There are two kinds of IPSEC available for Linux these days. For 2.2 and 2.4, there is FreeS/WAN, which was the first major implementation. They have an official site and an unofficial one that is actually maintained. FreeS/WAN has traditionally not been merged with the mainline kernel for a number of reasons. Most often mentioned are 'political' issues with Americans working on crypto tainting its exportability. Furthermore, it does not integrate too well with the Linux kernel, leading it to be a bad candidate for actual merging.
As of Linux 2.5.47, there is a native IPSEC implementation in the kernel. It was written by Alexey Kuznetsov and Dave Miller, inspired by the work of the USAGI IPv6 group. With its merge, James Morris' CrypoAPI also became part of the kernel - it does the actual crypting.
This HOWTO will only document the 2.5+ version of IPSEC. FreeS/WAN is recommended for Linux 2.4 users for now, but be aware that its configuration will differ from the native IPSEC.
As of 2.5.49, IPSEC works without further patches.
Userspace tools appear are available here, they are actively being maintained.
When compiling your kernel, be sure to turn on 'PF_KEY', 'AH', 'ESP' and everything in the CryptoAPI! TCP_MSS netfilter target is currently broken, you need to turn it off.
The author of this chapter is a complete IPSEC nitwit! If you find the inevitable mistakes, please email bert hubert <firstname.lastname@example.org>.
First, we'll show how to manually setup secure communication between two hosts. A large part of this process can also be automated, but here we'll do it by hand so as to acquaint ourselves with what is going on 'under the hood'.
Feel free to skip the following section if you are only interested in automatic keying but be aware that some understanding of manual keying is useful.
IPSEC is a complicated subject. A lot of information is available online, this HOWTO will concentrate on getting you up and running and explaining the basic principles.
Many iptables configurations drop IPSEC packets! To pass IPSEC, use: 'iptables -A xxx -p 50 -j ACCEPT' and 'iptables -A xxx -p 51 -j ACCEPT'
IPSEC offers a secure version of the Internet Protocol. Security in this context means two different things: encryption and authentication. A naive vision of security offers only encryption but it can easily be shown that is insufficient - you may be communicating encyphered, but no guarantee is offered that the remote party is the one you expect it to be.
IPSEC supports 'Encapsulated Security Payload' (ESP) for encryption and 'Authentication Header' (AH) for authenticating the remote partner. You can configure both of them, or decided to do only either.
Both ESP and AH rely on security associations. A security association (SA) consists of a source, a destination and an instruction. A sample authentication SA may look like this:
add 10.0.0.11 10.0.0.216 ah 15700 -A hmac-md5 "1234567890123456";This says 'traffic going from 10.0.0.11 to 10.0.0.216 that needs an AH can be signed using HMAC-MD5 using secret 1234567890123456'. This instruction is labelled with SPI ('Security Parameter Index') id '15700', more about that later. The interesting bit about SAs is that they are symmetrical. Both sides of a conversation share exactly the same SA, it is not mirrored on the other side. Do note however that there is no 'autoreverse' rule - this SA only describes a possible authentication from 10.0.0.11 to 10.0.0.216. For two-way traffic, two SAs are needed.
A sample ESP SA:
add 10.0.0.11 10.0.0.216 esp 15701 -E 3des-cbc "123456789012123456789012";This says 'traffic going from 10.0.0.11 to 10.0.0.216 that needs encryption can be encyphered using 3des-cbc with key 123456789012123456789012'. The SPI id is '15701'.
So far, we've seen that SAs describe possible instructions, but do not in fact describe policy as to when these need to be used. In fact, there could be an arbitrary number of nearly identical SAs with only differing SPI ids. Incidentally, SPI stands for Security Parameter Index. To do actual crypto, we need to describe a policy. This policy can include things as 'use ipsec if available' or 'drop traffic unless we have ispec'.
A typical simple Security Policy (SP) looks like this:
spdadd 10.0.0.216 10.0.0.11 any -P out ipsec esp/transport//require ah/transport//require;If entered on host 10.0.0.216, this means that all traffic going out to 10.0.0.11 must be encrypted and be wrapped in an AH authenticating header. Note that this does not describe which SA is to be used, that is left as an exercise for the kernel to determine.
In other words, a Security Policy specifies WHAT we want; a Security Association describes HOW we want it.
