is available).

@subsubheading Starting

After each branch has finished configuration and they have distributed
the host configuration files amongst them, they can start their vpe daemons.
They don't necessarily have to wait for the other branches to have started
their daemons, vpe will try connecting until they are available.


@c ==================================================================
@node    Running vpe, Technical information, Configuration, Top
@chapter Running vpe

If everything else is done, you can start vpe by typing the following command:

@example
vped -n @emph{netname}
@end example

@cindex daemon
vpe will detach from the terminal and continue to run in the background like a good daemon.
If there are any problems however you can try to increase the debug level
and look in the syslog to find out what the problems are.

@menu
* Runtime options::
* Error messages::
@end menu


@c ==================================================================
@node    Runtime options, Error messages,  , Running vpe
@section Runtime options

Besides the settings in the configuration file, vpe also accepts some
command line options.

@cindex command line
@cindex runtime options
@cindex options
@c from the manpage
@table @samp
@item --bypass-security
Disables encryption and authentication.
Only useful for debugging.

@item -c, --config=PATH
Read configuration options from the directory PATH.  The default is
@file{/etc/vpe/netname/}.

@cindex debug level
@item -d, --debug=LEVEL
Set debug level to LEVEL.  The higher the debug level, the more gets
logged.  Everything goes via syslog.

@item -K, --generate-keys[=BITS]
Generate public/private keypair of BITS length. If BITS is not specified,
1024 is the default. vpe will ask where you want to store the files,
but will default to the configuration directory (you can use the -c or -n option
in combination with -K). After that, vpe will quit.

@item --help
Display a short reminder of these runtime options and terminate.

@item -k, --kill[=SIGNAL]
Attempt to kill a running vped (optionally with the specified SIGNAL instead of SIGTERM) and exit.
Use it in conjunction with the -n option to make sure you kill the right vpe daemon.

@item -n, --net=NETNAME
Connect to net NETNAME.  @xref{Multiple networks}.

@item -D, --no-detach
Don't fork and detach.
This will also disable the automatic restart mechanism for fatal errors.

@item -L, --mlock
Lock vpe into main memory.
This will prevent sensitive data like shared private keys to be written to the system swap files/partitions.

@item --version
Output version information and exit.

@end table


@c ==================================================================
@node    Error messages,  , Runtime options, Running vpe
@section Error messages

What follows is a list of the most common error messages you can see
when configuring vpe.  Most of these messages are visible in the syslog
only, so keep an eye on it!

@table @strong
@item Could not open /dev/tap0: No such device

@itemize
@item You forgot to `modprobe netlink_dev' or `modprobe ethertap'.
@item You forgot to compile `Netlink device emulation' in the kernel.
@end itemize

@item Can't write to /dev/net/tun: No such device

@itemize
@item You forgot to `modprobe tun'.
@item You forgot to compile `Universal TUN/TAP driver' in the kernel.
@end itemize

@item Packet with destination 1.2.3.4 is looping back to us!

@itemize
@item Something is not configured right. Packets are being sent out to the
virtual network device, but according to the Subnet directives in your host configuration
file, those packets should go to your own host. Most common mistake is that
you have a Subnet line in your host configuration file with a prefix length which is
just as large as the prefix of the virtual network interface. The latter should in almost all
cases be larger. Rethink your configuration.
Note that you will only see this message if you specified a debug
level of 5 or higher!
@item Chances are that a `Subnet = ...' line in the host configuration file of this vpe daemon is wrong.
Change it to a subnet that is accepted locally by another interface,
or if that is not the case, try changing the prefix length into /32. 
@end itemize

@item Network doesn't work, syslog shows only packets of length 46

@item Network address and prefix length do not match!

