1 \input texinfo @c -*-texinfo-*-
2 @c $Id: tinc.texi,v 1.8.4.41 2003/08/02 21:55:12 guus Exp $
10 @dircategory Networking tools
12 * tinc: (tinc). The tinc Manual.
15 This is the info manual for tinc, a Virtual Private Network daemon.
17 Copyright @copyright{} 1998-2003 Ivo Timmermans
18 <ivo@@o2w.nl>, Guus Sliepen <guus@@sliepen.eu.org> and
19 Wessel Dankers <wsl@@nl.linux.org>.
21 $Id: tinc.texi,v 1.8.4.41 2003/08/02 21:55:12 guus Exp $
23 Permission is granted to make and distribute verbatim copies of this
24 manual provided the copyright notice and this permission notice are
25 preserved on all copies.
27 Permission is granted to copy and distribute modified versions of this
28 manual under the conditions for verbatim copying, provided that the
29 entire resulting derived work is distributed under the terms of a
30 permission notice identical to this one.
36 @subtitle Setting up a Virtual Private Network with tinc
37 @author Ivo Timmermans and Guus Sliepen
40 @vskip 0pt plus 1filll
42 Copyright @copyright{} 1998-2003 Ivo Timmermans
43 <ivo@@o2w.nl>, Guus Sliepen <guus@@sliepen.eu.org> and
44 Wessel Dankers <wsl@@nl.linux.org>.
46 $Id: tinc.texi,v 1.8.4.41 2003/08/02 21:55:12 guus Exp $
48 Permission is granted to make and distribute verbatim copies of this
49 manual provided the copyright notice and this permission notice are
50 preserved on all copies.
52 Permission is granted to copy and distribute modified versions of this
53 manual under the conditions for verbatim copying, provided that the
54 entire resulting derived work is distributed under the terms of a
55 permission notice identical to this one.
59 @c ==================================================================
60 @node Top, Introduction, (dir), (dir)
63 * Introduction:: Introduction
68 * Technical information::
70 * Concept Index:: All used terms explained
76 @c ==================================================================
77 @node Introduction, Preparations, Top, Top
81 tinc is a Virtual Private Network (VPN) daemon that uses tunneling and
82 encryption to create a secure private network between hosts on the
85 Because the tunnel appears to the IP level network code as a normal
86 network device, there is no need to adapt any existing software.
87 The encrypted tunnels allows VPN sites to share information with each other
88 over the Internet without exposing any information to others.
90 This document is the manual for tinc. Included are chapters on how to
91 configure your computer to use tinc, as well as the configuration
92 process of tinc itself.
95 * VPNs:: Virtual Private Networks in general
97 * Supported platforms::
100 @c ==================================================================
101 @node VPNs, tinc, Introduction, Introduction
102 @section Virtual Private Networks
105 A Virtual Private Network or VPN is a network that can only be accessed
106 by a few elected computers that participate. This goal is achievable in
107 more than just one way.
110 Private networks can consist of a single stand-alone Ethernet LAN. Or
111 even two computers hooked up using a null-modem cable. In these cases,
113 obvious that the network is @emph{private}, no one can access it from the
114 outside. But if your computers are linked to the Internet, the network
115 is not private anymore, unless one uses firewalls to block all private
116 traffic. But then, there is no way to send private data to trusted
117 computers on the other end of the Internet.
120 This problem can be solved by using @emph{virtual} networks. Virtual
121 networks can live on top of other networks, but they use encapsulation to
122 keep using their private address space so they do not interfere with
123 the Internet. Mostly, virtual networks appear like a singe LAN, even though
124 they can span the entire world. But virtual networks can't be secured
125 by using firewalls, because the traffic that flows through it has to go
126 through the Internet, where other people can look at it.
128 As is the case with either type of VPN, anybody could eavesdrop. Or
129 worse, alter data. Hence it's probably advisable to encrypt the data
130 that flows over the network.
132 When one introduces encryption, we can form a true VPN. Other people may
133 see encrypted traffic, but if they don't know how to decipher it (they
134 need to know the key for that), they cannot read the information that flows
135 through the VPN. This is what tinc was made for.
138 @c ==================================================================
139 @node tinc, Supported platforms, VPNs, Introduction
143 I really don't quite remember what got us started, but it must have been
144 Guus' idea. He wrote a simple implementation (about 50 lines of C) that
145 used the ethertap device that Linux knows of since somewhere
146 about kernel 2.1.60. It didn't work immediately and he improved it a
147 bit. At this stage, the project was still simply called @samp{vpnd}.
149 Since then, a lot has changed---to say the least.
152 tinc now supports encryption, it consists of a single daemon (tincd) for
153 both the receiving and sending end, it has become largely
154 runtime-configurable---in short, it has become a full-fledged
155 professional package.
157 @cindex Traditional VPNs
159 tinc also allows more than two sites to connect to eachother and form a single VPN.
160 Traditionally VPNs are created by making tunnels, which only have two endpoints.
161 Larger VPNs with more sites are created by adding more tunnels.
162 tinc takes another approach: only endpoints are specified,
163 the software itself will take care of creating the tunnels.
164 This allows for easier configuration and improved scalability.
166 A lot can---and will be---changed. We have a number of things that we would like to
167 see in the future releases of tinc. Not everything will be available in
168 the near future. Our first objective is to make tinc work perfectly as
169 it stands, and then add more advanced features.
171 Meanwhile, we're always open-minded towards new ideas. And we're
175 @c ==================================================================
176 @node Supported platforms, , tinc, Introduction
177 @section Supported platforms
180 tinc has been verified to work under Linux, FreeBSD, OpenBSD, NetBSD, MacOS/X (Darwin), Solaris, and Windows (in a Cygwin environment),
181 with various hardware architectures. These are some of the platforms
182 that are supported by the universal tun/tap device driver or other virtual network device drivers.
183 Without such a driver, tinc will most
184 likely compile and run, but it will not be able to send or receive data
188 For an up to date list of supported platforms, please check the list on
190 @uref{http://tinc.nl.linux.org/platforms}.
193 @c ==================================================================
197 tinc was first written for Linux running on an intel x86 processor, so
198 this is the best supported platform. The protocol however, and actually
199 anything about tinc, has been rewritten to support random byte ordering
200 and arbitrary word length. So in theory it should run on other
201 processors that Linux runs on. It has already been verified to run on
202 alpha and sparc processors as well.
204 tinc uses the ethertap device or the universal tun/tap driver. The former is provided in the standard kernel
205 from version 2.1.60 up to 2.3.x, but has been replaced in favour of the tun/tap driver in kernel versions 2.4.0 and later.
208 @c ==================================================================
212 tinc on FreeBSD relies on the universal tun/tap driver for its data
213 acquisition from the kernel. Therefore, tinc will work on the same platforms
214 as this driver. These are: FreeBSD 3.x, 4.x, 5.x.
217 @c ==================================================================
221 tinc on OpenBSD relies on the tun driver for its data
222 acquisition from the kernel. It has been verified to work under at least OpenBSD 2.9.
224 Tunneling IPv6 packets may not work on OpenBSD.
227 @c ==================================================================
230 @c ==================================================================
234 tinc on NetBSD relies on the tun driver for its data
235 acquisition from the kernel. It has been verified to work under at least NetBSD 1.5.2.
237 Tunneling IPv6 does not work on OpenBSD.
240 @c ==================================================================
244 tinc on Solaris relies on the universal tun/tap driver for its data
245 acquisition from the kernel. Therefore, tinc will work on the same platforms
246 as this driver. These are: Solaris 8 (SunOS 5.8).
248 IPv6 packets cannot be tunneled on Solaris.
250 @c ==================================================================
251 @subsection Darwin (MacOS/X)
255 tinc on Darwin relies on the tunnel driver for its data
256 acquisition from the kernel. This driver is not part of Darwin but can be
257 downloaded from @uref{http://chrisp.de/en/projects/tunnel.html}.
259 IPv6 packets cannot be tunneled on Darwin.
261 @c ==================================================================
262 @subsection Cygwin (Windows)
266 tinc on Windows, in a Cygwin environment, relies on the CIPE driver for its data
267 acquisition from the kernel. This driver is not part of Windows but can be
268 downloaded from @uref{http://cipe-win32.sourceforge.net/}.
