System Administration Guide: IP Services

How to Protect a VPN With an IPsec Tunnel in Transport Mode Over IPv4

This procedure extends the procedure, How to Secure Traffic Between Two Systems With IPsec. In addition to connecting two systems, you are connecting two intranets that connect to these two systems. The systems in this procedure function as gateways.

This procedure assumes the same setup as in How to Protect a VPN With an IPsec Tunnel in Tunnel Mode Over IPv6.

On the system console on one of the systems, assume the Primary Administrator role or become superuser.
The Primary Administrator role includes the Primary Administrator profile. To create the role and assign the role to a user, see Chapter 2, "Working With the Solaris Management Console (Tasks)," in System Administration Guide: Basic Administration.

Note - Logging in remotely exposes security-critical traffic to eavesdropping. Even if you somehow protect the remote login, the security of the system is reduced to the security of the remote login session.

Control the flow of packets before configuring IPsec.

Ensure that IP forwarding and IP dynamic routing are disabled.

# routeadm
Configuration       Current         Current
       Option       Configuration  System State
--------------------------------------------------
IPv4 forwarding     disabled           disabled
   IPv4 routing     default (enabled)   enabled
...

If forwarding and routing are enabled, you can disable them by typing:

# routeadm -d ipv4-routing -d ipv4-forwarding
# routeadm -u

Turning off IP forwarding prevents packets from being forwarded from one network to another network through this system. For a description of the routeadm command, see the routeadm(1M) man page.

Turn on IP strict destination multihoming.
# ndd -set /dev/ip ip_strict_dst_multihoming 1
Turning on IP strict destination multihoming ensures that packets for one of the system's destination addresses arrive at the correct destination address.

When you turn off IP forwarding and turn on IP strict destination multihoming, fewer packets flow all the way through the system. When strict destination multihoming is enabled, packets that arrive on a particular interface must be addressed to one of the local IP addresses of that interface. All other packets, even ones that are addressed to other local addresses of the system, are dropped.

Disable most network services, and possibly all network services.

Note - If your system was installed with the "limited" SMF profile, then you can skip this step. Network services, with the exception of Solaris Secure Shell, are disabled.

The disabling of network services prevents IP packets from doing any harm to the system. For example, an SNMP daemon, a telnet connection, or an rlogin connection could be exploited.

You have two options.

Run the "limited" SMF profile.
# netservices limited

Or, individually disable network services.

# svcadm disable network/ftp:default
# svcadm disable network/finger:default
# svcadm disable network/login:rlogin
# svcadm disable network/nfs/server:default
# svcadm disable network/rpc/rstat:default
# svcadm disable network/smtp:sendmail
# svcadm disable network/telnet:default
# svcs | grep network
online         Aug_02   svc:/network/loopback:default
...
online         Aug_09   svc:/network/ssh:default
...

On each system, add a pair of SAs between the two systems.
Choose one of the following options:
- Configure IKE to manage the keys for the SAs. Use one of the procedures in Configuring IKE (Task Map) to configure IKE for the VPN.
- If you have an overriding reason to manually manage the keys, see How to Manually Create IPsec Security Associations.

On each system, add IPsec policy.

Edit the /etc/inet/ipsecinit.conf file to add the IPsec policy for the VPN. To strengthen the policy, see Example 20-10.

For example, on the enigma system, type the following entry into the ipsecinit.conf file:

# LAN traffic to and from this host can bypass IPsec.
{laddr 10.1.3.3 dir both} bypass {}

# WAN traffic uses ESP with 3DES and MD5.
{tunnel ip.tun0 negotiate transport} 
 ipsec {encr_algs 3des encr_auth_algs md5 sa shared}

On the partym system, type the following entry into the ipsecinit.conf file:

# LAN traffic to and from this host can bypass IPsec.
{laddr 10.1.3.3 dir both} bypass {}

# WAN traffic uses ESP with 3DES and MD5.
{tunnel ip.tun0 negotiate transport} 
 ipsec {encr_algs 3des encr_auth_algs md5 sa shared}

On each system, configure the tunnel, ip.tun0.

Use the following ifconfig commands to create the point-to-point interface:

# ifconfig ip.tun0 plumb

# ifconfig ip.tun0 system1-point system2-point \
tsrc system1-taddr tdst system2-taddr

For example, on the enigma system, type the following commands:

# ifconfig ip.tun0 plumb

# ifconfig ip.tun0 10.16.16.6 10.1.3.3 \
tsrc 192.168.116.16 tdst 192.168.13.213

On the partym system, type the following commands:

# ifconfig ip.tun0 plumb

# ifconfig ip.tun0 10.1.3.3 10.16.16.6  \
tsrc 192.168.13.213 tdst 192.168.116.16

Protect the tunnel with the IPsec policy that you created.
# ipsecconf

Bring up the router for the tunnel.
# ifconfig ip.tun0 router up
On each system, turn on IP forwarding for the hme1 interface.
# ifconfig hme1 router
IP forwarding means that packets that arrive from somewhere else can be forwarded. IP forwarding also means that packets that leave this interface might have originated somewhere else. To successfully forward a packet, both the receiving interface and the transmitting interface must have IP forwarding turned on.