@itemize
@item The Subnet field must contain a @emph{network} address.
@item If you only want to use one IP address, set the netmask to /32.
@end itemize

@item This is a bug: net.c:253: 24: Some error

@itemize
@item This is something that should not have happened.
Please report this, and tell us exactly what went wrong before you got
this message.  In normal operation, these errors should not occur.
@end itemize

@item Error reading RSA key file `rsa_key.priv': No such file or directory

@itemize
@item You must specify the complete pathname.
Specifying a relative path does not make sense here.  vpe changes its
directory to / when starting (to avoid keeping a mount point busy).
@end itemize

@end table

@c ==================================================================
@node    Technical information, About us, Running vpe, Top
@chapter Technical information


@menu
* The connection::
* The meta-protocol::
* Security::
@end menu


@c ==================================================================
@node    The connection, The meta-protocol, Technical information, Technical information
@section The connection

@cindex connection
vpe is a daemon that takes VPN data and transmit that to another host
computer over the existing Internet infrastructure.

@menu
* The UDP tunnel::
* The meta-connection::
@end menu


@c ==================================================================
@node    The UDP tunnel, The meta-connection, The connection, The connection
@subsection The UDP tunnel

@cindex virtual network device
@cindex frame type
The data itself is read from a character device file, the so-called
@emph{virtual network device}.  This device is associated with a network
interface.  Any data sent to this interface can be read from the device,
and any data written to the device gets sent from the interface.  Data to
and from the device is formatted as if it were a normal Ethernet card,
so a frame is preceded by two MAC addresses and a @emph{frame type}
field.

So when vpe reads an Ethernet frame from the device, it determines its
type. When vpe is in it's default routing mode, it can handle IPv4 and IPv6
packets. Depending on the Subnet lines, it will send the packets off to their destination.
In the `switch' and `hub' mode, vpe will use broadcasts and MAC address discovery
to deduce the destination of the packets.
Since the latter modes only depend on the link layer information,
any protocol that runs over Ethernet is supported (for instance IPX and Appletalk).

After the destination has been determined,
the packet will be compressed (optionally),
a sequence number will be added to the packet,
the packet will then be encrypted
and a message authentication code will be appended.

@cindex encapsulating
@cindex UDP
When that is done, time has come to actually transport the
packet to the destination computer.  We do this by sending the packet
over an UDP connection to the destination host.  This is called
@emph{encapsulating}, the VPN packet (though now encrypted) is
encapsulated in another IP datagram.

When the destination receives this packet, the same thing happens, only
in reverse.  So it checks the message authentication code, decrypts the contents of the UDP datagram,
checks the sequence number
and writes the decrypted information to its own virtual network device.

To let the kernel on the receiving end accept the packet, the destination MAC
address must match that of the virtual network interface.
If vpe is in it's default routing mode, ARP does not work, so the correct destination MAC cannot be set
by the sending daemons.
vpe solves this by letting the receiving end detect the MAC address
and overwriting the destination MAC address of the received packet.
However, the MAC address of the network interface at the receiver might not always be known to vpe.
That is the reason why you should set the MAC address of your tap interface to that address
when in routing mode.

In switch or hub modes ARP does work so the sender already knows the correct destination MAC address.
In those modes every interface should have a unique MAC address, so make sure they are not the same.


@c ==================================================================
@node    The meta-connection,  , The UDP tunnel, The connection
@subsection The meta-connection

Having only a UDP connection available is not enough.  Though suitable
for transmitting data, we want to be able to reliably send other
information, such as routing and session key information to somebody.

@cindex TCP
TCP is a better alternative, because it already contains protection
against information being lost, unlike UDP.

So we establish two connections.  One for the encrypted VPN data, and one
for other information, the meta-data.  Hence, we call the second
connection the meta-connection.  We can now be sure that the
meta-information doesn't get lost on the way to another computer.

@cindex data-protocol
@cindex meta-protocol
Like with any communication, we must have a protocol, so that everybody
knows what everything stands for, and how she should react.  Because we
have two connections, we also have two protocols.  The protocol used for
the UDP data is the ``data-protocol,'' the other one is the
``meta-protocol.''