270 @c ==================================================================
271 @subsection MinGW (Windows)
275 tinc on Windows (native), compiled using MinGW, relies on the CIPE driver for its data
276 acquisition from the kernel. This driver is not part of Windows but can be
277 downloaded from @uref{http://cipe-win32.sourceforge.net/}.
286 @c Preparing your system
293 @c ==================================================================
294 @node Preparations, Installation, Introduction, Top
295 @chapter Preparations
297 This chapter contains information on how to prepare your system to
301 * Configuring the kernel::
306 @c ==================================================================
307 @node Configuring the kernel, Libraries, Preparations, Preparations
308 @section Configuring the kernel
315 If you are running Linux, chances are good that your kernel already supports
316 all the devices that tinc needs for proper operation. For example, the
317 standard kernel from Redhat Linux already has support for ethertap and netlink
318 compiled in. Debian users can use the modconf utility to select the modules.
319 If your Linux distribution supports this method of selecting devices, look out
320 for something called `ethertap', and `netlink_dev' if it is using a kernel
321 version prior to 2.4.0. In that case you will need both these devices. If you
322 are using kernel 2.4.0 or later, you need to select `tun'.
325 If you can install these devices in a similar manner, you may skip this section.
326 Otherwise, you will have to recompile the kernel in order to turn on the required features.
327 If you are unfamiliar with the process of configuring and compiling a new kernel,
328 you should read the @uref{http://howto.linuxberg.com/LDP/HOWTO/Kernel-HOWTO.html, Kernel HOWTO} first.
331 * Configuration of Linux kernels 2.1.60 up to 2.4.0::
332 * Configuration of Linux kernels 2.4.0 and higher::
333 * Configuration of FreeBSD kernels::
334 * Configuration of OpenBSD kernels::
335 * Configuration of NetBSD kernels::
336 * Configuration of Solaris kernels::
337 * Configuration of Darwin (MacOS/X) kernels::
338 * Configuration of Cygwin (Windows)::
339 * Configuration of MinGW (Windows)::
343 @c ==================================================================
344 @node Configuration of Linux kernels 2.1.60 up to 2.4.0, Configuration of Linux kernels 2.4.0 and higher, Configuring the kernel, Configuring the kernel
345 @subsection Configuration of Linux kernels 2.1.60 up to 2.4.0
347 Here are the options you have to turn on when configuring a new kernel:
350 Code maturity level options
351 [*] Prompt for development and/or incomplete code/drivers
353 [*] Kernel/User netlink socket
354 <M> Netlink device emulation
355 Network device support
356 <M> Ethertap network tap
359 If you want to run more than one instance of tinc or other programs that use
360 the ethertap, you have to compile the ethertap driver as a module, otherwise
361 you can also choose to compile it directly into the kernel.
363 If you decide to build any of these as dynamic kernel modules, it's a good idea
364 to add these lines to @file{/etc/modules.conf}:
367 alias char-major-36 netlink_dev
369 options tap0 -o tap0 unit=0
371 options tap1 -o tap1 unit=1
373 alias tap@emph{N} ethertap
374 options tap@emph{N} -o tap@emph{N} unit=@emph{N}
377 Add as much alias/options lines as necessary.
380 @c ==================================================================
381 @node Configuration of Linux kernels 2.4.0 and higher, Configuration of FreeBSD kernels, Configuration of Linux kernels 2.1.60 up to 2.4.0, Configuring the kernel
382 @subsection Configuration of Linux kernels 2.4.0 and higher
384 Here are the options you have to turn on when configuring a new kernel:
387 Code maturity level options
388 [*] Prompt for development and/or incomplete code/drivers
389 Network device support
390 <M> Universal tun/tap device driver support
393 It's not necessary to compile this driver as a module, even if you are going to
394 run more than one instance of tinc.
396 If you have an early 2.4 kernel, you can choose both the tun/tap driver and the
397 `Ethertap network tap' device. This latter is marked obsolete, and chances are
398 that it won't even function correctly anymore. Make sure you select the
399 universal tun/tap driver.
401 If you decide to build the tun/tap driver as a kernel module, add these lines
402 to @file{/etc/modules.conf}:
405 alias char-major-10-200 tun
409 @c ==================================================================
410 @node Configuration of FreeBSD kernels, Configuration of OpenBSD kernels, Configuration of Linux kernels 2.4.0 and higher, Configuring the kernel
411 @subsection Configuration of FreeBSD kernels
413 For FreeBSD version 4.1 and higher, the tap driver is included in the default kernel configuration, for earlier
414 systems (4.0 and earlier), you need to install the universal tun/tap driver
418 @c ==================================================================
419 @node Configuration of OpenBSD kernels, Configuration of NetBSD kernels, Configuration of FreeBSD kernels, Configuring the kernel
420 @subsection Configuration of OpenBSD kernels
422 For OpenBSD version 2.9 and higher,
423 the tun driver is included in the default kernel configuration.
426 @c ==================================================================
427 @node Configuration of NetBSD kernels, Configuration of Solaris kernels, Configuration of OpenBSD kernels, Configuring the kernel
428 @subsection Configuration of NetBSD kernels
430 For NetBSD version 1.5.2 and higher,
431 the tun driver is included in the default kernel configuration.
434 @c ==================================================================
435 @node Configuration of Solaris kernels, Configuration of Darwin (MacOS/X) kernels, Configuration of NetBSD kernels, Configuring the kernel
436 @subsection Configuration of Solaris kernels
438 For Solaris 8 (SunOS 5.8) and higher,
439 the tun driver is included in the default kernel configuration.
442 @c ==================================================================
443 @node Configuration of Darwin (MacOS/X) kernels, Configuration of Cygwin (Windows), Configuration of Solaris kernels, Configuring the kernel
444 @subsection Configuration of Darwin (MacOS/X) kernels
446 Darwin does not come with a tunnel driver. You must download it at
447 @uref{http://chrisp.de/en/projects/tunnel.html}. If compiling the source fails,
448 try the binary module. The tunnel driver must be loaded before starting tinc
449 with the following command:
455 Once loaded, the tunnel driver will automatically create @file{/dev/tun0}..@file{/dev/tun3}
456 and the corresponding network interfaces.
459 @c ==================================================================
460 @node Configuration of Cygwin (Windows), Configuration of MinGW (Windows), Configuration of Darwin (MacOS/X) kernels, Configuring the kernel
461 @subsection Configuration of Cygwin (Windows)
463 You will need to install the CIPE driver, you can download it from
464 @uref{http://cipe-win32.sourceforge.net}. Configure the CIPE network device in
465 the same way as you would do from the tinc-up script.
468 @c ==================================================================
469 @node Configuration of MinGW (Windows), , Configuration of Cygwin (Windows), Configuring the kernel
470 @subsection Configuration of MinGW (Windows)
472 You will need to install the CIPE driver, you can download it from
473 @uref{http://cipe-win32.sourceforge.net}. Configure the CIPE network device in
474 the same way as you would do from the tinc-up script.
477 @c ==================================================================
478 @node Libraries, , Configuring the kernel, Preparations
483 Before you can configure or build tinc, you need to have the OpenSSL,
484 zlib and lzo libraries installed on your system. If you try to configure tinc without
485 having them installed, configure will give you an error message, and stop.
494 @c ==================================================================
495 @node OpenSSL, zlib, Libraries, Libraries
499 For all cryptography-related functions, tinc uses the functions provided
500 by the OpenSSL library.
502 If this library is not installed, you wil get an error when configuring
503 tinc for build. Support for running tinc without having OpenSSL
504 installed @emph{may} be added in the future.
506 You can use your operating system's package manager to install this if
507 available. Make sure you install the development AND runtime versions
510 If you have to install OpenSSL manually, you can get the source code
511 from @url{http://www.openssl.org/}. Instructions on how to configure,
512 build and install this package are included within the package. Please
513 make sure you build development and runtime libraries (which is the
516 If you installed the OpenSSL libraries from source, it may be necessary
517 to let configure know where they are, by passing configure one of the
518 --with-openssl-* parameters.