Because the hme1 interface is inside the intranet, IP forwarding must be turned on for hme1. Because ip.tun0 connects the two systems through the Internet, IP forwarding must be turned on for ip.tun0.

The hme0 interface has its IP forwarding turned off to prevent an outside adversary from injecting packets into the protected intranet. The outside refers to the Internet.
On each system, ensure that routing protocols do not advertise the default route within the intranet.
# ifconfig hme0 private
Even if hme0 has IP forwarding turned off, a routing protocol implementation might still advertise the interface. For example, the in.routed protocol might still advertise that hme0 is available to forward packets to its peers inside the intranet. By setting the interface's private flag, these advertisements are prevented.
Manually, add a default route over hme0.
The default route should be a router with direct access to the Internet.
# route add default router-on-hme0-subnet
1. For example, on the enigma system, add the following route:
  # route add default 192.168.116.4
2. On the partym system, add the following route:
  # route add default 192.168.13.5
  Even though the hme0 interface is not part of the intranet, hme0 does need to reach across the Internet to its peer system. To find its peer, hme0 needs information about Internet routing. The VPN system appears to be a host, rather than a router, to the rest of the Internet. Therefore, you can use a default router or run the router discovery protocol to find a peer system. For more information, see the route(1M) and in.routed(1M) man pages.

Ensure that the VPN starts after a reboot by adding an entry to the /etc/hostname.ip.tun0 file.

system1-point system2-point tsrc system1-taddr \
tdst system2-taddr encr_algs 3des encr_auth_algs md5 router up

For example, on the enigma system, add the following entry to the hostname.ip.tun0 file:
10.16.16.6 10.1.3.3 tsrc 192.168.116.16 \ tdst 192.168.13.213 router up

On the partym system, add the following entry to the hostname.ip.tun0 file:

10.1.3.3 10.16.16.6 tsrc 192.168.13.213 \
tdst 192.168.116.16 router up

On each system, configure the interface files to pass the correct parameters to the routing daemon.
1. On the enigma system, modify the /etc/hostname.interface files.
  # cat enigma hostname.hme0 10.16.16.6 private
  # cat enigma hostname.hme1 192.168.116.16 router
2. On the partym system, modify the /etc/hostname.interface files.
  # cat partym hostname.hme0 10.1.3.3 private
  # cat partym hostname.hme1 192.168.13.213 router
On each system, run a routing protocol.
You might need to configure the routing protocol before enabling routing. For more information, see Routing Protocols in the Solaris OS. For a procedure, see How to Configure an IPv4 Router.
# routeadm -e ipv4-routing # routeadm -u

Example 20-10 Requiring IPsec Policy on All Systems in Transport Mode

In this example, the administrator comments out the bypass policy that was configured in Step 4, thereby strengthening the protection. With this policy configuration, each system on the LAN must activate IPsec to communicate with the router.

# LAN traffic must implement IPsec.
# {laddr 10.1.3.3 dir both} bypass {}

# WAN traffic uses ESP with 3DES and MD5.
{tunnel ip.tun0 negotiate transport} ipsec {encr_algs 3des encr_auth_algs md5}

Example 20-11 Using Deprecated Syntax to Configure an IPsec Tunnel in Transport Mode

In this example, administrator is connecting a Solaris Express system with a system that is running the Solaris 10 release. Therefore, the administrator uses Solaris 10 syntax in the configuration file, and includes the IPsec algorithms in the ifconfig command.

In Step 4, the syntax is the following:

# LAN traffic to and from this address can bypass IPsec.
{laddr 10.1.3.3 dir both} bypass {}

# WAN traffic uses ESP with 3DES and MD5.
{} ipsec {encr_algs 3des encr_auth_algs md5}

For Step 5 to Step 7, the syntax is the following:

# ifconfig ip.tun0 plumb

# ifconfig ip.tun0 10.16.16.6 10.1.3.3 \
tsrc 192.168.116.16 tdst 192.168.13.213 \
encr_algs 3des encr_auth_algs md5

# ifconfig ip.tun0 router up

# ifconfig ip.tun0 plumb

# ifconfig ip.tun0 10.16.16.6 10.1.3.3 \
tsrc 192.168.116.16 tdst 192.168.13.213 \
encr_algs 3des encr_auth_algs md5

The IPsec policy that is passed to the ifconfig commands must be the same as the IPsec policy in the ipsecinit.conf file. Upon reboot, each system reads the ipsecinit.conf file for its policy.

In Step 11, the syntax is the following:

10.16.16.6 10.1.3.3 tsrc 192.168.116.16 \
tdst 192.168.13.213 encr_algs 3des encr_auth_algs md5 router up