The reason we don't use TCP for both protocols is that UDP is much
better for encapsulation, even while it is less reliable.  The real
problem is that when TCP would be used to encapsulate a TCP stream
that's on the private network, for every packet sent there would be
three ACKs sent instead of just one.  Furthermore, if there would be
a timeout, both TCP streams would sense the timeout, and both would
start re-sending packets.


@c ==================================================================
@node    The meta-protocol, Security, The connection, Technical information
@section The meta-protocol

The meta protocol is used to tie all vpe daemons together, and
exchange information about which vpe daemon serves which virtual
subnet.

The meta protocol consists of requests that can be sent to the other
side.  Each request has a unique number and several parameters.  All
requests are represented in the standard ASCII character set.  It is
possible to use tools such as telnet or netcat to connect to a vpe
daemon started with the --bypass-security option
and to read and write requests by hand, provided that one
understands the numeric codes sent.

The authentication scheme is described in @ref{Authentication protocol}. After a
successful authentication, the server and the client will exchange all the
information about other vpe daemons and subnets they know of, so that both
sides (and all the other vpe daemons behind them) have their information
synchronised.

@cindex ADD_EDGE
@cindex ADD_SUBNET
@example
daemon  message
--------------------------------------------------------------------------
origin  ADD_EDGE node1 node2 21.32.43.54 655 222 0
                   |     |        |       |   |  +-> options
                   |     |        |       |   +----> weight
                           |     |        |       +--------> UDP port of node2
                           |     |        +----------------> real address of node2
                           |     +-------------------------> name of destination node
                   +-------------------------------> name of source node

origin  ADD_SUBNET node 192.168.1.0/24
                     |         |     +--> prefixlength
                     |         +--------> network address
                     +------------------> owner of this subnet
--------------------------------------------------------------------------
@end example

The ADD_EDGE messages are to inform other vpe daemons that a connection between
two nodes exist. The address of the destination node is available so that
VPN packets can be sent directly to that node.

The ADD_SUBNET messages inform other vpe daemons that certain subnets belong
to certain nodes. vpe will use it to determine to which node a VPN packet has
to be sent.

@cindex DEL_EDGE
@cindex DEL_SUBNET
@example
message
------------------------------------------------------------------
DEL_EDGE node1 node2
                   |     +----> name of destination node
           +----------> name of source node

DEL_SUBNET node 192.168.1.0/24
             |         |     +--> prefixlength
             |         +--------> network address
             +------------------> owner of this subnet
------------------------------------------------------------------
@end example

In case a connection between two daemons is closed or broken, DEL_EDGE messages
are sent to inform the other daemons of that fact. Each daemon will calculate a
new route to the the daemons, or mark them unreachable if there isn't any.

@cindex REQ_KEY
@cindex ANS_KEY
@cindex KEY_CHANGED
@example
message
------------------------------------------------------------------
REQ_KEY origin destination
           |       +--> name of the vpe daemon it wants the key from
           +----------> name of the daemon that wants the key      

ANS_KEY origin destination 4ae0b0a82d6e0078 91 64 4
           |       |       \______________/ |  |  +--> MAC length
           |       |               |        |  +-----> digest algorithm
           |       |               |        +--------> cipher algorithm
           |       |               +--> 128 bits key
           |       +--> name of the daemon that wants the key
           +----------> name of the daemon that uses this key

KEY_CHANGED origin
              +--> daemon that has changed it's packet key
--------------------------------------------------------------------------
@end example

The keys used to encrypt VPN packets are not sent out directly. This is
because it would generate a lot of traffic on VPNs with many daemons, and
chances are that not every vpe daemon will ever send a packet to every
other daemon. Instead, if a daemon needs a key it sends a request for it
via the meta connection of the nearest hop in the direction of the
destination.

@cindex PING
@cindex PONG
@example
daemon  message
--------------------------------------------------------------------------
origin  PING
dest.   PONG
--------------------------------------------------------------------------
@end example

There is also a mechanism to check if hosts are still alive. Since network
failures or a crash can cause a daemon to be killed without properly
shutting down the TCP connection, this is necessary to keep an up to date
connection list. PINGs are sent at regular intervals, except when there
is also some other traffic. A little bit of salt (random data) is added
with each PING and PONG message, to make sure that long sequences of PING/PONG
messages without any other traffic won't result in known plaintext.