521 --with-openssl=DIR OpenSSL library and headers prefix
522 --with-openssl-include=DIR OpenSSL headers directory
523 (Default is OPENSSL_DIR/include)
524 --with-openssl-lib=DIR OpenSSL library directory
525 (Default is OPENSSL_DIR/lib)
529 @subsubheading License
532 Since the license under which OpenSSL is distributed is not directly
533 compatible with the terms of the GNU GPL
534 @uref{http://www.openssl.org/support/faq.html#LEGAL2}, therefore we
535 include an addition to the GPL (see also the file COPYING.README):
538 This program is released under the GPL with the additional exemption
539 that compiling, linking, and/or using OpenSSL is allowed. You may
540 provide binary packages linked to the OpenSSL libraries, provided that
541 all other requirements of the GPL are met.
545 @c ==================================================================
546 @node zlib, lzo, OpenSSL, Libraries
550 For the optional compression of UDP packets, tinc uses the functions provided
553 If this library is not installed, you wil get an error when configuring
554 tinc for build. Support for running tinc without having zlib
555 installed @emph{may} be added in the future.
557 You can use your operating system's package manager to install this if
558 available. Make sure you install the development AND runtime versions
561 If you have to install zlib manually, you can get the source code
562 from @url{http://www.gzip.org/zlib/}. Instructions on how to configure,
563 build and install this package are included within the package. Please
564 make sure you build development and runtime libraries (which is the
568 @c ==================================================================
569 @node lzo, , zlib, Libraries
573 Another form of compression is offered using the lzo library.
575 If this library is not installed, you wil get an error when configuring
576 tinc for build. Support for running tinc without having lzo
577 installed @emph{may} be added in the future.
579 You can use your operating system's package manager to install this if
580 available. Make sure you install the development AND runtime versions
583 If you have to install lzo manually, you can get the source code
584 from @url{http://www.oberhumer.com/opensource/lzo/}. Instructions on how to configure,
585 build and install this package are included within the package. Please
586 make sure you build development and runtime libraries (which is the
599 @c ==================================================================
600 @node Installation, Configuration, Preparations, Top
601 @chapter Installation
603 If you use Debian, you may want to install one of the
604 precompiled packages for your system. These packages are equipped with
605 system startup scripts and sample configurations.
607 If you cannot use one of the precompiled packages, or you want to compile tinc
608 for yourself, you can use the source. The source is distributed under
609 the GNU General Public License (GPL). Download the source from the
610 @uref{http://tinc.nl.linux.org/download, download page}, which has
611 the checksums of these files listed; you may wish to check these with
612 md5sum before continuing.
614 tinc comes in a convenient autoconf/automake package, which you can just
615 treat the same as any other package. Which is just untar it, type
616 `./configure' and then `make'.
617 More detailed instructions are in the file @file{INSTALL}, which is
618 included in the source distribution.
621 * Building and installing tinc::
626 @c ==================================================================
627 @node Building and installing tinc, System files, Installation, Installation
628 @section Building and installing tinc
630 Detailed instructions on configuring the source, building tinc and installing tinc
631 can be found in the file called @file{INSTALL}.
633 @cindex binary package
634 If you happen to have a binary package for tinc for your distribution,
635 you can use the package management tools of that distribution to install tinc.
636 The documentation that comes along with your distribution will tell you how to do that.
639 * Darwin (MacOS/X) build environment::
640 * Cygwin (Windows) build environment::
641 * MinGW (Windows) build environment::
645 @c ==================================================================
646 @node Darwin (MacOS/X) build environment, Cygwin (Windows) build environment, , Building and installing tinc
647 @subsection Darwin (MacOS/X) build environment
649 In order to build tinc on Darwin, you need to install the MacOS/X Developer Tools
650 from @uref{http://developer.apple.com/tools/macosxtools.html} and
651 a recent version of Fink from @uref{http://fink.sourceforge.net/}.
653 After installation use fink to download and install the following packages:
654 autoconf25, automake, dlcompat, m4, openssl, zlib and lzo.
656 @c ==================================================================
657 @node Cygwin (Windows) build environment, MinGW (Windows) build environment, Darwin (MacOS/X) build environment, Building and installing tinc
658 @subsection Cygwin (Windows) build environment
660 If Cygwin hasn't already been installed, install it directly from
661 @uref{http://www.cygwin.com/}.
663 When tinc is compiled in a Cygwin environment, it can only be run in this environment,
664 but all programs, including those started outside the Cygwin environment, will be able to use the VPN.
665 It will also support all features.
667 @c ==================================================================
668 @node MinGW (Windows) build environment, , Cygwin (Windows) build environment, Building and installing tinc
669 @subsection MinGW (Windows) build environment
671 You will need to install the MinGW environment from @uref{http://www.mingw.org}.
673 When tinc is compiled using MinGW it runs natively under Windows,
674 it is not necessary to keep MinGW installed.
676 When running natively, tinc is not able to start scripts,
677 nor is tinc able to receive signals.
678 When detaching, tinc will install itself as a service,
679 which will be restarted automatically after reboots.
682 @c ==================================================================
683 @node System files, , Building and installing tinc, Installation
684 @section System files
686 Before you can run tinc, you must make sure you have all the needed
687 files on your system.
695 @c ==================================================================
696 @node Device files, Other files, System files, System files
697 @subsection Device files
700 First, you'll need the special device file(s) that form the interface
701 between the kernel and the daemon.
703 The permissions for these files have to be such that only the super user
704 may read/write to this file. You'd want this, because otherwise
705 eavesdropping would become a bit too easy. This does, however, imply
706 that you'd have to run tincd as root.
708 If you use Linux and have a kernel version prior to 2.4.0, you have to make the
712 mknod -m 600 /dev/tap0 c 36 16
713 mknod -m 600 /dev/tap1 c 36 17
715 mknod -m 600 /dev/tap@emph{N} c 36 @emph{N+16}
718 There is a maximum of 16 ethertap devices.
720 If you use the universal tun/tap driver, you have to create the
721 following device file (unless it already exist):
724 mknod -m 600 /dev/tun c 10 200
727 If you use Linux, and you run the new 2.4 kernel using the devfs filesystem,
728 then the tun/tap device will probably be automatically generated as
731 Unlike the ethertap device, you do not need multiple device files if
732 you are planning to run multiple tinc daemons.
735 @c ==================================================================
736 @node Other files, , Device files, System files
737 @subsection Other files
739 @subsubheading @file{/etc/networks}
741 You may add a line to @file{/etc/networks} so that your VPN will get a
742 symbolic name. For example:
748 @subsubheading @file{/etc/services}
751 You may add this line to @file{/etc/services}. The effect is that you
752 may supply a @samp{tinc} as a valid port number to some programs. The
753 number 655 is registered with the IANA.
758 # Ivo Timmermans <ivo@@o2w.nl>
773 @c ==================================================================
774 @node Configuration, Running tinc, Installation, Top
775 @chapter Configuration
778 * Configuration introduction::
779 * Multiple networks::
780 * How connections work::
781 * Configuration files::
782 * Generating keypairs::
783 * Network interfaces::
784 * Example configuration::
787 @c ==================================================================
788 @node Configuration introduction, Multiple networks, Configuration, Configuration
789 @section Configuration introduction
791 @cindex Network Administrators Guide
792 Before actually starting to configure tinc and editing files,
793 make sure you have read this entire section so you know what to expect.
794 Then, make it clear to yourself how you want to organize your VPN:
795 What are the nodes (computers running tinc)?
796 What IP addresses/subnets do they have?
797 What is the network mask of the entire VPN?
798 Do you need special firewall rules?
799 Do you have to set up masquerading or forwarding rules?
800 These questions can only be answered by yourself,
801 you will not find the answers in this documentation.
802 Make sure you have an adequate understanding of networks in general.
803 A good resource on networking is the
804 @uref{http://www.linuxdoc.org/LDP/nag2/, Linux Network Administrators Guide}.
806 If you have everything clearly pictured in your mind,
807 proceed in the following order:
808 First, generate the configuration files (@file{tinc.conf}, your host configuration file, @file{tinc-up} and perhaps @file{tinc-down}).
809 Then generate the keypairs.
810 Finally, distribute the host configuration files.
811 These steps are described in the subsections below.
814 @c ==================================================================
815 @node Multiple networks, How connections work, Configuration introduction, Configuration
816 @section Multiple networks
818 @cindex multiple networks
820 In order to allow you to run more than one tinc daemon on one computer,
821 for instance if your computer is part of more than one VPN,
822 you can assign a ``netname'' to your VPN.