Outgoing packets are labelled with the SA SPI ('the how') which the kernel used for encryption and authentication so the remote can lookup the corresponding verification and decryption instruction.
What follows is a very simple configuration for talking from host 10.0.0.216 to 10.0.0.11 using encryption and authentication. Note that the reverse path is plaintext in this first version and that this configuration should not be deployed.
On host 10.0.0.216:
#!/sbin/setkey -f add 10.0.0.216 10.0.0.11 ah 24500 -A hmac-md5 "1234567890123456"; add 10.0.0.216 10.0.0.11 esp 24501 -E 3des-cbc "123456789012123456789012"; spdadd 10.0.0.216 10.0.0.11 any -P out ipsec esp/transport//require ah/transport//require;
On host 10.0.0.11, the same Security Associations, no Security Policy:
#!/sbin/setkey -f add 10.0.0.216 10.0.0.11 ah 24500 -A hmac-md5 "1234567890123456"; add 10.0.0.216 10.0.0.11 esp 24501 -E 3des-cbc "123456789012123456789012";
With the above configuration in place (these files can be executed if 'setkey' is installed in /sbin), 'ping 10.0.0.11' from 10.0.0.216 looks like this using tcpdump:
22:37:52 10.0.0.216 > 10.0.0.11: AH(spi=0x00005fb4,seq=0xa): ESP(spi=0x00005fb5,seq=0xa) (DF) 22:37:52 10.0.0.11 > 10.0.0.216: icmp: echo replyNote how the ping back from 10.0.0.11 is indeed plainly visible. The forward ping cannot be read by tcpdump of course, but it does show the Security Parameter Index of AH and ESP, which tells 10.0.0.11 how to verify the authenticity of our packet and how to decrypt it.
A few things must be mentioned however. The configuration above is shown in a lot of IPSEC examples and it is very dangerous. The problem is that the above contains policy on how 10.0.0.216 should treat packets going to 10.0.0.11, and that it explains how 10.0.0.11 should treat those packets but it does NOT instruct 10.0.0.11 to discard unauthenticated or unencrypted traffic!
Anybody can now insert spoofed and completely unencrypted data and 10.0.0.11 will accept it. To remedy the above, we need an incoming Security Policy on 10.0.0.11, as follows:
#!/sbin/setkey -f spdadd 10.0.0.216 10.0.0.11 any -P IN ipsec esp/transport//require ah/transport//require;This instructs 10.0.0.11 that any traffic coming to it from 10.0.0.216 is required to have valid ESP and AH.
Now, to complete this configuration, we need return traffic to be encrypted and authenticated as well of course. The full configuration on 10.0.0.216:
#!/sbin/setkey -f flush; spdflush; # AH add 10.0.0.11 10.0.0.216 ah 15700 -A hmac-md5 "1234567890123456"; add 10.0.0.216 10.0.0.11 ah 24500 -A hmac-md5 "1234567890123456"; # ESP add 10.0.0.11 10.0.0.216 esp 15701 -E 3des-cbc "123456789012123456789012"; add 10.0.0.216 10.0.0.11 esp 24501 -E 3des-cbc "123456789012123456789012"; spdadd 10.0.0.216 10.0.0.11 any -P out ipsec esp/transport//require ah/transport//require; spdadd 10.0.0.11 10.0.0.216 any -P in ipsec esp/transport//require ah/transport//require;
And on 10.0.0.11:
#!/sbin/setkey -f flush; spdflush; # AH add 10.0.0.11 10.0.0.216 ah 15700 -A hmac-md5 "1234567890123456"; add 10.0.0.216 10.0.0.11 ah 24500 -A hmac-md5 "1234567890123456"; # ESP add 10.0.0.11 10.0.0.216 esp 15701 -E 3des-cbc "123456789012123456789012"; add 10.0.0.216 10.0.0.11 esp 24501 -E 3des-cbc "123456789012123456789012"; spdadd 10.0.0.11 10.0.0.216 any -P out ipsec esp/transport//require ah/transport//require; spdadd 10.0.0.216 10.0.0.11 any -P in ipsec esp/transport//require ah/transport//require;
Note that in this example we used identical keys for both directions of traffic. This is not in any way required however.
To examine the configuration we just created, execute setkey -D, which shows the Security Associations or setkey -DP which shows the configured policies.