This basically covers what is sent over the meta connection by
vpe.


@c ==================================================================
@node    Security,  , The meta-protocol, Technical information
@section About vpe's encryption and other security-related issues.

@cindex TINC
@cindex Cabal
vpe got its name from ``TINC,'' short for @emph{There Is No Cabal}; the
alleged Cabal was/is an organisation that was said to keep an eye on the
entire Internet.  As this is exactly what you @emph{don't} want, we named
the vpe project after TINC.

@cindex SVPN
But in order to be ``immune'' to eavesdropping, you'll have to encrypt
your data.  Because vpe is a @emph{Secure} VPN (SVPN) daemon, it does
exactly that: encrypt.
vpe by default uses blowfish encryption with 128 bit keys in CBC mode, 32 bit
sequence numbers and 4 byte long message authentication codes to make sure
eavesdroppers cannot get and cannot change any information at all from the
packets they can intercept. The encryption algorithm and message authentication
algorithm can be changed in the configuration. The length of the message
authentication codes is also adjustable. The length of the key for the
encryption algorithm is always the default length used by OpenSSL.

@menu
* Authentication protocol::
* Encryption of network packets::
@end menu


@c ==================================================================
@node    Authentication protocol, Encryption of network packets, Security, Security
@subsection Authentication protocol

@cindex authentication
A new scheme for authentication in vpe has been devised, which offers some
improvements over the protocol used in 1.0pre2 and 1.0pre3. Explanation is
below.

@cindex ID
@cindex META_KEY
@cindex CHALLENGE
@cindex CHAL_REPLY
@cindex ACK
@example
daemon  message
--------------------------------------------------------------------------
client  <attempts connection>

server  <accepts connection>

client  ID client 12
              |   +---> version
              +-------> name of vpe daemon

server  ID server 12
              |   +---> version
              +-------> name of vpe daemon

client  META_KEY 5f0823a93e35b69e...7086ec7866ce582b
                 \_________________________________/
                                 +-> RSAKEYLEN bits totally random string S1,
                                     encrypted with server's public RSA key

server  META_KEY 6ab9c1640388f8f0...45d1a07f8a672630
                 \_________________________________/
                                 +-> RSAKEYLEN bits totally random string S2,
                                     encrypted with client's public RSA key

From now on:
 - the client will symmetrically encrypt outgoing traffic using S1
 - the server will symmetrically encrypt outgoing traffic using S2

client  CHALLENGE da02add1817c1920989ba6ae2a49cecbda0
                  \_________________________________/
                                 +-> CHALLEN bits totally random string H1

server  CHALLENGE 57fb4b2ccd70d6bb35a64c142f47e61d57f
                  \_________________________________/
                                 +-> CHALLEN bits totally random string H2

client  CHAL_REPLY 816a86
                      +-> 160 bits SHA1 of H2

server  CHAL_REPLY 928ffe
                      +-> 160 bits SHA1 of H1

After the correct challenge replies are received, both ends have proved
their identity. Further information is exchanged.

client  ACK 655 123 0
             |   |  +-> options
                 |   +----> estimated weight
                 +--------> listening port of client

server  ACK 655 321 0
             |   |  +-> options
                 |   +----> estimated weight
                 +--------> listening port of server
--------------------------------------------------------------------------
@end example

This new scheme has several improvements, both in efficiency and security.

First of all, the server sends exactly the same kind of messages over the wire
as the client. The previous versions of vpe first authenticated the client,
and then the server. This scheme even allows both sides to send their messages
simultaneously, there is no need to wait for the other to send something first.
This means that any calculations that need to be done upon sending or receiving
a message can also be done in parallel. This is especially important when doing
RSA encryption/decryption. Given that these calculations are the main part of
the CPU time spent for the authentication, speed is improved by a factor 2.