823 It is not required if you only run one tinc daemon,
824 it doesn't even have to be the same on all the sites of your VPN,
825 but it is recommended that you choose one anyway.
827 We will asume you use a netname throughout this document.
828 This means that you call tincd with the -n argument,
829 which will assign a netname to this daemon.
831 The effect of this is that the daemon will set its configuration
832 ``root'' to /etc/tinc/netname/, where netname is your argument to the -n
833 option. You'll notice that it appears in syslog as ``tinc.netname''.
835 However, it is not strictly necessary that you call tinc with the -n
836 option. In this case, the network name would just be empty, and it will
837 be used as such. tinc now looks for files in /etc/tinc/, instead of
838 /etc/tinc/netname/; the configuration file should be /etc/tinc/tinc.conf,
839 and the host configuration files are now expected to be in /etc/tinc/hosts/.
841 But it is highly recommended that you use this feature of tinc, because
842 it will be so much clearer whom your daemon talks to. Hence, we will
843 assume that you use it.
846 @c ==================================================================
847 @node How connections work, Configuration files, Multiple networks, Configuration
848 @section How connections work
850 When tinc starts up, it parses the command-line options and then
851 reads in the configuration file tinc.conf.
852 If it sees one or more `ConnectTo' values pointing to other tinc daemons in that file,
853 it will try to connect to those other daemons.
854 Whether this succeeds or not and whether `ConnectTo' is specified or not,
855 tinc will listen for incoming connection from other deamons.
856 If you did specify a `ConnectTo' value and the other side is not responding,
857 tinc will keep retrying.
858 This means that once started, tinc will stay running until you tell it to stop,
859 and failures to connect to other tinc daemons will not stop your tinc daemon
860 for trying again later.
861 This means you don't have to intervene if there are temporary network problems.
865 There is no real distinction between a server and a client in tinc.
866 If you wish, you can view a tinc daemon without a `ConnectTo' value as a server,
867 and one which does specify such a value as a client.
868 It does not matter if two tinc daemons have a `ConnectTo' value pointing to each other however.
871 @c ==================================================================
872 @node Configuration files, Generating keypairs, How connections work, Configuration
873 @section Configuration files
875 The actual configuration of the daemon is done in the file
876 @file{/etc/tinc/netname/tinc.conf} and at least one other file in the directory
877 @file{/etc/tinc/netname/hosts/}.
879 These file consists of comments (lines started with a #) or assignments
886 The variable names are case insensitive, and any spaces, tabs, newlines
887 and carriage returns are ignored. Note: it is not required that you put
888 in the `=' sign, but doing so improves readability. If you leave it
889 out, remember to replace it with at least one space character.
891 In this section all valid variables are listed in alphabetical order.
892 The default value is given between parentheses,
893 other comments are between square brackets and
894 required directives are given in @strong{bold}.
897 * Main configuration variables::
898 * Host configuration variables::
903 @c ==================================================================
904 @node Main configuration variables, Host configuration variables, Configuration files, Configuration files
905 @subsection Main configuration variables
908 @cindex AddressFamily
909 @item AddressFamily = <ipv4|ipv6|any> (any)
910 This option affects the address family of listening and outgoing sockets.
911 If "any" is selected, then depending on the operating system
912 both IPv4 and IPv6 or just IPv6 listening sockets will be created.
914 @cindex BindToAddress
915 @item BindToAddress = <address> [experimental]
916 If your computer has more than one IPv4 or IPv6 address, tinc
917 will by default listen on all of them for incoming connections.
918 It is possible to bind only to a single address with this variable.
920 This option may not work on all platforms.
922 @cindex BindToInterface
923 @item BindToInterface = <interface> [experimental]
924 If you have more than one network interface in your computer, tinc will
925 by default listen on all of them for incoming connections. It is
926 possible to bind tinc to a single interface like eth0 or ppp0 with this
929 This option may not work on all platforms.
932 @item @strong{ConnectTo = <name>}
933 Specifies which other tinc daemon to connect to on startup.
934 Multiple ConnectTo variables may be specified,
935 in which case outgoing connections to each specified tinc daemon are made.
936 The names should be known to this tinc daemon
937 (i.e., there should be a host configuration file for the name on the ConnectTo line).
939 If you don't specify a host with ConnectTo,
940 tinc won't try to connect to other daemons at all,
941 and will instead just listen for incoming connections.
944 @item @strong{Device = <device>} (/dev/tap0 or /dev/net/tun)
945 The virtual network device to use. Note that you can only use one device per
946 daemon. See also @ref{Device files}.
949 @item Hostnames = <yes|no> (no)
950 This option selects whether IP addresses (both real and on the VPN)
951 should be resolved. Since DNS lookups are blocking, it might affect
952 tinc's efficiency, even stopping the daemon for a few seconds everytime
953 it does a lookup if your DNS server is not responding.
955 This does not affect resolving hostnames to IP addresses from the
959 @item Interface = <interface>
960 Defines the name of the interface corresponding to the virtual network device.
961 Depending on the operating system and the type of device this may or may not actually set the name of the interface
962 or choose the device corresponding to this interface.
965 @item Mode = <router|switch|hub> (router)
966 This option selects the way packets are routed to other daemons.
972 variables in the host configuration files will be used to form a routing table.
973 Only unicast packets of routable protocols (IPv4 and IPv6) are supported in this mode.
975 This is the default mode, and unless you really know you need another mode, don't change it.
979 In this mode the MAC addresses of the packets on the VPN will be used to
980 dynamically create a routing table just like an Ethernet switch does.
981 Unicast, multicast and broadcast packets of every protocol that runs over Ethernet are supported in this mode
982 at the cost of frequent broadcast ARP requests and routing table updates.
984 This mode is primarily useful if you want to bridge Ethernet segments.
988 This mode is almost the same as the switch mode, but instead
989 every packet will be broadcast to the other daemons
990 while no routing table is managed.
994 @item KeyExpire = <seconds> (3600)
995 This option controls the time the encryption keys used to encrypt the data
996 are valid. It is common practice to change keys at regular intervals to
997 make it even harder for crackers, even though it is thought to be nearly
998 impossible to crack a single key.
1001 @item MACExpire = <seconds> (600)
1002 This option controls the amount of time MAC addresses are kept before they are removed.
1003 This only has effect when Mode is set to "switch".
1006 @item @strong{Name = <name>}
1007 This is a symbolic name for this connection. It can be anything
1010 @item PingTimeout = <seconds> (60)
1011 The number of seconds of inactivity that tinc will wait before sending a
1012 probe to the other end. If that other end doesn't answer within that
1013 same amount of seconds, the connection is terminated, and the others
1014 will be notified of this.
1016 @cindex PriorityInheritance
1017 @item PriorityInheritance = <yes|no> (no) [experimental]
1018 When this option is enabled the value of the TOS field of tunneled IPv4 packets
1019 will be inherited by the UDP packets that are sent out.
1022 @item PrivateKey = <key> [obsolete]
1023 This is the RSA private key for tinc. However, for safety reasons it is
1024 advised to store private keys of any kind in separate files. This prevents
1025 accidental eavesdropping if you are editting the configuration file.
1027 @cindex PrivateKeyFile
1028 @item @strong{PrivateKeyFile = <path>} [recommended]
1029 This is the full path name of the RSA private key file that was
1030 generated by ``tincd --generate-keys''. It must be a full path, not a
1033 Note that there must be exactly one of PrivateKey
1035 specified in the configuration file.
1040 @c ==================================================================
1041 @node Host configuration variables, How to configure, Main configuration variables, Configuration files
1042 @subsection Host configuration variables
1046 @item @strong{Address = <IP address|hostname>} [recommended]
1047 This variable is only required if you want to connect to this host. It
1048 must resolve to the external IP address where the host can be reached,
1049 not the one that is internal to the VPN.
1052 @item Cipher = <cipher> (blowfish)
1053 The symmetric cipher algorithm used to encrypt UDP packets.
1054 Any cipher supported by OpenSSL is recognized.
1057 @item Compression = <level> (0)
1058 This option sets the level of compression used for UDP packets.
1059 Possible values are 0 (off), 1 (fast zlib) and any integer up to 9 (best zlib),
1060 10 (fast lzo) and 11 (best lzo).
1063 @item Digest = <digest> (sha1)
1064 The digest algorithm used to authenticate UDP packets.
1065 Any digest supported by OpenSSL is recognized.
1066 Furthermore, specifying "none" will turn off packet authentication.
1068 @cindex IndirectData
1069 @item IndirectData = <yes|no> (no)
1070 This option specifies whether other tinc daemons besides the one you
1071 specified with ConnectTo can make a direct connection to you. This is
1072 especially useful if you are behind a firewall and it is impossible to
1073 make a connection from the outside to your tinc daemon. Otherwise, it
1074 is best to leave this option out or set it to no.
1077 @item MACLength = <length> (4)
1078 The length of the message authentication code used to authenticate UDP packets.
1079 Can be anything from 0
1080 up to the length of the digest produced by the digest algorithm.
1083 @item Port = <port> (655)
1084 This is the port this tinc daemon listens on.
1085 You can use decimal portnumbers or symbolic names (as listed in /etc/services).
1088 @item PublicKey = <key> [obsolete]
1089 This is the RSA public key for this host.
1091 @cindex PublicKeyFile
1092 @item PublicKeyFile = <path> [obsolete]
1093 This is the full path name of the RSA public key file that was generated
1094 by ``tincd --generate-keys''. It must be a full path, not a relative
1098 From version 1.0pre4 on tinc will store the public key directly into the
1099 host configuration file in PEM format, the above two options then are not
1100 necessary. Either the PEM format is used, or exactly
1101 @strong{one of the above two options} must be specified
1102 in each host configuration file, if you want to be able to establish a
1103 connection with that host.
1106 @item Subnet = <address[/prefixlength]>
1107 The subnet which this tinc daemon will serve.
1108 tinc tries to look up which other daemon it should send a packet to by searching the appropiate subnet.
1109 If the packet matches a subnet,
1110 it will be sent to the daemon who has this subnet in his host configuration file.
1111 Multiple subnet lines can be specified for each daemon.
1113 Subnets can either be single MAC, IPv4 or IPv6 addresses,
1114 in which case a subnet consisting of only that single address is assumed,
1115 or they can be a IPv4 or IPv6 network address with a prefixlength.
1116 Shorthand notations are not supported.
1117 For example, IPv4 subnets must be in a form like 192.168.1.0/24,
1118 where 192.168.1.0 is the network address and 24 is the number of bits set in the netmask.
1119 Note that subnets like 192.168.1.1/24 are invalid!
1120 Read a networking HOWTO/FAQ/guide if you don't understand this.
1121 IPv6 subnets are notated like fec0:0:0:1:0:0:0:0/64.
1122 MAC addresses are notated like 0:1a:2b:3c:4d:5e.
1124 @cindex CIDR notation
1125 prefixlength is the number of bits set to 1 in the netmask part; for
1126 example: netmask 255.255.255.0 would become /24, 255.255.252.0 becomes
1127 /22. This conforms to standard CIDR notation as described in
1128 @uref{ftp://ftp.isi.edu/in-notes/rfc1519.txt, RFC1519}
1131 @item TCPonly = <yes|no> (no) [experimental]
1132 If this variable is set to yes, then the packets are tunnelled over a
1133 TCP connection instead of a UDP connection. This is especially useful
1134 for those who want to run a tinc daemon from behind a masquerading
1135 firewall, or if UDP packet routing is disabled somehow.
1136 Setting this options also implicitly sets IndirectData.
1140 @c ==================================================================
1141 @node How to configure, , Host configuration variables, Configuration files
1142 @subsection How to configure
1144 @subsubheading Step 1. Creating the main configuration file
1146 The main configuration file will be called @file{/etc/tinc/netname/tinc.conf}.
1147 Adapt the following example to create a basic configuration file:
1150 Name = @emph{yourname}
1151 Device = @emph{/dev/tap0}
1152 PrivateKeyFile = /etc/tinc/@emph{netname}/rsa_key.priv
1155 Then, if you know to which other tinc daemon(s) yours is going to connect,
1156 add `ConnectTo' values.
1158 @subsubheading Step 2. Creating your host configuration file
1160 If you added a line containing `Name = yourname' in the main configuarion file,
1161 you will need to create a host configuration file @file{/etc/tinc/netname/hosts/yourname}.
1162 Adapt the following example to create a host configuration file:
1165 Address = @emph{your.real.hostname.org}
1166 Subnet = @emph{192.168.1.0/24}
1169 You can also use an IP address instead of a hostname.
1170 The `Subnet' specifies the address range that is local for @emph{your part of the VPN only}.
1171 If you have multiple address ranges you can specify more than one `Subnet'.
1172 You might also need to add a `Port' if you want your tinc daemon to run on a different port number than the default (655).
1175 @c ==================================================================
1176 @node Generating keypairs, Network interfaces, Configuration files, Configuration
1177 @section Generating keypairs
1179 @cindex key generation
1180 Now that you have already created the main configuration file and your host configuration file,
1181 you can easily create a public/private keypair by entering the following command:
1184 tincd -n @emph{netname} -K
1187 tinc will generate a public and a private key and ask you where to put them.
1188 Just press enter to accept the defaults.
1191 @c ==================================================================
1192 @node Network interfaces, Example configuration, Generating keypairs, Configuration
1193 @section Network interfaces
1195 Before tinc can start transmitting data over the tunnel, it must
1196 set up the virtual network interface.
1198 First, decide which IP addresses you want to have associated with these
1199 devices, and what network mask they must have.
1201 tinc will open a virtual network device (@file{/dev/tun}, @file{/dev/tap0} or similar),
1202 which will also create a network interface called something like `tun0', `tap0', or,
1203 if you are using the Linux tun/tap driver, the network interface will by default have the same name as the netname.
1206 You can configure the network interface by putting ordinary ifconfig, route, and other commands
1207 to a script named @file{/etc/tinc/netname/tinc-up}. When tinc starts, this script
1208 will be executed. When tinc exits, it will execute the script named
1209 @file{/etc/tinc/netname/tinc-down}, but normally you don't need to create that script.
1211 An example @file{tinc-up} script:
1215 ifconfig $INTERFACE 192.168.1.1 netmask 255.255.0.0
1218 This script gives the interface an IP address and a netmask.
1219 The kernel will also automatically add a route to this interface, so normally you don't need
1220 to add route commands to the @file{tinc-up} script.
1221 The kernel will also bring the interface up after this command.
1223 The netmask is the mask of the @emph{entire} VPN network, not just your
1227 @c ==================================================================
1228 @node Example configuration, , Network interfaces, Configuration
1229 @section Example configuration
1233 Imagine the following situation. Branch A of our example `company' wants to connect
1234 three branch offices in B, C and D using the Internet. All four offices
1235 have a 24/7 connection to the Internet.
1237 A is going to serve as the center of the network. B and C will connect
1238 to A, and D will connect to C. Each office will be assigned their own IP
1242 A: net 10.1.0.0 mask 255.255.0.0 gateway 10.1.54.1 internet IP 1.2.3.4
1243 B: net 10.2.0.0 mask 255.255.0.0 gateway 10.2.1.12 internet IP 2.3.4.5
1244 C: net 10.3.0.0 mask 255.255.0.0 gateway 10.3.69.254 internet IP 3.4.5.6
1245 D: net 10.4.0.0 mask 255.255.0.0 gateway 10.4.3.32 internet IP 4.5.6.7
1248 ``gateway'' is the VPN IP address of the machine that is running the
1249 tincd. ``internet IP'' is the IP address of the firewall, which does not
1250 need to run tincd, but it must do a port forwarding of TCP&UDP on port
1251 655 (unless otherwise configured).
1253 In this example, it is assumed that eth0 is the interface that points to
1254 the inner (physical) LAN of the office, although this could also be the
1255 same as the interface that leads to the Internet. The configuration of
1256 the real interface is also shown as a comment, to give you an idea of
1257 how these example host is set up. All branches use the netname `company'
1258 for this particular VPN.