Second, only one RSA encrypted message is sent instead of two. This reduces the
amount of information attackers can see (and thus use for a cryptographic
attack). It also improves speed by a factor two, making the total speedup a
factor 4.

Third, and most important:
The symmetric cipher keys are exchanged first, the challenge is done
afterwards. In the previous authentication scheme, because a man-in-the-middle
could pass the challenge/chal_reply phase (by just copying the messages between
the two real vpe daemons), but no information was exchanged that was really
needed to read the rest of the messages, the challenge/chal_reply phase was of
no real use. The man-in-the-middle was only stopped by the fact that only after
the ACK messages were encrypted with the symmetric cipher. Potentially, it
could even send it's own symmetric key to the server (if it knew the server's
public key) and read some of the metadata the server would send it (it was
impossible for the mitm to read actual network packets though). The new scheme
however prevents this.

This new scheme makes sure that first of all, symmetric keys are exchanged. The
rest of the messages are then encrypted with the symmetric cipher. Then, each
side can only read received messages if they have their private key. The
challenge is there to let the other side know that the private key is really
known, because a challenge reply can only be sent back if the challenge is
decrypted correctly, and that can only be done with knowledge of the private
key.

Fourth: the first thing that is send via the symmetric cipher encrypted
connection is a totally random string, so that there is no known plaintext (for
an attacker) in the beginning of the encrypted stream.


@c ==================================================================
@node    Encryption of network packets,  , Authentication protocol, Security
@subsection Encryption of network packet
@cindex encryption

A data packet can only be sent if the encryption key is known to both
parties, and the connection is  activated. If the encryption key is not
known, a request is sent to the destination using the meta connection
to retrieve it. The packet is stored in a queue while waiting for the
key to arrive.

@cindex UDP
The UDP packet containing the network packet from the VPN has the following layout:

@example
... | IP header | UDP header | seqno | VPN packet | MAC | UDP trailer
                             \___________________/\_____/
                                       |             |
                                       V             +---> digest algorithm
                         Encrypted with symmetric cipher
@end example

So, the entire VPN packet is encrypted using a symmetric cipher, including a 32 bits
sequence number that is added in front of the actual VPN packet, to act as a unique
IV for each packet and to prevent replay attacks. A message authentication code
is added to the UDP packet to prevent alteration of packets. By default the
first 4 bytes of the digest are used for this, but this can be changed using
the MACLength configuration variable.

@c ==================================================================
@node    About us, Concept Index, Technical information, Top
@chapter About us


@menu
* Contact Information::
* Authors::
@end menu


@c ==================================================================
@node    Contact Information, Authors, About us, About us
@section Contact information

@cindex website
vpe's website is at @url{http://vpe.nl.linux.org/},
this server is located in the Netherlands.

@cindex IRC
We have an IRC channel on the FreeNode IRC network. Connect to
@uref{http://www.freenode.net/, irc.freenode.net}
and join channel #vpe.


@c ==================================================================
@node    Authors,  , Contact Information, About us
@section Authors

@table @asis
@item Ivo Timmermans (zarq) (@email{ivo@@o2w.nl})
@item Guus Sliepen (guus) (@email{guus@@sliepen.eu.org})
@end table

We have received a lot of valuable input from users.  With their help,
vpe has become the flexible and robust tool that it is today.  We have
composed a list of contributions, in the file called @file{THANKS} in
the source distribution.


@c ==================================================================
@node    Concept Index,  , About us, Top
@c        node-name,    next, previous,        up
@unnumbered Concept Index

@c ==================================================================
@printindex cp


@c ==================================================================
@contents
@bye
Revision:	1.2
Committed:	Fri Mar 28 19:46:47 2003 UTC (21 years, 2 months ago) by pcg
Content type:	application/x-texinfo
Branch:	MAIN
CVS Tags:	HEAD
Changes since 1.1:	+0 -0 lines
State:	*FILE REMOVED*
Log Message:	* empty log message *