1260 @subsubheading For Branch A
1262 @emph{BranchA} would be configured like this:
1264 In @file{/etc/tinc/company/tinc-up}:
1267 # Real interface of internal network:
1268 # ifconfig eth0 10.1.54.1 netmask 255.255.0.0 broadcast 10.1.255.255
1270 ifconfig $INTERFACE 10.1.54.1 netmask 255.0.0.0
1273 and in @file{/etc/tinc/company/tinc.conf}:
1277 PrivateKeyFile = /etc/tinc/company/rsa_key.priv
1281 On all hosts, /etc/tinc/company/hosts/BranchA contains:
1284 Subnet = 10.1.0.0/16
1287 -----BEGIN RSA PUBLIC KEY-----
1289 -----END RSA PUBLIC KEY-----
1292 Note that the IP addresses of eth0 and tap0 are the same.
1293 This is quite possible, if you make sure that the netmasks of the interfaces are different.
1294 It is in fact recommended to give give both real internal network interfaces and tap interfaces the same IP address,
1295 since that will make things a lot easier to remember and set up.
1298 @subsubheading For Branch B
1300 In @file{/etc/tinc/company/tinc-up}:
1303 # Real interface of internal network:
1304 # ifconfig eth0 10.2.43.8 netmask 255.255.0.0 broadcast 10.2.255.255
1306 ifconfig $INTERFACE 10.2.1.12 netmask 255.0.0.0
1309 and in @file{/etc/tinc/company/tinc.conf}:
1314 PrivateKeyFile = /etc/tinc/company/rsa_key.priv
1317 Note here that the internal address (on eth0) doesn't have to be the
1318 same as on the tap0 device. Also, ConnectTo is given so that no-one can
1319 connect to this node.
1321 On all hosts, in @file{/etc/tinc/company/hosts/BranchB}:
1324 Subnet = 10.2.0.0/16
1327 -----BEGIN RSA PUBLIC KEY-----
1329 -----END RSA PUBLIC KEY-----
1333 @subsubheading For Branch C
1335 In @file{/etc/tinc/company/tinc-up}:
1338 # Real interface of internal network:
1339 # ifconfig eth0 10.3.69.254 netmask 255.255.0.0 broadcast 10.3.255.255
1341 ifconfig $INTERFACE 10.3.69.254 netmask 255.0.0.0
1344 and in @file{/etc/tinc/company/tinc.conf}:
1352 C already has another daemon that runs on port 655, so they have to
1353 reserve another port for tinc. It knows the portnumber it has to listen on
1354 from it's own host configuration file.
1356 On all hosts, in @file{/etc/tinc/company/hosts/BranchC}:
1360 Subnet = 10.3.0.0/16
1363 -----BEGIN RSA PUBLIC KEY-----
1365 -----END RSA PUBLIC KEY-----
1369 @subsubheading For Branch D
1371 In @file{/etc/tinc/company/tinc-up}:
1374 # Real interface of internal network:
1375 # ifconfig eth0 10.4.3.32 netmask 255.255.0.0 broadcast 10.4.255.255
1377 ifconfig $INTERFACE 10.4.3.32 netmask 255.0.0.0
1380 and in @file{/etc/tinc/company/tinc.conf}:
1385 Device = /dev/net/tun
1386 PrivateKeyFile = /etc/tinc/company/rsa_key.priv
1389 D will be connecting to C, which has a tincd running for this network on
1390 port 2000. It knows the port number from the host configuration file.
1391 Also note that since D uses the tun/tap driver, the network interface
1392 will not be called `tun' or `tap0' or something like that, but will
1393 have the same name as netname.
1395 On all hosts, in @file{/etc/tinc/company/hosts/BranchD}:
1398 Subnet = 10.4.0.0/16
1401 -----BEGIN RSA PUBLIC KEY-----
1403 -----END RSA PUBLIC KEY-----
1406 @subsubheading Key files
1408 A, B, C and D all have generated a public/private keypair with the following command:
1414 The private key is stored in @file{/etc/tinc/company/rsa_key.priv},
1415 the public key is put into the host configuration file in the @file{/etc/tinc/company/hosts/} directory.
1416 During key generation, tinc automatically guesses the right filenames based on the -n option and
1417 the Name directive in the @file{tinc.conf} file (if it is available).
1419 @subsubheading Starting
1421 After each branch has finished configuration and they have distributed
1422 the host configuration files amongst them, they can start their tinc daemons.
1423 They don't necessarily have to wait for the other branches to have started
1424 their daemons, tinc will try connecting until they are available.
1427 @c ==================================================================
1428 @node Running tinc, Technical information, Configuration, Top
1429 @chapter Running tinc
1431 If everything else is done, you can start tinc by typing the following command:
1434 tincd -n @emph{netname}
1438 tinc will detach from the terminal and continue to run in the background like a good daemon.
1439 If there are any problems however you can try to increase the debug level
1440 and look in the syslog to find out what the problems are.
1448 @c ==================================================================
1449 @node Runtime options, Error messages, , Running tinc
1450 @section Runtime options
1452 Besides the settings in the configuration file, tinc also accepts some
1453 command line options.
1455 @cindex command line
1456 @cindex runtime options
1460 @item -c, --config=PATH
1461 Read configuration options from the directory PATH. The default is
1462 @file{/etc/tinc/netname/}.
1464 @item -D, --no-detach
1465 Don't fork and detach.
1466 This will also disable the automatic restart mechanism for fatal errors.
1469 @item -d, --debug=LEVEL
1470 Set debug level to LEVEL. The higher the debug level, the more gets
1471 logged. Everything goes via syslog.
1473 @item -k, --kill[=SIGNAL]
1474 Attempt to kill a running tincd (optionally with the specified SIGNAL instead of SIGTERM) and exit.
1475 Use it in conjunction with the -n option to make sure you kill the right tinc daemon.
1476 Under native Windows the optional argument is ignored,
1477 the service will always be stopped and removed.
1479 @item -n, --net=NETNAME
1480 Connect to net NETNAME. @xref{Multiple networks}.
1482 @item -K, --generate-keys[=BITS]
1483 Generate public/private keypair of BITS length. If BITS is not specified,
1484 1024 is the default. tinc will ask where you want to store the files,
1485 but will default to the configuration directory (you can use the -c or -n option
1486 in combination with -K). After that, tinc will quit.
1489 Lock tinc into main memory.
1490 This will prevent sensitive data like shared private keys to be written to the system swap files/partitions.
1492 @item --logfile[=FILE]
1493 Write log entries to a file instead of to the system logging facility.
1494 If FILE is omitted, the default is /var/log/tinc.NETNAME.log.
1496 @item --pidfile=FILE
1497 Write PID to FILE instead of /var/run/tinc.NETNAME.pid.
1499 @item --bypass-security
1500 Disables encryption and authentication.
1501 Only useful for debugging.
1504 Display a short reminder of these runtime options and terminate.
1507 Output version information and exit.
1512 @c ==================================================================
1513 @node Error messages, , Runtime options, Running tinc
1514 @section Error messages
1516 What follows is a list of the most common error messages you can see
1517 when configuring tinc. Most of these messages are visible in the syslog
1518 only, so keep an eye on it!
1521 @item Could not open /dev/tap0: No such device
1524 @item You forgot to `modprobe netlink_dev' or `modprobe ethertap'.
1525 @item You forgot to compile `Netlink device emulation' in the kernel.
1528 @item Can't write to /dev/net/tun: No such device
1531 @item You forgot to `modprobe tun'.
1532 @item You forgot to compile `Universal TUN/TAP driver' in the kernel.
1535 @item Packet with destination 1.2.3.4 is looping back to us!
1538 @item Something is not configured right. Packets are being sent out to the
1539 virtual network device, but according to the Subnet directives in your host configuration
1540 file, those packets should go to your own host. Most common mistake is that
1541 you have a Subnet line in your host configuration file with a prefix length which is
1542 just as large as the prefix of the virtual network interface. The latter should in almost all
1543 cases be larger. Rethink your configuration.
1544 Note that you will only see this message if you specified a debug
1545 level of 5 or higher!
1546 @item Chances are that a `Subnet = ...' line in the host configuration file of this tinc daemon is wrong.
1547 Change it to a subnet that is accepted locally by another interface,
1548 or if that is not the case, try changing the prefix length into /32.
1551 @item Network doesn't work, syslog shows only packets of length 46
1553 @item Network address and prefix length do not match!
1556 @item The Subnet field must contain a @emph{network} address.
1557 @item If you only want to use one IP address, set the netmask to /32.
1560 @item This is a bug: net.c:253: 24: Some error
1563 @item This is something that should not have happened.
1564 Please report this, and tell us exactly what went wrong before you got
1565 this message. In normal operation, these errors should not occur.
1568 @item Error reading RSA key file `rsa_key.priv': No such file or directory
1571 @item You must specify the complete pathname.
1572 Specifying a relative path does not make sense here. tinc changes its
1573 directory to / when starting (to avoid keeping a mount point busy).
1578 @c ==================================================================
1579 @node Technical information, About us, Running tinc, Top
1580 @chapter Technical information
1585 * The meta-protocol::
1590 @c ==================================================================
1591 @node The connection, The meta-protocol, Technical information, Technical information
1592 @section The connection
1595 tinc is a daemon that takes VPN data and transmit that to another host
1596 computer over the existing Internet infrastructure.
1600 * The meta-connection::
1604 @c ==================================================================
1605 @node The UDP tunnel, The meta-connection, The connection, The connection
1606 @subsection The UDP tunnel
1608 @cindex virtual network device
1610 The data itself is read from a character device file, the so-called
1611 @emph{virtual network device}. This device is associated with a network
1612 interface. Any data sent to this interface can be read from the device,
1613 and any data written to the device gets sent from the interface.
1614 There are two possible types of virtual network devices:
1615 `tun' style, which are point-to-point devices which can only handle IPv4 and/or IPv6 packets,
1616 and `tap' style, which are Ethernet devices and handle complete Ethernet frames.
1618 So when tinc reads an Ethernet frame from the device, it determines its
1619 type. When tinc is in it's default routing mode, it can handle IPv4 and IPv6
1620 packets. Depending on the Subnet lines, it will send the packets off to their destination IP address.
1621 In the `switch' and `hub' mode, tinc will use broadcasts and MAC address discovery
1622 to deduce the destination of the packets.
1623 Since the latter modes only depend on the link layer information,
1624 any protocol that runs over Ethernet is supported (for instance IPX and Appletalk).
1625 However, only `tap' style devices provide this information.
1627 After the destination has been determined,
1628 the packet will be compressed (optionally),
1629 a sequence number will be added to the packet,
1630 the packet will then be encrypted
1631 and a message authentication code will be appended.
1633 @cindex encapsulating
1635 When that is done, time has come to actually transport the
1636 packet to the destination computer. We do this by sending the packet
1637 over an UDP connection to the destination host. This is called
1638 @emph{encapsulating}, the VPN packet (though now encrypted) is
1639 encapsulated in another IP datagram.
1641 When the destination receives this packet, the same thing happens, only
1642 in reverse. So it checks the message authentication code, decrypts the contents of the UDP datagram,
1643 checks the sequence number
1644 and writes the decrypted information to its own virtual network device.
1646 If the virtual network device is a `tun' device (a point-to-point tunnel),
1647 there is no problem for the kernel to accept a packet.
1648 However, if it is a `tap' device (this is the only available type on FreeBSD),
1649 the destination MAC address must match that of the virtual network interface.
1650 If tinc is in it's default routing mode, ARP does not work, so the correct destination MAC
1651 can not be known by the sending host.
1652 tinc solves this by letting the receiving end detect the MAC address of its own virtual network interface
1653 and overwriting the destination MAC address of the received packet.
1655 In switch or hub modes ARP does work so the sender already knows the correct destination MAC address.
1656 In those modes every interface should have a unique MAC address, so make sure they are not the same.
1657 Because switch and hub modes rely on MAC addresses to function correctly,
1658 these modes cannot be used on the following operating systems which don't have a `tap' style virtual network device:
1659 OpenBSD, NetBSD, Darwin and Solaris.
1662 @c ==================================================================
1663 @node The meta-connection, , The UDP tunnel, The connection
1664 @subsection The meta-connection
1666 Having only a UDP connection available is not enough. Though suitable
1667 for transmitting data, we want to be able to reliably send other
1668 information, such as routing and session key information to somebody.
1671 TCP is a better alternative, because it already contains protection
1672 against information being lost, unlike UDP.
1674 So we establish two connections. One for the encrypted VPN data, and one
1675 for other information, the meta-data. Hence, we call the second
1676 connection the meta-connection. We can now be sure that the
1677 meta-information doesn't get lost on the way to another computer.
1679 @cindex data-protocol
1680 @cindex meta-protocol
1681 Like with any communication, we must have a protocol, so that everybody
1682 knows what everything stands for, and how she should react. Because we
1683 have two connections, we also have two protocols. The protocol used for
1684 the UDP data is the ``data-protocol,'' the other one is the
1687 The reason we don't use TCP for both protocols is that UDP is much
1688 better for encapsulation, even while it is less reliable. The real
1689 problem is that when TCP would be used to encapsulate a TCP stream
1690 that's on the private network, for every packet sent there would be
1691 three ACKs sent instead of just one. Furthermore, if there would be
1692 a timeout, both TCP streams would sense the timeout, and both would
1693 start re-sending packets.
1696 @c ==================================================================
1697 @node The meta-protocol, Security, The connection, Technical information
1698 @section The meta-protocol
1700 The meta protocol is used to tie all tinc daemons together, and
1701 exchange information about which tinc daemon serves which virtual
1704 The meta protocol consists of requests that can be sent to the other
1705 side. Each request has a unique number and several parameters. All
1706 requests are represented in the standard ASCII character set. It is
1707 possible to use tools such as telnet or netcat to connect to a tinc
1708 daemon started with the --bypass-security option
1709 and to read and write requests by hand, provided that one
1710 understands the numeric codes sent.
1712 The authentication scheme is described in @ref{Authentication protocol}. After a
1713 successful authentication, the server and the client will exchange all the
1714 information about other tinc daemons and subnets they know of, so that both
1715 sides (and all the other tinc daemons behind them) have their information
1722 --------------------------------------------------------------------------
1723 origin ADD_EDGE node1 node2 21.32.43.54 655 222 0
1724 | | | | | +-> options
1725 | | | | +----> weight
1726 | | | +--------> UDP port of node2
1727 | | +----------------> real address of node2
1728 | +-------------------------> name of destination node
1729 +-------------------------------> name of source node
1731 origin ADD_SUBNET node 192.168.1.0/24
1732 | | +--> prefixlength
1733 | +--------> network address
1734 +------------------> owner of this subnet
1735 --------------------------------------------------------------------------
1738 The ADD_EDGE messages are to inform other tinc daemons that a connection between
1739 two nodes exist. The address of the destination node is available so that
1740 VPN packets can be sent directly to that node.
1742 The ADD_SUBNET messages inform other tinc daemons that certain subnets belong
1743 to certain nodes. tinc will use it to determine to which node a VPN packet has
1750 ------------------------------------------------------------------
1751 DEL_EDGE node1 node2
1752 | +----> name of destination node
1753 +----------> name of source node
1755 DEL_SUBNET node 192.168.1.0/24
1756 | | +--> prefixlength
1757 | +--------> network address
1758 +------------------> owner of this subnet
1759 ------------------------------------------------------------------
1762 In case a connection between two daemons is closed or broken, DEL_EDGE messages
1763 are sent to inform the other daemons of that fact. Each daemon will calculate a
1764 new route to the the daemons, or mark them unreachable if there isn't any.
1771 ------------------------------------------------------------------
1772 REQ_KEY origin destination
1773 | +--> name of the tinc daemon it wants the key from
1774 +----------> name of the daemon that wants the key
1776 ANS_KEY origin destination 4ae0b0a82d6e0078 91 64 4
1777 | | \______________/ | | +--> MAC length
1778 | | | | +-----> digest algorithm
1779 | | | +--------> cipher algorithm
1780 | | +--> 128 bits key
1781 | +--> name of the daemon that wants the key
1782 +----------> name of the daemon that uses this key
1785 +--> daemon that has changed it's packet key
1786 --------------------------------------------------------------------------
1789 The keys used to encrypt VPN packets are not sent out directly. This is
1790 because it would generate a lot of traffic on VPNs with many daemons, and
1791 chances are that not every tinc daemon will ever send a packet to every
1792 other daemon. Instead, if a daemon needs a key it sends a request for it
1793 via the meta connection of the nearest hop in the direction of the
1800 --------------------------------------------------------------------------
1803 --------------------------------------------------------------------------
1806 There is also a mechanism to check if hosts are still alive. Since network
1807 failures or a crash can cause a daemon to be killed without properly
1808 shutting down the TCP connection, this is necessary to keep an up to date
1809 connection list. PINGs are sent at regular intervals, except when there
1810 is also some other traffic. A little bit of salt (random data) is added
1811 with each PING and PONG message, to make sure that long sequences of PING/PONG
1812 messages without any other traffic won't result in known plaintext.
1814 This basically covers what is sent over the meta connection by
1818 @c ==================================================================
1819 @node Security, , The meta-protocol, Technical information
1820 @section About tinc's encryption and other security-related issues.
1824 tinc got its name from ``TINC,'' short for @emph{There Is No Cabal}; the
1825 alleged Cabal was/is an organisation that was said to keep an eye on the
1826 entire Internet. As this is exactly what you @emph{don't} want, we named
1827 the tinc project after TINC.
1830 But in order to be ``immune'' to eavesdropping, you'll have to encrypt
1831 your data. Because tinc is a @emph{Secure} VPN (SVPN) daemon, it does
1832 exactly that: encrypt.
1833 tinc by default uses blowfish encryption with 128 bit keys in CBC mode, 32 bit
1834 sequence numbers and 4 byte long message authentication codes to make sure
1835 eavesdroppers cannot get and cannot change any information at all from the
1836 packets they can intercept. The encryption algorithm and message authentication
1837 algorithm can be changed in the configuration. The length of the message
1838 authentication codes is also adjustable. The length of the key for the
1839 encryption algorithm is always the default length used by OpenSSL.
1842 * Authentication protocol::
1843 * Encryption of network packets::
1847 @c ==================================================================
1848 @node Authentication protocol, Encryption of network packets, Security, Security
1849 @subsection Authentication protocol
1851 @cindex authentication
1852 A new scheme for authentication in tinc has been devised, which offers some
1853 improvements over the protocol used in 1.0pre2 and 1.0pre3. Explanation is
1863 --------------------------------------------------------------------------
1864 client <attempts connection>
1866 server <accepts connection>
1870 +-------> name of tinc daemon
1874 +-------> name of tinc daemon
1876 client META_KEY 5f0823a93e35b69e...7086ec7866ce582b
1877 \_________________________________/
1878 +-> RSAKEYLEN bits totally random string S1,
1879 encrypted with server's public RSA key
1881 server META_KEY 6ab9c1640388f8f0...45d1a07f8a672630
1882 \_________________________________/
1883 +-> RSAKEYLEN bits totally random string S2,
1884 encrypted with client's public RSA key
1887 - the client will symmetrically encrypt outgoing traffic using S1
1888 - the server will symmetrically encrypt outgoing traffic using S2
1890 client CHALLENGE da02add1817c1920989ba6ae2a49cecbda0
1891 \_________________________________/
1892 +-> CHALLEN bits totally random string H1
1894 server CHALLENGE 57fb4b2ccd70d6bb35a64c142f47e61d57f
1895 \_________________________________/
1896 +-> CHALLEN bits totally random string H2
1898 client CHAL_REPLY 816a86
1899 +-> 160 bits SHA1 of H2
1901 server CHAL_REPLY 928ffe
1902 +-> 160 bits SHA1 of H1
1904 After the correct challenge replies are received, both ends have proved
1905 their identity. Further information is exchanged.
1907 client ACK 655 123 0
1909 | +----> estimated weight
1910 +--------> listening port of client
1912 server ACK 655 321 0
1914 | +----> estimated weight
1915 +--------> listening port of server
1916 --------------------------------------------------------------------------
1919 This new scheme has several improvements, both in efficiency and security.
1921 First of all, the server sends exactly the same kind of messages over the wire
1922 as the client. The previous versions of tinc first authenticated the client,
1923 and then the server. This scheme even allows both sides to send their messages
1924 simultaneously, there is no need to wait for the other to send something first.
1925 This means that any calculations that need to be done upon sending or receiving
1926 a message can also be done in parallel. This is especially important when doing
1927 RSA encryption/decryption. Given that these calculations are the main part of
1928 the CPU time spent for the authentication, speed is improved by a factor 2.
1930 Second, only one RSA encrypted message is sent instead of two. This reduces the
1931 amount of information attackers can see (and thus use for a cryptographic
1932 attack). It also improves speed by a factor two, making the total speedup a
1935 Third, and most important:
1936 The symmetric cipher keys are exchanged first, the challenge is done
1937 afterwards. In the previous authentication scheme, because a man-in-the-middle
1938 could pass the challenge/chal_reply phase (by just copying the messages between
1939 the two real tinc daemons), but no information was exchanged that was really
1940 needed to read the rest of the messages, the challenge/chal_reply phase was of
1941 no real use. The man-in-the-middle was only stopped by the fact that only after
1942 the ACK messages were encrypted with the symmetric cipher. Potentially, it
1943 could even send it's own symmetric key to the server (if it knew the server's
1944 public key) and read some of the metadata the server would send it (it was
1945 impossible for the mitm to read actual network packets though). The new scheme
1946 however prevents this.
1948 This new scheme makes sure that first of all, symmetric keys are exchanged. The
1949 rest of the messages are then encrypted with the symmetric cipher. Then, each
1950 side can only read received messages if they have their private key. The
1951 challenge is there to let the other side know that the private key is really
1952 known, because a challenge reply can only be sent back if the challenge is
1953 decrypted correctly, and that can only be done with knowledge of the private
1956 Fourth: the first thing that is send via the symmetric cipher encrypted
1957 connection is a totally random string, so that there is no known plaintext (for
1958 an attacker) in the beginning of the encrypted stream.
1961 @c ==================================================================
1962 @node Encryption of network packets, , Authentication protocol, Security
1963 @subsection Encryption of network packet
1966 A data packet can only be sent if the encryption key is known to both
1967 parties, and the connection is activated. If the encryption key is not
1968 known, a request is sent to the destination using the meta connection
1969 to retrieve it. The packet is stored in a queue while waiting for the
1973 The UDP packet containing the network packet from the VPN has the following layout:
1976 ... | IP header | UDP header | seqno | VPN packet | MAC | UDP trailer
1977 \___________________/\_____/
1979 V +---> digest algorithm
1980 Encrypted with symmetric cipher
1983 So, the entire VPN packet is encrypted using a symmetric cipher, including a 32 bits
1984 sequence number that is added in front of the actual VPN packet, to act as a unique
1985 IV for each packet and to prevent replay attacks. A message authentication code
1986 is added to the UDP packet to prevent alteration of packets. By default the
1987 first 4 bytes of the digest are used for this, but this can be changed using
1988 the MACLength configuration variable.
1990 @c ==================================================================
1991 @node About us, Concept Index, Technical information, Top
1996 * Contact Information::
2001 @c ==================================================================
2002 @node Contact Information, Authors, About us, About us
2003 @section Contact information
2006 tinc's website is at @url{http://tinc.nl.linux.org/},
2007 this server is located in the Netherlands.
2010 We have an IRC channel on the FreeNode and OFTC IRC networks. Connect to
2011 @uref{http://www.freenode.net/, irc.freenode.net}
2013 @uref{http://www.oftc.net/, irc.oftc.net}
2014 and join channel #tinc.
2017 @c ==================================================================
2018 @node Authors, , Contact Information, About us
2022 @item Ivo Timmermans (zarq) (@email{ivo@@o2w.nl})
2023 @item Guus Sliepen (guus) (@email{guus@@sliepen.eu.org})
2026 We have received a lot of valuable input from users. With their help,
2027 tinc has become the flexible and robust tool that it is today. We have
2028 composed a list of contributions, in the file called @file{THANKS} in
2029 the source distribution.
2032 @c ==================================================================
2033 @node Concept Index, , About us, Top
2034 @c node-name, next, previous, up
2035 @unnumbered Concept Index
2037 @c ==================================================================
2041 @c ==================================================================